Professional Documents
Culture Documents
Dow Vinerobot (610953) 2013-09-10
Dow Vinerobot (610953) 2013-09-10
Part A
Workplan Tables
Starting date
4
01/12/2013
Duration in months
5
36
Abstract
9
The aim of this project is the design, development, and deployment of a novel use-case agricultural robot under
the scope of Unmanned Ground Vehicles (UGV), and equipped with several non-invasive sensing technologies
to monitor: 1) of grape yield, 2) vegetative growth, 3) Water stress and 4) Grape composition in order to optimize
the vineyard management and improve grape composition and wine quality.
The use of UGVs to monitor vineyard physiological parameters and grape composition has several advantages
over simple hand-held devices for manual sampling and over UAV.
UGVs may provide key information regarding vineyard physiology and water stress, grape composition, and
pests and diseases’ incidence much faster than manual solutions and at higher resolution, in a more flexible way,
and lower cost than aerial scouting carried out by drones or planes.
The scope covers the integral monitoring of vineyards, over the entire season, by placing a ground robot along
the vineyards (endowed with artificial intelligence and machine learning techniques).
The final users will receive updated in mobile application (app).
The proposed use-case agricultural robots (where key proximal sensing technologies will be implemented) will
allow revolutionary and conclusive decision making to optimize vineyard management and to drive agronomical
fundamental decisions according to grape yield estimation, plant growth monitoring, water status, and berry
composition assessment. UGV proposed will incorporate an integrated system that includes machine vision,
thermography and fluorescence-based sensors. Canopy images and data acquisition processes will be executed
in real time through customized algorithms to compose specific production maps.
List of Beneficiaries
Note that the budget mentioned in this table is the total budget requested by the Beneficiary and associated Third Parties.
610953
Project title
VINEROBOT—VINEyardROBOT
Call (part) identifier
FP7-ICT-2013-10
Funding scheme
Collaborative project
WT1
List of work packages
Project Number
1
610953 Project Acronym
2
VINEROBOT
Delive- Estimated
WP Dissemi-
rable Lead benefi- indicative Delivery date
Deliverable Title number Nature
62 nation level
Number 53 ciary number person- 63
64
61
months
Report on robot
specifications
D1.1 defining what the 1 6 8.50 R CO 12
robot needs to
do (M12)
Rate of
achievement
protocol for
D1.2 1 6 0.50 R RE 6
vine robot
performance
(M6)
Report and
actual robot
D2.1 prototype 2 3 60.00 R RE 18
endowed with
safe mobility
Report on
physical
D2.2 integration of 2 3 60.00 R RE 24
perception unit
in prototype
Field
D2.3 2 3 25.50 R PU 19
demonstration
Instructions
D2.4 manual of 2 3 38.00 R RE 32
optimized robot
Report on
sensor selection,
D3.1 adaptation, 3 2 75.00 R RE 18
and field
performance
Final report
on algorithms,
models,
calibration
curves and
D3.2 3 2 65.70 R CO 36
validation
tests of
agronomical and
physiological
parameters
Report on
D4.1 data fusion of 4 7 40.00 R RE 24
perception and
61
months
localization
sensors in real
time
Multi-parameter
map generation
D4.2 4 4 34.50 R RE 32
and data
management
Application
of display of
D4.3 4 4 12.00 D RE 35
information on
mobile devices
User’s manual
D4.4 4 4 20.00 R RE 36
for prototype
Prototype demo
D4.5 and presentation 4 7 45.00 D PU 32
to users
Project website
D5.1 and presentation 5 8 4.00 O PU 1
video
First report of the
exploitation and
D5.2 dissemination 5 8 4.00 R RE 12
of knowledge
activity
Second
report of the
exploitation and
D5.3 5 8 6.00 R RE 24
dissemination
of knowledge
activity
Third report
of the
exploitation and
D5.4 5 8 6.00 R RE 36
dissemination
of knowledge
activity
E-mail and
data exchange
D6.1 network between 6 1 2.80 O CO 1
coordinator and
partners
First project
D6.2 6 1 6.00 R CO 6
progress report
First evaluation
D6.3 of the Advisory 6 1 3.00 R RE 12
Board
61
months
Project progress
D6.4 6 1 8.00 R CO 18
technical report
Project progress
D6.5 6 1 8.00 R CO 30
technical report
Total 532.50
Project Number
1
610953 Project Acronym
2
VINEROBOT
Objectives
The objective of WP1 is to establish the technical requirements for the multi-sensor scouting ROBOT from SMEs
end users.
Description of work (possibly broken down into tasks), and role of participants
WP1 will be led by GEIS and involves SMEs and RTD performers. WP1 has four tasks:
Several parameters can be measured and the final user requirements will be discussed according to the
expected techniques involved and the required accuracy. The agronomical variables involved in the
development of the project should be established and ranked to guarantee the success and the achievement
of the objectives. Success rate will be determined through the benchmarking protocol defined in D1.2. Using
this information, the need and way to obtain vineyard parameters of interest taken into account the technology
available will be completed. The result of this task should guide the research related to the creation of the
scouting robot capable of providing useful information available to the grower. The selection of parameters
to monitor will follow an iterative process in such a way that the estimation of physiological traits in vines will
increase their score in the benchmarking protocol of D1.2 as the project advances.
The agronomical and physiological requirements will be reported partly in D1.1 and D1.2 (benchmarking
protocol)
It will be established from the previous task what is exactly the current state of the technology to acquire the
wanted information from the vineyard on-the-go, and the need to perform adaptations or developments of the
current sensors to fit the problem under the real working conditions. Furthermore, the knowledge of VIBU will
bring the working environment and the real needs of the sector to determine the actual requirements and the
information to be provided. Using this information, engineering solutions will be identified for the problems of
installing all the sensors on a robot that is moving through the vineyard robustly obtaining data, and decisions
will be taken about the equipment that best suits in the project development.
The requirements for the robot will be reported in D1.1 and D1.2 (benchmarking protocol)
All the equipment used in the robot requires the definition of several experimental set-up parameters. Hence
it is necessary to determine the positioning, distance, plugs and on-the-go acquisition process. In this task,
the best approach to obtain the data that fulfil the grower needs will be analysed and determined, leading to
The requirements for the data fusion approach will be reported in D1.1 and D1.2 (benchmarking protocol)
It would be important to determine the protocol for the data acquisition while the vehicle advances to ensure that
the data acquired from all sensors actually belong to the same scene avoiding overlapping and gapping that may
cause confusion to the decision system. All data will be acquired, pre-processed, sent to the cloud, processed
in a remote server and offered in real-time to non-experienced growers in an easy and practical way by means
of different devices like smart phones, tablets or control desks using different operating systems. But also
the data should be stored for further management, double-checking or analysis. The task 1.1 is providing the
important parameters and how to present them to the final users, which will obtain the status of the vineyard and
mapping for the evaluated hectares. This information will be used to set the processing needs and platforms, the
transmission protocols, or the display systems and devices.
The new product is targeted to the vineyard manager, i.e., the person who takes decisions about the vineyard’s
management or provides consultancy to grape growers and/or winery managers. Therefore, end-users can be:
i) technicians of grower’s consortia, cooperatives, or large farm enterprises and ii) external consultants. The
robot is meant to work in any kind of environment and hopefully available at a moderate cost. The idea is to
keep complexity level below the difficulty of cellphone browsing, therefore the robot is aimed at non-experts in
information technologies. The verification of data can only be conducted by comparison to ground-truth data
measured with alternative (agronomical destructive) methods. In that respect, the participation of the winery
VIBU is key to assure the proper verification of the estimates automatically issued by the robot. If, for instance,
a prediction on yield is output in a map, manual harvesting will be necessary to assess the degree of precision
reached with the predictive model programmed in the robot. The final data will come in a map format, compatible
with all the maps generated by the robot regardless of the sensors used. The specific format of the maps still
needs to be investigated in WP1, WP3, and WP4, but it will be both visually (to help decision-making “on the
spot”) and numerically to permit statistical analysis and the creation of historical databases.
The requirements for data management and display will be reported in D1.1 and D1.2 (benchmarking protocol)
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 1.00
2 FORC 1.00
3 WALL 1.00
4 SIVI 1.00
5 VIBU 1.00
Participant number
10
Participant short name
11
Person-months per participant
6 GEIS 2.00
7 UPV 1.00
8 AVAN 1.00
Total 9.00
List of deliverables
Description of deliverables
D1.1) Report on robot specifications defining what the robot needs to do (M12): Report on robot specifications
defining what the robot needs to do. Agronomical, physiological, sensor, robot requirements. [month 12]
D1.2) Rate of achievement protocol for vine robot performance (M6): Benchmarking protocol under T 1.1. Rate
of achievement protocol for vine robot performance. [month 6]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
D 1.1. Definition of
robot specifications
in terms of sensing
MS1 Sensing needs for robot 6 12
behaviors and capacities
needed for meting user
requirements.
Objectives
In this WP, the integration of several sensing and navigation techniques into the scouting robot is aimed to
obtain a large amount of agronomical and physiological data at the vineyard scale.
It was not explicitly mentioned in the initial proposal but work-packages 2, 3, and 4 will require many iterations
until the platform can behave satisfactorily. In the new description of the work plan it is stated that the design
process of the platform (WP2) and its supporting sensors (WP3, WP4) will follow an iterative philosophy.
Description of work
WP2 will be led by WALL which is an SME with great experience with robots and involves SMEs and RTD
performers. WP2 has six tasks:
WALL in collaboration with the Technical Manager will design a four-wheel-drive medium size robotic platform.
After its physical construction it will be endowed with necessary algorithms for navigation and obstacle detection.
Once the robot can safely navigate along vineyard rows, crop detection sensors will be integrated in Task 2.2.
The accomplishment of Task 2.1 involves the selection, calibration, and integration of such navigation sensors
as GPS receiver, wheel optical encoders, and a laser rangefinder for safeguarding. Additionally, actuators for
braking and steering the wheels will be necessary, as well as emergency-stop buttons. Important key points are
autonomy and speed to provide large surface maps of the vineyard without human intervention.
A basic platform developed by WALL will be the initial point after which the design of the vine robot will improve
mobility in uncertain terrains and will endow the new robot with a high-performance safeguarding engine. Later
on, the perceptive abilities required to construct the maps will be added along tasks WP3 and WP4. Mapping,
therefore, will be accomplished by registering data with several sensors and fusing them in compatible maps
composed from GPS localization data. The change of row at the headlands will require the successful integration
of a global positioning device (GPS) with local perception sensors (machine vision or laser/lidar), all supervised
by a path-planner algorithm under the higher-level supervision of the safety unit. The path-planner algorithms
will be developed under WP2, but previous work conducted in robotics, with special attention paid at the
results published from DARPA Grand Challenge (autonomous vehicle competition) will be carefully examined.
Long-term autonomous navigation including headland turning is one of the key challenges that still remain to be
solved, and to which we hope to contribute with this project.
Task 2.2. Design of sensory system for the detection of canopy and fruit under dynamic conditions
The aim of this task is to design a sensory system and its supporting algorithms for detecting and localizing
canopy and fruits under dynamic conditions. For measurements on leaves, the solution will allow the adjustment
of the position (height, distance) of different agronomical sensors by taking into account the structure of the
canopy. A priori, to realize a map of ripeness only few bunches are necessary, so the solution will select
Task 2.3. Design of sensory system for physiological and agronomical parameters of vineyards under dynamic
conditions
Based on task 1.2, sensors used to determine physiological and agronomical parameters of vineyards will be
integrated on the robot. This integration requires improvement or adaptation of each sensor to take into account
mechanical, electronics, and communication constraints. In particular, autonomy and weight are important key
point to realize large map per day without human intervention.
At the end of this work package, all sensors must be developed and tested to facilitate their mechanical
integration and connection with the robot. The results of this task are:
a) To confirm and validate the operation of the sensors for acquiring data working altogether. Tests should span
an entire season.
b) To perform specific tests for each of the sensors separately. To check that the LPCU can control the sensors
so they are able to acquire data accurately and synchronized.
c) To ensure the coordination and sampling-rate compatibility among sensors.
d) To perform tests to guarantee repeatability and reproducibility of the entire system under vineyard conditions
to obtain a robust system capable of achieving the same results under the same conditions.
At the end of task 2.4, the robotic platform will be an inextricable mix of cables, sensors, and mechanical
components. The goal of this task is the optimized arrangement of the robot parts for its best performance in
terms of safety and usability for the other partners of the project. At the end of this task the robot must have a
professional commercially-oriented appearance.
Two public demonstrations of the robot will be scheduled in the middle (M19) and at the end of the project (M32).
The former will show the progress conducted on the platform, especially regarding navigation and safeguarding,
and the latter will focus on the integration of the perception sensors.
First progress demonstration in M19 (June 2nd year): This demo will basically involve WP2 and WP4, and its
main objective is to update all the partners of the progress made to gather feedback, as well as informing the
European Commission on the status of the VINEROBOT project. The demonstration will take place along the
all-partner mid-project general meeting, organized by the Project Coordinator and Technical Manager.
Second progress demonstration in M32 (July 3rd year). This demo will involve all the partners and will take
place in the vineyards of VIBU (France), where the robotic platform will be challenged for full capability in
commercial vineyards. Feedback will be gathered to introduce the final improvements on the platform for
enhancing functionality from a user-interface standpoint. Feedback will also be welcomed from EC auditors.
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 13.00
2 FORC 5.00
3 WALL 55.00
4 SIVI 18.00
5 VIBU 1.00
6 GEIS 8.00
7 UPV 37.50
8 AVAN 46.00
Total 183.50
List of deliverables
Description of deliverables
D2.1) Report and actual robot prototype endowed with safe mobility: Report and actual robot prototype endowed
with safe mobility [month 18]
D2.2) Report on physical integration of perception unit in prototype: Report on physical integration of perception
unit in prototype [month 24]
D2.3) Field demonstration: Field demonstration. Basic mobility and navigation of the robot. [month 19]
D2.4) Instructions manual of optimized robot: Instructions manual of optimized robot [month 32]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
D 2.1 Demo 1.
Deployment of robotic
MS2 Construction of robotic platform 3 18
platform with navigation
capabilities.
D 2.2 Integration of
Assembly of multi-perception head on robotic perception head in
MS3 3 24
platform robot, adding sensing
capacities to navigation
Objectives
The objective of WP3 involves the development of the algorithms, calibration curves and models that process
the data acquired by the multi-sensor system of the scouting robot by integrating and fusing data coming
from agronomical and physiological information. It was not explicitly mentioned in the initial proposal but this
work-package will require many iterations until the platform can behave satisfactorily.
Description of work
WP3 will be led by FORC and involves SMEs and RTD performers. WP3 has five tasks:
An initial evaluation of the parameters under study will be performed under laboratory conditions to obtain
reference data. Linear regression analysis will be used to find out correlations between the canopy parameters
under study measured manually in the field or in the laboratory, and the data obtained by the integrated sensors.
Multivariate analysis techniques will be used to infer results from the comparison between the features extracted
using destructive methods and the data extracted from the images, infrared, and fluorescence sensors. Principal
component analysis (PCA) will be used to group data coming from destructive and wet chemistry analysis,
along with the optical properties obtained by the sensors to make predictions about the vegetative status of the
vineyard. Discriminant analysis will be used to classify the different vineyard status regarding of the variables
measured by the sensors.
From the results obtained in the previous task, it will be possible to determine those variables extracted from
the measurement of the sensors that best correlate with the reference data, and hence to find out the need to
combine information from different sensors or to identify some variables that can be used directly from the raw
data to determine important properties of the vineyard. The aim would be to find the minimal set of data that
maximizes the result. To identify single variables and to estimate some important properties without combining
information from all sensors can save considerable processing time and data transmission load, which will
translate into better system performance. In other cases it will be necessary to build models using all variables
in the data set to build classifiers optimised using sensor fusion techniques as detailed in Task 3.3. Once the
models are generated and the classifiers built, they will be implemented in the control unit. A strategy to follow
will be the creation of look-up tables stored in the computer’s memory whose inputs will be the variables in the
system (data extracted from the sensor’s measurements) and the output the prediction of the parameters under
study.
Data provided by different sensors based on fluorescence, visible images, and infrared images, must be
combined to make a fast and unique decision on-the-go so the system can offer the user real-time information
about the vineyard. The sensor fusion techniques combine information from different heterogeneous sources to
provide a single response. These techniques are usually grouped into three categories: i) positioning techniques,
ii) identification techniques, and iii) pre-processing techniques. In order to meet the objectives of this task, a
three-step process will be followed: extraction of data from each sensor individually, individual feature extraction
from the data, and decision-making by combining the results obtained by the assembly of sensors. Several
statistical methods will be explored to help making this final decisions, such as classical inference, Bayesian
inference, the Dempster-Shafer method, or heuristics.
The final data to be presented to the end-user will have the format of compatible maps in a global-based
local-coordinated frame. The detection of grape bunches, for example, will likely result in yield maps
or bunch-size maps, which may be confronted to vegetation maps or irrigation prescription maps, all
fully-comparable as they will be given in the same coordinate system and sharing the same coordinate system.
For the agronomical and physiological parameters identified in previous tasks, calibration curves (covering the
ranges present in the vineyard) with data acquired by the different techniques embedded in the multi-sensor
platform will be built.
After the implementation of the models, adjustment and tuning will be performed in order to assess whether
the results meet the requirements specified in WP1. Usually the validation is achieved through calibration
of the model in an iterative comparison of the performance of the model compared to the measurement of
the real parameters and using the differences to improve the model until a good approximation with the real
system is achieved. This process will run iteratively in parallel with the development of models until the model
is acceptable. Data of the vineyard will be acquired on-the-go using the sensors integrated on the system. The
results will be analysed to determine the influence of each of the measured variables obtained after the data
pre-processing, which will provide information on the sensitivity of the final models to the different variables.
Thus, it is intended to obtain various types of results: the quantification of the goodness of the results obtained
from the sensors compared to reference methods and information about the influence of each of the variables
and sensors on the different types of resulting data that is relevant to the creation of maps of the vineyards.
These analyses will provide critical information for a potential reduction of the processing requirements or even
a reduction in costs, since the accurate determination of the most influential variables can simplify the additional
implementation of sensors and hence its cost. To perform the validation, three measures of deviation will be
used: modelling efficiency, coefficient of model determination, and error coefficient.
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 30.00
2 FORC 5.00
3 WALL 5.00
Participant number
10
Participant short name
11
Person-months per participant
4 SIVI 14.70
5 VIBU 2.00
6 GEIS 9.00
7 UPV 28.00
8 AVAN 47.00
Total 140.70
List of deliverables
Description of deliverables
D3.1) Report on sensor selection, adaptation, and field performance: Report on sensor selection, adaptation,
and field performance [month 18]
D3.2) Final report on algorithms, models, calibration curves and validation tests of agronomical and physiological
parameters: Final report on algorithms, models, calibration curves and validation tests of agronomical and
physiological parameters [month 36]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
D3.1 Design of
perception head for
Selection and adaptation of perception
MS4 2 18 robot, describing
sensors for on-the-go performance
individual performance of
each sensor
Objectives
The objectives for WP4 involve processing, transmission, display, and storage of all the data generated through
WP1, WP2, and WP3 for their further implementation in the scouting robot. This task will require many iterations
until the platform can behave satisfactorily.
Description of work
WP4 will be led by UPV and involves SMEs and RTD performers. WP4 has six tasks.
All the information coming from perceptual sensors will be centralized through the main computer onboard,
which will also run the navigation engine in parallel. The information delivered by each sensor will be processed
in real time and transferred to a text file for its final integration in a standardized map. To do so, east-north
coordinates and a common local origin will be used for representing the crop information. This will require the
optimum integration of the GPS receiver in the robot. Crop maps offer visual ready-to-use vital information for
the grower, but further analyses may be necessary on the crop map. For such purpose, a map in an image
format is not operational, but if the data embodied in the map can also be represented in a standard text file
easy to send and easy to read in Matlab or Excel, a more efficient use of information can be made, especially if
prediction models have to be built and later modified when new data is available.
The instructions for data communication and transmission will be reported in D4.1 and D4.2.
The type of data and statistics generated and made available to the user will be established to validate all the
work and verify that it meets the objectives of the project and the specifications of a real system. Once obtained
the most important variables in WP1 and WP3, inductive classifiers (neural networks, fuzzy logic, etc) will be
used for the generation of decision-making rules on the vineyard agronomical and physiological status. The
decision rules generated will be implemented on the onboard computer to make predictions about the vegetative
state of the vineyard.
A graphic user interface (GUI) will assist with the collection of the data. The final crop maps will be either
visually represented in the main software onboard the vehicle (or remotely sent to a mobile unit running the
same softawe) or output in a standardized text file to be represented in general-scope scientific software as
Matlab. Matlab is probably the most standardized platform to process data and plot graphics. M-files have very
low memory requirements and can be sent easily through email. Having so many partners and four countries
The final testing of the system will address two main objectives: to prove that the system is able to obtain data
on-the-go intensively, without problems for a long time and in real vineyard conditions, and to ensure that the
predicted results are true, reliable, and repeatable. Therefore three types of tests will be conducted: reliability
tests, performance tests, and usability tests that will be repeated for different environmental conditions.
Reliability tests will test the robustness of the system in real working conditions and will focus on the acquisition,
processing and delivery of data. The prototype will be run under different weather conditions and possible
system failures will be tested to evaluate the reaction of the multi-sensor system. These include failures in the
sensors, communication failures and the random errors in the components of the system.
During performance testing of the robot it will be checked that the multi-sensor perception system provides
consistent predictions of the vineyard with regard to manual estimations, being capable of repeating the results
when tested under different conditions. The trial will be satisfactory when there are no significant differences
between the previous and the new predictions, establishing certain tolerances.
The ultimate goal is to provide a useful tool to the growers so that the highest importance is given to non-expert
users, which can easily manage the applications developed and display the data obtained by the system in such
a way that they can use them to make their own decisions. Usability tests will be conducted in parallel with the
above and will be carried out by creating specific tests that will be performed by non-expert users. The users will
have a multiple-choice form to provide feedback on the system, adding their impressions and suggestions so
that improvements can start from the beginning of the tests.
After finishing all the validation tests and performing the adequate improvements, a prototype for a simplified
application will be constructed, so that the final maps acquired by the platform will be generated in a compatible
ready-to-use format.
Two demonstrations will be scheduled in months 19 and 36. The mid-project demo will show end-users the
design of the graphic user interface (GUI), allowing them to interact with it in order to gather suggestions and
comments. The final project demo will aim at teaching users to handle the data management system with the
purpose of distributing some experimental units among selected users from Vignerons de Buzet. This demo
will be scheduled within the general final project meeting where potential users will also be invited to attend,
operate, and criticize the prototype for prospective commercialization and upgrades. In addition to EC officers,
this meeting will also be opened to media press approved by the IPR and Dissemination Manager.
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 44.00
2 FORC 9.00
3 WALL 5.00
4 SIVI 8.50
5 VIBU 19.50
6 GEIS 10.50
7 UPV 36.50
8 AVAN 18.50
Total 151.50
List of deliverables
Description of deliverables
D4.1) Report on data fusion of perception and localization sensors in real time: Report on data fusion of
perception and localization sensors in real time [month 24]
D4.2) Multi-parameter map generation and data management: Multi-parameter map generation and data
management [month 32]
D4.3) Application of display of information on mobile devices: Application of display of information on mobile
devices [month 35]
D4.4) User’s manual for prototype: User’s manual for prototype [month 36]
D4.5) Prototype demo and presentation to users: Prototype demo and presentation to users [month 32]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
D4.1 Report. Demo
2 All data captured
with perception head
Data fusion of every component of perception
MS5 7 24 designed in M3.1 and
heal
integrated in M2.2 must
be fused in real time with
GPS.
D4.2 Software. Demos
2-3 All kind of data is
converted to compatible
standard maps, stored
Multi-parameter map generation and data
MS6 7 32 in the robot central
output (including app software)
processor, displayed
in the robot’s control
console, and transferred
wirelessly to remot
Objectives
The objective of WP5 is to disseminate the project results to (technicians and advisors) communities involved
in grapevine management, to growers, wineries, retailers, and consumers. Both traditional and innovative
information technologies will be used for dissemination.
A short video targeted to the possible final users within the wine industry will be produced and available at the
website. One presentation will be held at the beginning of the project and another one at the end. The website
will be the main vehicle for sharing and disseminating information on the VINEROBOT project, starting from
M1. It will have a private section only for partners and a public section where papers and news will be posted
in a timely manner. Demos will be announced beforehand and reports/videos of past demos will be accessible.
After the end of the project, the VINEROBOT website will be operative at least for three years and repository of
open source code will be available. Special care will be paid to the dissemination of the project through social
networks, such as Twitter and/or Facebook.
Dissemination and exploitation plans will be produced their reporting periods will be covered on annual basis
(M12, M24, M36). Results will be disseminated from the very first moment they become available, and the idea
is to attend major annual conferences, even if only preliminary results are obtained. However, the protection
of patentable material may delay the publication of interesting data. Leaflets in the four languages of the
VINEROBOT consortium will be produced and be available for the DEMO activities. In that respect, the IPR &
Dissemination Manager will have to deal with SMEs and find the optimum trade-off. The project will spread its
activities and disseminate the results to potential customers by forming an Industrial Interest Group (IIG) and
through a range of dissemination activities: press release, trade exhibitions, technical conferences, international
academic publications, and face to face meetings. In addition, the project will be open to upcoming opportunities
such as unscheduled demonstrations. Upon request, some conference organizers may allow us run our robot
if it is ready. This is an excellent way to disseminate our results. The IPR & Dissemination Manager will keep
in contact with other FP7 projects working either in the field of robotics or viticulture. Courtesy visits will be
organized for other FP7 teams interested in this project and potential collaborations will be established if deemed
appropriate for both projects.
A two-day international congress will be held at the end of the project. It will contain scientific content but mainly
addressed to the end-users such as wineries, consultants, vineyard managers, and grape growers. Proceedings
Relevant results from the project will be published in peer-reviewed scientific journals of contrasted relevance.
Publication efforts will normally be initiated and guided by the leaders of WPs. The IPR and Dissemination
Manager will assure consistency in all publications, from highly scientific communications to press releases.
General articles on the progress or results of the project may be motivated and led by one of the Operational
Managers. All project publications, including the web site, will contain the following statement: "The work leading
to these results has received funding from the European Union under grant agreement n° 610953". This will
be reinforced and routinely checked by the Management Board. A list of project publications searchable under
various criteria of interest will be maintained as well as a list of papers written outside the consortium that cite the
VINEROBOT project.
All related to journal publications will be reported in D5.2, D5.3 and D5.4
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 3.00
2 FORC 2.00
3 WALL 1.00
4 SIVI 1.00
5 VIBU 1.00
6 GEIS 1.00
7 UPV 3.00
8 AVAN 8.00
Total 20.00
List of deliverables
Description of deliverables
D5.1) Project website and presentation video: Project website and presentation video [month 1]
D5.2) First report of the exploitation and dissemination of knowledge activity: First report of the exploitation and
dissemination of knowledge activity [month 12]
D5.3) Second report of the exploitation and dissemination of knowledge activity: Second report of the
exploitation and dissemination of knowledge activity [month 24]
D5.4) Third report of the exploitation and dissemination of knowledge activity: . Third report of the exploitation
and dissemination of knowledge activity [month 36]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
D5.1 Website. Design
of professional website
MS7 Project website and presentation video 8 1
with intranet for internal
management of partners
D5.2 Report.
Manuscripts and
Publication manuscripts and drafts as well
drafts. Publications in
MS8 as presentations in conferences and press 8 12
journals, presentations
releases of the
in conferences, press
releases, etc.
D5.3 Report.
Manuscripts and
Publication manuscripts and drafts as well
drafts. Publications in
MS9 as presentations in conferences and press 8 24
journals, presentations
releases of the
in conferences, press
releases, etc.
D5.4 Report.
Manuscripts and
Publication manuscripts and drafts as well
drafts. Publications in
MS10 as presentations in conferences and press 8 36
journals, presentations
releases of the
in conferences, press
releases, etc.
Objectives
1) responding to the complexity of the project and the degree of integration required
2) guaranteeing consistency between the project work plan and the actual implementation
3) achieving the goals of the project in a cost-effective way according to the agreed time schedule and budget
4) ensuring an effective process of decision-making and internal communication
5) Managing knowledge circulation and intellectual property according to the Consortium Agreement.
Description of work
WP6 will be led by the project coordinator (UDLR). The management structure of the project is designed in such
a way to represent adequately the SMEs needs and their collective interest.
UDLR will be in constant connection with the leaders of the different WPs by e-mail and phone. Meetings will
be organised in order to plan the activities and to assess the results progress. A planning of meetings of the
management bodies (Steering Committee, Project Management Team and Advisory Board) and of technical
and financial reports required by the EC will be prepared. In principle, three meetings will be organised for each
management body; additional meetings will be done for specific needs arising during the project. All exchanges
of information among partners will be copied to UDLR to guarantee its update and give the possibility to check if
all aspects agreed in the Consortium Agreement and in the technical annex (with the EC) are pursued.
Reports will be prepared in due time by each partner, send to the coordinator and included in the annual report.
UDLR will guarantee a proper exchange among WPs, in the respect of intellectual property and according to the
level of dissemination agreed in the Consortium Agreement. Intermediate reports covering the managing and
technical progress of the VINEROBOT project will be reported in the Progress reports on M6, M18 and M30, to
be delivered to the PO. Mandatory technical and financial reports will be also delivered on M12, M24 and M36.
Partners will finalize the Consortium Agreement and manage all the matters related to IPR of the Consortium as
a whole and as a single entity for exploitation of the results.
Participant number
10
Participant short name
11
Person-months per participant
1 UDLR 21.00
2 FORC 0.50
3 WALL 0.50
4 SIVI 0.30
5 VIBU 0.50
6 GEIS 0.50
7 UPV 4.00
8 AVAN 0.50
Total 27.80
List of deliverables
Description of deliverables
D6.1) E-mail and data exchange network between coordinator and partners: E-mail and data exchange network
between coordinator and partners [month 1]
Lead Delivery
Milestone benefi-
Milestone name date from Comments
number
59
ciary
Annex I
60
number
Agreement.Kick-off
MS11 Signature of Consortium and kick-off meeting 1 1 meeting in Universidad
de La Rioja.
D 6.3 The AB has been
MS12 First evaluation of the Advisory Board 1 12 constituted and provides
feedback for first year.
D6.2 Location Burgundy
MS13 First project review meeting 1 12
(France)
Location Valencia
MS14 Second project review meeting 1 24
(Spain) or Siena (Italy)
Hosted by UDLR in
MS15 Final project review meeting 1 36
Logroño (Spain).
presentations in
in conferences, press
conferences and
releases, etc.
press releases of the
Publication
D5.4 Report. Manuscripts
manuscripts and
and drafts. Publications
drafts as well as
MS10 WP5 8 36 in journals, presentations
presentations in
in conferences, press
conferences and
releases, etc.
press releases of the
Signature of Agreement.Kick-off
MS11 Consortium and WP6 1 1 meeting in Universidad de
kick-off meeting La Rioja.
D 6.3 The AB has been
First evaluation of
MS12 WP6 1 12 constituted and provides
the Advisory Board
feedback for first year.
First project review D6.2 Location Burgundy
MS13 WP6 1 12
meeting (France)
Second project Location Valencia (Spain)
MS14 WP6 1 24
review meeting or Siena (Italy)
Final project review Hosted by UDLR in
MS15 WP6 1 36
meeting Logroño (Spain).
1. RTD/Innovation activities
WP 1 1.00 1.00 1.00 1.00 1.00 2.00 1.00 1.00 9.00
WP 2 13.00 5.00 55.00 18.00 1.00 8.00 37.50 46.00 183.50
WP 3 30.00 5.00 5.00 14.70 2.00 9.00 28.00 47.00 140.70
WP 4 44.00 9.00 5.00 8.50 19.50 10.50 36.50 18.50 151.50
WP 5 3.00 2.00 1.00 1.00 1.00 1.00 3.00 8.00 20.00
Total Research 91.00 22.00 67.00 43.20 24.50 30.50 106.00 120.50 504.70
2. Demonstration activities
Total Demo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
4. Other activities
Total other 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Total 112.00 22.50 67.50 43.50 25.00 31.00 110.00 121.00 532.50
Table of Contents
B1. Concept and objectives, progress beyond state-of-the-art, S/T methodology and work plan
................................................................................................................................................ 1
B 1.1 Concept and project objective(s)................................................................................. 1
B 1.2 Progress beyond the state of the art ........................................................................... 5
B 1.3 S/T Methodology and associated work plan.............................................................. 11
B 1.3.1 Overall strategy and general description .......................................................... 11
B 1.3.2 Timing of work packages and their components............................................... 19
B2. Implementation............................................................................................................... 25
B 2.1 Management structure and procedures .................................................................... 25
B 2.1.1 Description of project management structure and procedures ......................... 25
B 2.2 Beneficiaries ............................................................................................................. 30
2.2.1 Brief descriptions of the role of each partner in the project................................. 30
B 2.3 Consortium as a whole.............................................................................................. 40
B 2.4 Resources to be committed ...................................................................................... 53
B3. Impact............................................................................................................................. 56
B 3.1 Strategic impact........................................................................................................ 57
B 3.2 Plan for the use and dissemination of foreground..................................................... 64
B1. Concept and objectives, progress beyond state-of-the-art, S/T methodology and work plan
Introduction
VINEROBOT: VINEyard ROBOT
State of the art and motivation
Traditionally, European vineyards have not been technology driven. The human eye has been
the only “sensor” steering vineyard management decisions, providing subjective evaluations on
yield, vegetative growth and plant status regarding water stress, pests and diseases. The wine
industry has always sought the accurate and objective estimation of grape yield. Similarly, the
assessment of a vineyard’s wine grape quality potential has been regularly attempted by using
vineyard score sheets, requiring the visual subjective approach to several canopy variables such
as vigour, leaf status, exposed leaf area, canopy porosity and fruit exposure; all of which are
intrinsically related to grape and wine composition and quality.
Sustainable viticulture requires objective and continuous key parameter monitoring for rational
decision-making with advanced technologies and sensors in the fields. These would allow for
observing crops and quantifying important factors in vineyards such as grape yield, leaf
development, disease incidence and detection of various stress issues (water, nutrition, etc.).
Additionally, all this information can be globally referenced with GPS tags, enabling the
assessment of spatial variability for key agronomical parameters in the vineyard which may
facilitate precision viticulture strategies, aimed at differentiated and optimized vineyard
management.
In agriculture, optical technology has been mainly used in post-harvest operations but, in
viticulture, the use of computer vision to characterize aspects of the canopy and estimate
parameters affecting the production or health of grapes is virtually unknown. As a fact, there is
no practical system that integrates multiple sensors capable of acquiring information on
agronomical, physiological and fruit composition simultaneously and on-the-go. Only some
remote sensing solutions from aerial platforms or satellites have included the simultaneous
acquisition of spectral information in the visible and infrared ranges allowing for the assessment
of grapevine vigour and water status. However, the small spatial resolution of multispectral
devices together with the discrete architecture of grapevine cultivation in rows rather than
continuous crop, the limited weather flexibility and the elevated cost of aerial monitoring are
important drawbacks, which have practically led to discard the remote sensing technique for
most of the small- and medium-size European vineyards. An alternative to long-distance remote
sensing from airplanes and satellites has been the use of UAVs but these present serious
limitations: payload; severe restrictions in terms of flight autonomy, flight control and safety;
lack of official regulations for these aerial vehicles, etc.
The use of UGV to monitor vineyard physiological parameters and grape composition has many
advantages over simple hand-held devices for manual sampling, and over unmanned aerial
vehicles (UAV). UGV may provide key information regarding vineyard physiology and water
stress, grape composition, and the incidence of pests and diseases much faster than manual
solutions, and at higher resolution, in a more flexible way, and at lower costs than aerial
scouting carried out by drones or planes. The scope of the present project covers the integral
monitoring of vineyards over the entire season by placing a ground robot along the vineyards.
Robot behavior will be determined by artificial intelligence and machine learning techniques to
plan how to operate in a changing environment in order to construct real-time precision
viticulture maps of important production factors acquired with non-invasive sensors at high rate
and high resolution.
FORC
SENSORS AVAN
VIBU
UDLR
UPV LABORATORY
GEIS
WIFI
VINEYARD SIVI 3G
UDLR
AVAN UPV UPV
GEIS
WALL
VINEYARD
ROBOTS
The principal objective of this project is the design and construction of a medium-size scouting
robot capable of navigating autonomously along vineyard rows to gather key information of
crops, which will be output through standardized maps to be used by vineyard managers or
end-users. The successful deployment of a robot like this requires the effective solution of each
of the following challenges and specific objectives:
1. Physical constraints for mobility in off-road environments. The robot will use a four-
wheel-drive system to rove, and will provide support and protection for the rest of
equipment onboard. WALL-YE will lead WP2 and assemble the enhanced robot over
which the prototype will be constructed. The Technical Manager (UPV) will visit Wall-Ye
facilities to assist in the early design of the field robot. The following milestones are
devoted to accomplish this objective: MS1, MS2
2. Navigation planning and safeguarding. The platform will be endowed with a path
planner based on GPS positioning and real-time steering control assisted by perception
sensors granting obstacle detection and avoidance. WALL-YE and the UPV (Technical
Manager and leader of WP4) will work together to design a stable control system for
the robot’s mobility and integrate the navigation sensors in the platform. The following
milestones are devoted to accomplish this objective: MS2, MS3
3. Principal perception engine. Although the robot will be equipped with reactive
perception sensors for safeguarding and obstacle detection, the principal perception
engine of the robot is the set of advanced sensors in charge of capturing the key
information from the crop. They are listed in Table 1.1a and will be selected, calibrated,
implemented and validated by FORC (leader of WP3) in a joint effort with the three
RTDs participating in the project: UDLR (WP6 leader), GEIS (WP1 leader) and UPV
(WP4 leader). The following milestones are devoted to accomplish this objective: MS1,
MS3, MS4, MS5
4. Global localization system. The scouting robot will use a differential GPS for navigation
and mapping. The leader team will be UPV, which will implement the conditioning
routines (Rovira-Más and Banerjee, 2013) and a grid-based approach (Rovira-Más,
2012) developed in previous research projects. The coordinate system in which the
robot will navigate will be the local tangent plane (LTD) with Cartesian coordinates
north, east and height. SIVIS will assist with the integration of the GPS receiver in the
robot. The following milestones are involved in the accomplishment of this objective:
MS2, MS6
5. Mapping engine. An ad hoc algorithm implemented in the main processor will combine
the perceptual data coming from the perception engine with the GPS information to
yield global-referenced crop maps. This is a delicate step because the maps must be
generated in real-time while keeping a common local origin and global axes; therefore
all teams will be involved including the winery VIBU, which will be the end-user. The
following milestones are devoted to accomplish this objective: MS1, MS2, MS5, MS6
6. Communication engine. Optimized packages of information will be sent wireless upon
user demand. Potential reception units will be tablet computers and smartphones. The
leader developing this electronic system will be the enterprise SIVIS. The following
milestone is involved with this objective: MS1, MS5, MS6
7. Validation tests. The nature of this project demands many iterations and heavy testing
in actual fields under real conditions. Validation tests need the constant checking of
ground-truth verification. The RTDs GEIS (WP1 leader) and FORC (leader of WP3) will
supply the necessary equipment and expertise for the validation of agronomical
properties. Field tests will take place in the vineyards of VIBU. The following milestones
are devoted to accomplish this objective: MS2, MS4, MS5
8. Prototype construction and demo agenda. All the systems and technology developed
will be integrated in a self-propelled prototype, which will be tested and demonstrated
according to an established agenda involving all the EU states participating in the
project. The following milestones are devoted to accomplish this objective: MS2,MS3,
MS4, MS5
The main capability expected from the VINEROBOT is the possibility of working autonomously
and safely over long periods of time under uncertain environmental conditions of vineyards. At
present there are no commercial robots that can be left alone in fields for a timeframe longer
than demo duration. In addition to navigation, the VINEROBOT will gather key information from
the vines and will convey it through compatible maps. The variety of maps will depend on the
producers’ needs and technology developed by the project partners, but the crucial goal will be
the reliability of the robot to navigate on its own without provoking any risk or danger in the
field to plants, vehicles, structures and principally field workers. The autonomy of the robot will
depend on the selection of optimum hardware (navigation skills and safeguarding excellence) as
well as a good decision-making design based on artificial intelligence algorithms. The major goal
will be twofold: the robot must navigate in any standard vineyard for at least four hours without
human intervention, and at the end of the mission the robot must provide a set of crop maps in
a compatible format. This will involve a lot of iterations and the mobility of researchers to work
together for intensive periods of time. Several scheduled demos will provide the necessary
feedback to meet the main objectives. The numerous iterations will necessarily result in various
degrees of evolution. The first platforms will basically perform navigation and safeguarding
tasks. The final prototype is expected to evolve from a four-wheel medium-size unmanned
ground vehicle to one endowed with different perception sensors such as digital cameras and
laser rangefinders, a global localization engine and a central processing unit capable of
assembling, storing and optionally sending (wirelessly) the resulting crop maps. The interaction
with the various development platforms will be through its main computer. The next generation
of the scouting robot will incorporate a touch-screen or LCD messaging display to manipulate
the robot from the outside (front panel) and retrieve the maps without accessing to the inner
computer hosting the main algorithms and control of the sensors. It will have safety emergency
buttons installed in the external chassis to halt its operations should an emergency arise.
and continuous monitoring of key parameters for rational decision making, which will only
become a useful tool by applying new technologies and advanced sensors in the vineyard.
Zerger et al. (2010) and Lee et al. (2010) have illustrated the enormous potential of new
technologies to monitor crops and quantify important parameters in agriculture such as grape
yield, leaf development, disease incidence and the objective detection of various stress factors
(water, nutrition. etc.). Furthermore, all this information can be globally referenced with GPS
tags enabling the assessment of spatial variability for key agronomical parameters of the
vineyard.
Proximal sensing includes all detecting technologies that retrieve information from an object
when the distance between the sensor and the object is less than, or comparable to, some of
the dimensions of the sensor. In agriculture, these sensors have been widely used, either
individually or in combination, for evaluating and characterizing various soil properties such as
water content, salinity, nutrients, etc. (Zerger et al., 2010). However, the characterization of
biophysical and physiological parameters of plants by these sensors has been less explored. The
main sensors used for monitoring the cultivation and / or composition of fruits in the field are
the following: RGB cameras, thermal cameras, multi and hyperspectral cameras, fluorescence-
based sensors, LIDAR (light detection and ranging) sensors and NIR spectroscopy. An advanced
and non-invasive technology to detect visible and invisible properties of objects is represented
by “machine vision.” In the field of agriculture, this technology has been mainly applied in post-
harvest for quality inspection and colour detection to determine the incidence of defects in
fruits such as citrus (Blasco et al., 2007; Blasco et al., 2009a; López-García et al., 2010), apple
(Xiao-bo et al., 2010) and pomegranate (Blasco et al., 2009b), among others. In viticulture, the
use of computer vision to characterize aspects of the canopy and estimate factors affecting the
production or health of the grapes is virtually unexplored. Some authors have begun using
methods based on the analysis of RGB images with classification algorithms for the detection
and quantification of various parameters of interest such as the exposed leaf surface, the
porosity of the canopy, the number of clusters exposed and the proportion of damaged or
senescent leaves (Diago et al. 2012). Other approaches have used near infrared (NIR) filters to
segment healthy vegetation and compose vigour maps of vineyards with the aid of a differential
GPS (Sáiz-Rubio and Rovira-Más, 2012). Another type of advanced detection sensors with a wide
range of applications in agriculture are the thermal cameras. Also called IR cameras, they
measure the radiation emitted by an object, which is proportional to its temperature, allowing
the registration of the temperature distribution in the field or in individual plants. One of the
main applications of thermography in viticulture is assessing the grapevine water status needed
in irrigation scheduling (Sela et al. 2007). The versatility of thermography and its discriminatory
power, based on the emission of radiation from objects, has led this technique to complement
conventional machine vision applications (Bulanon et al., 2009).
The fluorescence emitted by various compounds present in leaves and fruits (mainly
chlorophyll) has been used to study the physiological behaviour of plants (Chaerle et al., 2009),
and recently has been proposed to monitor different crops and estimate the composition of
fruits in a non-invasive way (Agati et al., 2007). When combined with thermography,
fluorescence has been used for monitoring and selecting plants tolerant to different stress
factors (Chaerle et al., 2009). The Multiplex™ device is a multiparametric portable sensor based
on fluorescence technology that was developed in France by Force-A (www.force-a.eu) and Dr.
Cerovic’s group. This sensor is commercially available and has been successfully used to
determine, quickly and non-invasively, anthocyanin and phenolic compound contents of grapes
(Cerovic et al., 2008; Ben Ghozlen et al., 2010). In addition, the Multiplex™ sensor has also been
used to study the spatial variability of the phenolic composition of the grapes in a Tempranillo
(Vitis vinifera L.) vineyard (Baluja et al., 2011). Coupled with a differential GPS, this sensor is able
to built maps, by hand or on-the-go, of different types of information such as canopy porosity,
nitrogen status, chlorophyll content or grape quality. Spectral sensors are optical devices
capable of measuring specific spectral indices, which are related to certain physiological
features. More complex hyperspectral cameras have been used to provide spectral information
over a range of wavelengths at each single pixel of a 2D image. These hyperspectral cameras
have been applied for determining the content of various compounds in fruits (water, sugars,
acids and polyphenols), but its use has mostly been conducted indoors (Nicolai et al. 2007).
Technical limitations of existing systems. There is no practical system at present that integrates
multiple sensors capable of acquiring information on agronomical, physiological, and fruit
composition simultaneously and on-the-go. Only some remote sensing solutions include the
simultaneous acquisition of spectral information in the visible and infrared ranges allowing for
the assessment of grapevine vigour and water status. In viticulture, remote sensing has been
applied through aerial platforms or satellites (Johnson, 2003; Lamb et al., 2004; Sepulcre-Canto
et al., 2009). Likewise, most of the commercial attempts to gather agronomical and
physiological data on vast vineyard plots rely on remote sensing, either from aerial such as ITK
(France) and Ager (Italy), or satellite-based platforms like Vineview in California (USA) and
Deimos (Spain). However, the small spatial resolution of multispectral devices together with the
discrete architecture of grapevine cultivation in rows rather than continuous crop, the limited
temporal flexibility (cloudy and foggy weather conditions are not suitable), and the elevated
cost of aerial monitoring are important drawbacks, which practically lead to discard this
technique for the actual situation of European vineyards, most of them small- and medium-size
producers. An alternative to long-distance remote sensing from airplanes and satellites could be
the use of unmanned aerial vehicles (UAV or drones) but these systems present other serious
limitations: i) their payload (maximum weight that can be transported) is still too low to support
all the sensors needed for a complete monitoring of the crops; ii) there are severe restrictions in
terms of flight autonomy, flight control, safety; and iii) there is an uncomfortable lack of official
regulations for this kind of aerial vehicles, which may become very dangerous if improperly
used.
At field level, there is a wide range of sensors capable of determining spectral indexes related to
vigour (GreenSeeker™, PlantPen NDVI 300™) and water stress (PlantPen PRI 200™) and sensors
based on chlorophyll fluorescence (DualexTM and MultiplexTM), mainly used to estimate vine
vigour and berry grape composition. In the same way, multi-parameter sensors based on
portable near infrared technology have emerged to estimate grape composition and retrieve
plant physiology information (MicroPhazirTM). Additionally, there are other sensors mainly used
for research purposes such as gas-exchange meters, thermal imagers, sap-flow meters,
dendrometers and Scholander pressure chambers. Nevertheless, the main pitfall of these
sensors and monitoring devices is that they are slow, too high labour-demanding (as they are
not automated), destructive in many cases, and most of the times only capable of measuring a
small number of vines per hectare in a reasonable time period making impossible the real-time
generation of crop maps and the application of precision agriculture tools.
In conclusion, a promising and solid alternative to remote sensing for vineyard monitoring
seems to rely on proximal sensing, through which all the disadvantages mentioned above are
certainly expected to be overcome by the multisensor scouting robot, as it includes advanced
image analyses involving machine vision, infrared and fluorescence acquisition data, which will
provide objective and accurate information on agronomical, physiological and grape
composition aspects of vast vineyards at any time, without timing or weather constraints.
An innovative approach for modernizing vineyard management in Europe. The present project
proposes the integration of three fundamental sensing technologies—chlorophyll-based
fluorescence, RGB machine vision, and IR thermography—to map important agronomical and
physiological parameters of vineyards, non-invasively and in real time, from a novel scouting
robot capable of covering large fields autonomously and safely. The end-user will improve the
vineyard’s management by means of standardized multi-parameter crop maps, which will be
downloaded from the robot upon the end of the mapping mission, or alternatively transferred
wirelessly from the robot to smartphones or tablet computers. The use of complementary
detection sources (RGB vision, infrared thermography, and fluorescence) to expand the
capabilities of monitoring systems in terms of discrimination accuracy and diagnostics potential
is a strategy that has been mentioned by several authors (Chaerle et al., 2009; Cubero et al.,
2010a; Lee et al., 2010; Zerger et al., 2010). Additionally, the continuous and globally-
referenced measurement of key parameters (on-the-go mapping) from an innovative farm
robot especially designed to navigate along vineyard rows will add valuable insights to the
management of wineries and grape producers.
The proposed solution faces one of the principal problems found by the wine production
industry: inefficient management as a consequence of data-less decision-making. The
philosophy envisioned to improve the management of vineyards is the implementation of
precision farming (PF) techniques in such a way that spatial variability—a phenomenon of
extreme importance to produce premier wine—is taken into account in the wine-making
process. The careful selection of perception sensors will try to gather as much optic information
as possible from the crop. The combination of the vegetation sensing unit with the GPS receiver
will enable the creation of global crop-management maps, which will be comparable among
them and with maps generated in other seasons.
The final design of the robot power system will be the result of various iterations and numberless
field tests. Nevertheless, our initial idea is to incorporate two easily-rechargeable lithium
batteries: the primary battery for mobility (transmission and steering) and the secondary battery
for sensing, computing, and displaying. Some of the reasons for this choice are:
1. Navigation in off-road environments requires a significant amount of energy, as friction
is important for uneven terrains. The scenario worsens in sloping terrains where the
weight of the vehicle must be overcome.
2. If wheel slip is excessive, or the required torque exhausts the batteries, we still want to
run the communication system to send a message to the closest operator warning of
the vehicle’s incapacity to move. Warning lights should also work to warn approaching
vehicles in the dark.
3. The GPS receiver should be isolated from potential fluctuations in power caused by
navigation needs. If GPS power drops, the position lock will be lost and the system will
need several minutes to find the minimum number of satellites before resuming its
mission. Even if the robot stops its motion, the GPS must always be powered to keep
the number of satellites locked.
4. Solar panels may help power the system for emergency warning lights or keeping the
robot display on, as an additional resource.
5. Both batteries may be connected through a smart system that allows the transmission
system be powered by the secondary battery only if a hazard is detected, or the main
processor and communication system be powered by the primary battery to save all
data and send an SOS message before running out. The following conceptual schematic
shows this idea:
Annex 1 – Description of Work – Part B Page 8 of 70
Version N°3 agreed with the EC services on 28-06-2013
FP7-ICT-2013-10-610953
All the issues associated to remote sensing and aerial monitoring such as low resolution, low
update rate, high cost and little controllability and accessibility will be surmounted by the
deployment of a ground robot, big enough to carry the necessary payload, but light enough to
be safe even at the unlikely occurrence of a severe failure in the navigation system. The crop
maps constructed by the robot will be updatable in real-time with the information extracted
instantaneously by the perception and localization engines of the robot. The number and nature
of the final maps will be determined by the research conducted under WP1 and WP3, but the
scope of the perception engine of the robot will include the following key parameters for a
successful vineyard management:
Vine growth
Grape yield estimation
Vineyard water status
Grape composition
road applications, including planetary exploration, may benefit from the algorithms and control
laws devised in the project.
The concept of precision farming (PF), including precision viticulture, has been around for more
than two decades and is regarded as the only sustainable solution for agriculture in
industrialized counties. However, the adoption rate of PF has globally slowed down in recent
years (Pölling et al., 2010), mainly due to the complexity of solutions when taken to the field.
Part of this difficulty rests on the lack of sensors for automatic acquisition of data, as crops need
to be properly monitored, and manual sampling is prohibitive at the elevated cost of labor in
Europe, Japan and North America. This project aims at the core of this problem; as the scouting
robot will be endowed with several non-invasive techniques (WP1 & WP3) to assess crop status.
Furthermore, not only we expect to contribute in sensing but also in data handling (WP4),
another vital issue in the standardization of solutions based on precision agriculture
Performance research indicators for the project. The Consortium defines the following
indicators to evaluate the project’s results and impact.
Results
The success of the research product (i.e., the VINEROBOT) is measured by using concrete
indicators, in particular: i) indicators defined in D1.2 Benchmarking protocol, which defines the
rate of achievement protocol for assessing VINEROBOT performance (Table B1.1); ii) usability of
the UGR by end-users; iii) reliability and precision in decision-making; iv) vineyard management
costs.
Table B1.1. Rate-of-achievement protocol for assessing VINEROBOT performance (D1.2)
ROBOT PERFORMANCE
Specification
MODEST AVERAGE EXCELLENT
Mapping pace along rows (km/h) 1-2 3 4-5
Mapping autonomy (hours) 1–2 2-3 4-6
Presence of data-less gaps (%) >30 15 - 30 < 15
Human interventions per 4-hour period >3 1-3 None
Average error predicting yield (%) To be defined To be defined To be defined
Average error estimating grapevine leaf area (%) To be defined To be defined To be defined
Average error estimating canopy temperature (⁰C) To be defined To be defined To be defined
Average error estimating leaf Nitrogen (%) To be defined To be defined To be defined
Average error estimating grape anthocyanins (%) To be defined To be defined To be defined
Average error estimating grape phenols (%) To be defined To be defined To be defined
Usability will be assessed in WP5 (Task. 5.2. Exploitation plans and dissemination). In this task
the creation of an industrial interest group (IIG) involving potential end-users is envisaged. The
IIG will be invited to attend seminars, technical conferences and demonstration activities of the
project. In the context of these activities, participants will be asked to fill-in a questionnaire
where specific questions regarding the usability of the VINEROBOT in comparison to actual
systems for vineyard monitoring will be included. To answer these questions, participants will
have to score among 5 possible options (<25%, <50%, 0%, >25%, >50%), which represent lower,
equal or higher usability than actual systems. The expected results will be an increase of
usability > 25%. The members of VIBU will provide an excellent group of growers to which the
questionnaire will be addressed.
The reliability and precision of the robot operations will be evaluated in both WP2 and WP4
(Task 2.6. Progress field demonstrations; Task 4.6. Prototype demo and presentation). The
potential end-users attending the prototype demos will evaluate the performance of the robot
when making decisions. The trajectory of the vehicle will be tracked with a GPS receiver, and
safety will be assessed by navigating both in straight paths and over the headland turns in the
presence of static and dynamic obstacles (people moving around the field and interacting with
the robot’s path). The accuracy of maps will be analyzed in the office by comparing
automatically-generated maps with manual estimations.
Vineyard management costs will be evaluated in WP4 (Task 4.5. Deployment of prototype; Task
4.6. Prototype demo and presentation). The end-users chosen from VIBU to evaluate the system
will be invited to use the robot with the purpose of applying precision viticulture techniques in
their plot fields. A comparative study will be carried out to estimate the benefits of making
decision based on the new data source. The expected outcome is a cost reduction of at least 30
%. Similarly, yield and quality of grape production will be evaluated in WP4 (Task 4.6. Prototype
demo and presentation). An increase in the quantity and quality of grape production derived
from the close monitoring of the vineyard and the implementation of vineyard delineation
zones traced by the VINEROBOT is expected, with improvements in the range 5-25%.
Impact
Since this research project aims at designing and manufacturing a novel use-case unmanned
ground robot (UGR) to be marketed, the impact of RTD activities is evaluated on the basis of the
quantitative indicator “return on investment.” Section B3.1 explores this aspect in greater
detail. In particular, a hypothesis relative to the financial data of the first five operative years
following the end of the project is drawn with the purpose of estimating sales and profits
resulting from the investment in RTD. Other indicators such as patents, bibliometrics and peer
reviews are also discussed in section B3.2.
will be done in full agreement with the coordinator, the steering committee and the project
management board. Mobility among different partners will be granted whenever the
development of the activities advises to do so.
The risks to fail expected results can be considered low. However, whenever the results
obtained with the planned approaches are not completely satisfactory, alternative approaches
will be followed and maximum efforts will be applied to fulfil the objectives of this proposal,
following suggestions from the steering committee, the project management board and the
advisory board (see section B.2.1) and the contingency plan reported in Table B.1.2. In
particular, the risk that the system will not reach the main goals of autonomous navigation and
grapevine monitoring of the vineyard providing information on vegetative growth, grape yield,
physiological status and grape composition is very low. On one hand, the potential processes to
gather crop information from RGB/NIR imaging, thermography, and fluorescence have been
already investigated and developed by some partners separately. On the other hand, the
achievement of autonomous and safe navigation of the UGV will depend on the selection of
optimum hardware (navigation skills and safeguarding excellence) as well as a good decision-
making design based on artificial intelligence algorithms, topics that fall in the field of expertise
of several partners of the consortium. However, even if the challenging target of 4-5 hours of
autonomous and safe navigation involving efficient monitoring and mapping is not finally
reached, self-navigation for periods of time longer than those achieved by current off-road
robots will be a very important outcome for viticultural monitoring and optimization of key
viticultural practices (in agreement with the European Directives about the Sustainable Use of
Water and Pesticides), as there are no commercial solutions even close to meet these
objectives, and the vine robot will definitely set the starting point for improved prototypes.
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
WP1 Technical and agronomical requirements
Task 1.1 Agronomical and physiological Failure gathering 5% Critical: lack of Efforts to gather the relevant information on GEIS
requirements information on the relevant the agronomical and physiological
agronomical and information for requirements within the following month
physiological requirements later progress of form the delivery date.
in the required timeline the project
Task 1.2 Multi-sensor robot Failure gathering 5% Critical: lack of Efforts to gather the relevant information GEIS
requirements information on the multi- relevant on the multi sensor robot requirements
sensor robot requirements information for within the following month form the
in the required timeline later progress of delivery date.
the project
Task 1.3 Sensor Fusion and modelling Failure gathering 5% Critical: lack of Efforts to gather the relevant information on GEIS
requirements information on the fusion relevant the fusion and modelling requirements
and modelling requirements information for within the following month form the
in the required timeline later progress of delivery date.
the project
Task 1.4 Data management and display Failure gathering 5% Critical: lack of Efforts to gather the relevant information on GEIS
requirements information on the data relevant data management and display requirements
management and display information for within the following month form the
requirements in the later progress of delivery date.
required timeline the project
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
WP2 Multi-sensor scouting robot
Task 2.1 Task 2.1. Design of the robot Failure of integration of 20% Increase of the Find an off-the-shelf platform with WALL
navigation sensors and difficulty to get an navigation capabilities.
control systems. integrated,
compact multi
sensor system
Task 2.2 Task 2.2. Design of sensory The sensor integration may 10% Increase of the Assembling of an external driver or memory WALL
system for the detection of not be achievable in the timing for data module to the multi sensor robot for data
canopy and fruit under timeline of the project acquisition and storage.
dynamic conditions transmission
Failure in the 15% Failure in data -To adjust the acquisition speed and settings WALL
synchronisation of the data homogeneity for all sensors
acquisition with all the -To implement synchronisation strategies
sensors and the advance of (external encoder, trigger, etc)
the robot
Task 2.3 Task 2.3. Design of sensory Failure of real-time data 5% Critical: complete Slowing down of the robot at a first stage. WALL
system for physiological and acquisition due to high lack of data for Secondly, modification of the acquisition
agronomical parameters of speed of the robot later processing settings.
vineyards under dynamic Failure of real-time pre- 15% Increase of the Installation of additional processing units to WALL
conditions processing time needed for accelerate the processing operation
data processing
Task 2.4 Task 2.4 In-field real-time data The types of sensors and 25% Will delay time -Substitution of devices and/other WALL
acquisition associated components may associated components for other of
not be immediately improved performance in the required, but
matching applications insufficient capability.
requirements -Additional tests may be performed in
vineyards in partners’ countries once
device(s)/component(s) are substituted.
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
Task 2.5 Design and construction of The robot does not meet 15% Delay in the Intensify the following 1-2 months in getting WALL
optimized robotic platform commercial specifications. progress of the the commercial needs and appearance of
project the robot
Task 2.6 Progress field demonstrations No people attend the 10% No dissemination is Organize an additional date for another VIBU/AVAN
demonstrations, or the attained or bad demonstration and increase the advertising
robot fails in its missions. reputation is of the event. It may also be necessary to
spread facilitate transportation to the vineyard.
Heavy preparation and testing.
WP3 Fusion and modelling
Task 3.1 Data analysis The calculation 20% Decrease of the - Investigate to improve the numerical UPV/UDLR
performance is not system usability for algorithm optimization
satisfactory on-demand services - Adopt parallel computing strategies to
speed up the analysis
- Find out for an improved data processing
configuration
- Rewriting of the processing procedure
Task 3.2 Building of algorithms and Failure of real-time data 15% Reduce the speed of -Optimise the algorithms for not providing UPV/FORC
models processing due to high the on-the-go not essential data
speed of the robot or lack of system robot -Use of faster processing units
processing speed of the Decrease of the -Parallelise the algorithms on several
computers used global performance processing units
Task 3.3 Data integration and fusion Failure in combining 10% Lack of robustness -Revise the information provided from each UPV/FORC
information from the of the decision sensor
different sensors models -Revise the acquisition settings and
Failure of sensor Lack of accuracy of procedure for all sensors
compatibility the system -Revise the local pre-processing process to
obtain proper data
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
Task 3.4 Establishment of agricultural Risk of insufficient accuracy 10% Calibration curves - Use data already available and validated UDLR
calibration curves of the calibration curves will lack of acquired with hand-held devices similar to
robustness those mounted on the robot under
controlled conditions
Task 3.5 Models’ validation Risk of insufficient accuracy 10% Models will lack of - Use data already available and validated UDLR/FORC
of the models robustness acquired with hand-held devices similar to
those mounted on the robot.
Damage to experimental 5% Models will lack of -Choose of a new location for testing or UDLR/FORC
vineyards (hail, heavy storm, in-field validation manual and on-the-go data collection
large fungal infection...)
WP4 Sensory data management & display
Task 4.1 Data transmission and storage Failure of data storage 10% Critical: Lack of data -Revise the storage unit UPV
for further -Should the storage unit fail, substitute it for
processing another one
Lack of GPRS coverage to 15% Failure of the - Temporary storage of data already SIVIS
send data to the remote transmission process implemented (data buffering)
server/cloud and lack of data for -Installation of a data concentrator with
further processing manual and local data connection through a PC
Task 4.2 Implementation of fusion The variables used to build 5% Creation of erroneous To revise the data acquisition process, to UPV
models the models cannot predict models, impossibility search for indexes, ratios or combination of
the real-world of create proper variables. In extreme cases, to search for
models new variables
Decisions taken by the 10% Lack of the accuracy To check out the training data, re-build the UPV
system do not conform the in the decisions models or use different statistical
reality techniques
Information provided by the 5% Critical: loss of To revise the requirements of the end users UPV
models is not relevant for interest of end to adapt the system properly to satisfy the
the users users end user demands
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
Task 4.3 Display of agricultural The user interface is not 5% Increase the Evolve the user interface agree to the UPV/SIVIS
information appreciated from a difficulty for the external user indications
significant number of users product marketing
outside the project and selling after the
project completion
Task 4.4 Complete system validations Individual 20% System will lack For those device(s) and/or model(s) not UPV/SIVIS
device(s)/model(s) do not some robustness completely validated use data already
fully work with the desired available and validated in previous research
accuracy works with similar, but hand-held devices
under controlled conditions.
Validation of the complete 10% Delay to market Use data already available and validated in UPV/SIVIS
system is not fulfilled on previous research works with similar, but
time hand-held devices under controlled
conditions.
The prototype is not 10% Delay to demo and Efforts in having the prototype delivered UPV
achieved on time market within the next month. Intensive effort and
Task 4.5 Deployment of prototype team mobility.
No people attend the 10% No dissemination is Organize an additional date for another VIBU/AVAN
demonstrations, or the attained, or bad demonstration and increase the advertising
Prototype demo and
Task 4.6 robot fails in accomplishing reputation is of the event. It may also be necessary to
presentation
the mission. conveyed. facilitate transportation to the vineyard.
Heavy testing.
WP5 Dissemination & Exploitation
Task 5.1 Dissemination activities: video, Video or website is not 10% Delay in the Efforts in having the video/website AVAN
website, presentation complete on time dissemination of the completed within the following month to
onset of the project the delivery date.
Beneficiary
Estimated Estimated impact
Contingency plan (action to be taken should responsible for
Work package and Task Risk occurrence of the project if the
the risk occur) risk managing
probability risk happens
(short name)
Task 5.2 Exploitation plant and The parts do not agree with 5% Delay in the An independent expert will be consulted AVAN
dissemination the rest of the Consortium exploitation
Task 5.3 International congress Failure in the organization 5% Significant Attendance to international trade shows to AVAN
of the international diminishment of the present the multi sensor robot
congress project visibility and
final users’
awareness
Task 5.4 Publications in scientific and Manuscripts are rejected by 15% No publication of Reviewing the manuscripts and AVAN
technical journals editors the project results resubmission to other journals
WP6 Management
Task 6.1 Management of the WPs Meeting foreseen are not 10% Reduced interaction Additional occasions of discussions will be UDLR
enough for managing the between partners arranged as conference calls, web
project resulting in conferences, etc...
diminished
integration between
tasks and WPs
Task 6.2 Interaction among WPs No risk - - - UDLR
Task 6.3 Interaction with the Advisory People invited to AB 10% Low feedback from Selection of new candidates for the AB. UDLR
Board declines or does not the AB Efforts to stimulate participation and
participate actively feedback through personal contact
Task 6.4 IPR handling and knowledge Conflict between partners 5% Problems related to Efforts in solving conflicts on the basis of UDLR
management IPR the principles agreed in the Consortium
Agreement
Table B.1.3. VINEROBOT project Gantt chart (milestones are in bold and deliverables in shadow cells)
Months
Work packages and Tasks
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
1
2
3
4
5
6
7
8
9
D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N
WP1 Technical and agronomical requirements
Task 1.1 Agronomical and physiological
X X X X X X X X X X X X
requirements
Task 1.2 Multi-sensor robot requirements X X X X X X X X X X X X
Task 1.3 Sensor Fusion and modelling
X X X X X X X X X X X X
requirements
Task 1.4 Data management and display
X X X X X X X X X X X X
requirements
WP2 Multi-sensor scouting robot
Task 2.1 Design of the robot X X X X X X X X X X X X X X X X X X X X X X X X X X X X
Task 2.2 Design of sensory system for the
detection of canopy and fruit X X X X X X X X X X X X X X X X X X X X X X X X X X X X
under dynamic conditions
Task2.3 Design of sensory system for
physiological and agronomical
X X X X X X X X X X X X X X X X X X X X X X X X X X X X
parameters of vineyards under
dynamic conditions
Task 2.4 In field real-time data acquisition X X X X X X X X X X X X X X X X
Task 2.5 Design & construct. optimized ver. X X X X
Task 2.6 Progress field demonstrations X X
WP3 Fusion and modelling
Task 3.1 Data analysis X X X X X X X X X X X X X X X X X X X
Task 3.2 Building of algorithms and models X X X X X X X X X X X X X X X X
Task 3.3 Data integration and fusion X X X X X X X X X X X X X X X X X X X
Task 3.4 Establish. of agricult. calibr. curves X X X X X X X X X X X X X X X X X X X X X
Task 3.5 Models’ validation X X X X X X X X X X
Annex 1 – Description of Work – Part B Page 20 of 70
Version N°3 agreed with the EC services on 28-06-2013
FP7-ICT-2013-10-610953
Months
Work packages and Tasks
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
1
2
3
4
5
6
7
8
9
D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N
WP4 Data management & display
Task 4.1 Data transmission and storage X X X X X X X X X X
X X X X X X X X X
Task 4.2 Implementation of fusion models X X X X X X X X
X X X X X X X X X
Task 4.3 Display of agricultural information X X X X X X X X X X X X X X X X
X X X X X X X X X
Task 4.4 Complete system validation X X X X X X X X X X
X X X X X X X X X
Task 4.5 Deployment of prototype X X X X X X X X X X X X X X X X X X X X X
Task 4.6 Prototype demo and presentation X X
WP5 Dissemination & exploitation
Task 5.1 Video, presentation and website X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
Task 5.2 Exploitation plans and
X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
dissemination
Task 5.3 International congresses X X X X X X X X X X X X
Task 5.4 Publications in scientific and
X X X X X X X X X X X X X X X X X X X X X X
technical journals
WP6 Management
Task 6.1 Management of the WPs X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
Task 6.2 Interaction among WPs X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
Task 6.3 Interaction with the Advisory
X X X X X X X X X X X X X X X x X X X X X X X X X X X X X X X X X X X
Board
Task 6.4 IPR handling and knowledge
X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
management
References cited
Agati, G., Meyer, S., Matteini, P. and Cerovic, Z.G. 2007. Assessment of anthocyanins in grape
(Vitis vinifera L.) berries using a non-invasive chlorophyll fluorescence method. Journal of
Agricultural and Food Chemistry 55, 1053-1061.
Arnó, J., Martínez Casasnovas, J., Ribes-Dasi, M. and Rosell, J.R. 2009. Review. Precision
Viticulture. Research topics, challenges and opportunities in site-specific vineyard
management. Spanish Journal of Agricultural Research 7,779-790.
Baluja, J. M.P. Diago, L. Rocchi, H. Ochagavia and J. Tardaguila.2011. Using a fluorescence
proximal sensor to study spatial variability of grape phenols in a Tempranillo vineyard. 8th
European Conference on Precision Agriculture. Prague, July, 2011.
Ben Ghozlen, N., Cerovic, Z.G., Germain, C., Toutain, S. and Latouche, G. 2010.Non-invasive
optical monitoring of grape maturation by proximal sensing. Sensors 10 (11), 10040-10068.
Berenstein, R., Shahar, O.B., Shapiro, A. and Edan, Y. 2010. Grape clusters and foliage detection
algorithms for autonomous selective vineyard sprayer. Intel. Serv. Robotics 3, 233-243.
Blasco J., Aleixos N. and Moltó E. 2007. Computer vision detection of peel defects in citrus by
means of a region oriented segmentation algorithm. Journal of Food Engineering 81(3), 535-
543.
Blasco, J., Aleixos, N., Gómez-Sanchís, J.andMoltó E. 2009a.Recognition and classification of
external skin damages in citrus fruits using multispectral data and morphological
features.Biosystems Engineering. 103(2), 137-145.
Blasco, J., Cubero, S., Gómez-Sanchís, J., Mira, P. and Moltó E. 2009b. Development of a
machine for the automatic sorting of pomegranate (Punica granatum) arils based on
computer vision.Journal of Food Engineering 90(1), 27-34.
Bulanon, D.M., Burks, T.F. and Alchanatis, V. 2009. Image fusion of visible and thermal images
for fruit detection.Biosystems Engineering 103, 12-22.
Cerovic, Z.G., Moise, N., Agati, G., Latouche, G., Ben Ghozlen, N. and Meyer, S. 2008. New
portable optical sensors for the assessment of winegrape phenolic maturity based on berry
fluorescence.Journal of Food Composition and Analysis. 21, 650-654.
Chaerle, L., Lenk, S., Leinonen, I., Jones, H.G., Van Der Straeten, D. and Buschmann, C. 2009.
Multi- sensor plant Imaging: Towards the development of a stress catalogue. Biotechnology
Journal 4, 1152-1167.
Cubero, S., Aleixos, N., Moltó, E., Gómez-Sanchis, J. and Blasco, J. 2010a.Advances in machine
vision applications for automatic inspection and quality evaluation of fruits and vegetables.
Food and Bioprocess Technology, 1-18.
Diago, M.P., Correa, C., Millan, B., Barreriro, P., Valero, C., Tardaguila, J. 2012. Grapevine yield
and leaf area estimation using supervised classification methodology on RGB images taken
under field conditions. Sensors, 22, 16988-17006.
Johnson, L.F. 2003.Temporal stability of an NDVI-LAI relationship in a Napa Valley vineyard.
Australian Journal Grape Wine Research 9, 96-101.
Lamb, D.W., Weedon, M.M. and Bramley, R.G.V., 2004. Using remote sensing to predict
phenolics and colour at harvest in a Cabernet Sauvignon vineyard: Timing observations
against vine phenology and optimising image resolution. Australian Journal Grape Wine
Research 10,46-54.
Lee, W.S., Alchanatis, V., Yang, C., Hirafuji, M., Moshou, D. and Li, C. 2010.Sensing technologies
for precision specialty crop production. Computers and Electronics in Agriculture 74, 2-33.
López-García, F., Andreu-García, G., Blasco, J., Aleixos, N. and Valiente, J.M. 2010.Automatic
detection of skin defects in citrus fruits using a multivariate image analysis approach.
Computer and Electronics in Agriculture 71, 189-197.
Nicolaï, B.M., Beullens, K., Bobelyn, E., Peirs, A., Saeys, W., Theron, K.I. and Lammertyn, J. 2007.
Nondestructive measurement of fruit and vegetable quality by means of NIR spectroscopy:
A review. Postharvest Biology and Technology, 46, 99-118.
Pölling, B; Herold, L.; Volgmann, A.; Wurbs, A.; Werner, A. 2010. Assessing the potential use of
Precision Farming- Technologies in the EU. In Proceedings of the Robotic Conference
(EuroAgEng). Clremont-Ferrand (France), September 2010.
Rovira-Más, F. 2012. Global-referenced navigation grids for off-road vehicles and environments.
Robotics and Autonomous Systems 60: 278-287.
Rovira-Más F; Banerjee, R. 2013. GPS data conditioning for enhancing reliability of automated
off-road vehicles. Journal of Automobile Engineering 227(4): 78-92.
Sáiz-Rubio, V; Rovira-Más, F. Dynamic segmentation to estímate vine vigor from ground images.
Spanish Journal of Agricultural Research 10(3): 596-604.
Sela, E., Cohen, Y., Alchanatis, V., Saranga, Y., Cohen, S., and Möller, M. 2007. Thermal Imaging
for estimating and mapping crop water stress in cotton.In J.V. Stafford (Ed), European
conference in precision agriculture, 365-371.
Sepulcre-Cantó, G, M.P. Diago, P. Balda, F.Martínez de Toda, F. Morales, J. and Tardaguila. 2009.
Monitoring vineyard spatial variability of vegetative growth and physiological status using an
unmanned aerial vehicle (UAV). International Symposium of Viticulture Giesco, Davis,
EE.UU.
Tardaguila, J. and Diago, M.P. 2008.Precision viticulture: Principles and technologies for the
vineyard. World Wine Forum, Logroño, Spain.
Xiao-bo, Z., Jie-wen, Z., Yanxiao, Li. and Holmes, M.2010.In line detection of apple defects using
three color cameras system.Computers and electronics in Agriculture 70,129-134.
Zerger, A., Viscarra-Rossel, R.A., Swain, D.L., Wark, T., Handcock, R.N., Doerr, V.A.J, Bishop-
Hurley, G.J., Doerr, E.D., Gibbons, P.G. and Lobsey, C. 2010. Environmental sensor networks
for vegetation, animal and soil sciences. International Journal of Applied Earth Observation
and Geoinformation 12, 303-316
B2. Implementation
To strengthen the PMS, a matrix is created to cross reference structure components, personnel
involved, scheduled meetings, issues and responsibilities as shown in Table B2.1.
The management board will meet every 6 months. If required video/call conferences will be
organized to update information to each individual partner. Meetings will involve formal
presentations from each WP or tasks by partners actively involved with them. The board will
pay particular attention to the subcontracted activities performed by the RTDs, managing both
technical progress and value for money. The board will suggest corrections or changes to WP
and tasks based on the reports and presentations. Three intermediate reviews at months M7,
M12, M24 and a final one at M36 are scheduled (See WT5). At the mid-term reviews, the
project will be assessed against milestones and deliverables listed in the Work Programme.
At the final meeting, the board will assess all deliverables and reports with particular focus on
the reports outlining post- project work on exploitation and use of the project results. The
board will take an active role in suggesting strategic changes and corrections to the project, if
needed to fulfil the overall goals. This implies that the board will seek to amend the DoW, if
needed, and duly informing the PO of such changes. The board will closely monitor risks and
suggest how to cope with these according to the risk management plan.
Apart from the mandatory annual technical and financial reports to the EC, three additional
Progress reports are scheduled in M6, M18 and M30 to be delivered to the PO summarizing the
managerial and technical progress of the project. These progress reports constitute deliverables
D6.2, D6.5 and D6.7.
The operational management of the project will be performed by the Coordinator, the Technical
Manager, and the IPR and Dissemination Manager. These managers will have clearly defined
roles and report to the management board.
Project Coordinator
Prof. Javier Tardaguila, associate professor of Viticulture at the University of La Rioja, has
conducted many research activities in precision viticulture and canopy management, and
possesses an extensive experience in the wine sector.
Project progress review and management against scientific and technological objectives.
Workflow scheduling and Work Plan change control procedure.
Communication between partners within the work package.
Provision of the minutes taken at these meetings.
Coordination of technical activities between partners within work packages.
Reviews and management of impact on economic and societal issues.
Eventually, the Technical Manager may have to spend a 2-4 week period in Japan
(Sapporo and Kyoto), on a research stay (see specification below).
ACTIVITY LOCATION DATE REASON
Month 18 is the stipulated date for accomplishing
milestone M2.1 (robot delivery) and work must be
in progress to fulfill M2.2 (month 24) in which
perception sensors must be integrated. This is a
critical point for the success of the robot
Research Sapporo and July performance, and it would be advisable to confront
leave Kyoto, Japan 2015 robot behaviors and navigation laws with two
recognized experts: Prof. Noboru Noguchi at
Hokkaido University and Prof. Naoshi Kondo in
Kyoto University. The idea is to spend two weeks in
each lab and learn how their robots have solved
similar problems, visiting also the BRAIN research
center where several agricultural robots have been
The trip to Japan would meet two goals: 1) VINEROBOT would be widely introduced in the
major institutions visited and 2) Key knowledge is expected to be gained to improve and
optimize algorithms performance. The ROI of the Japan trip is estimated in 50000 €, as the
increase of sales by Wall Ye and the consortium will potentially increase.
Technical Manager 1
Prof. Francisco Rovira-Más, associate professor and head of the Agricultural Robotics Lab at the
Polytechnic University of Valencia (UPV), with an ample experience abroad in machine vision,
robotics and intelligent vehicles.
Dr. Julio Gomez, CEO and founder of Avanzare, with 16 years of nanomaterials’ experience at
different universities and research centres in Spain, UK, France and Portugal.
Decision making and conflict resolution mechanisms and responsibilities for each individual
partner
The principle focus on strategic decision making falls on the management board as already
mentioned. Also all decisions that are controversial and/or disagreed upon among partners will
be addressed by the board. All partners may bring up issues to the management board level if a
solution is deemed necessary. This will ensure that all partners have fair play and a chance to be
heard before a decision is made. The board is obliged to hear all partners and discuss alternative
solutions to reach an agreement. If discussions do not lead to an agreement, then the issue is
decided by simple voting. In the board each partner has exactly one vote. The Grant Agreement
overrules the Consortium Agreement and any decision that might be taken at the consortium
level.
If an urgent solution to a problem is required, the management board meets via video/call-
conferencing to discuss and solve the problem using the rules above. Daily work decisions are
1
In the event of problems or unexpected deviations in a work package, the Technical Manager will flag the issue to the Coordinator
and, if needed, will bring it to the Management Board. The board will then determine the appropriate course of action such as
further development iterations, adoption of alternative technologies or alternative development routes.
made at the appropriate project activity level. The coordinator, the technical manager or the
IPR/exploitation manager should, however, be consulted and informed promptly on decisions
taken at the WP leaders’ level. To this extent decisions made by WP leaders may be brought up
to the management board, which in turn can overrule such decisions.
It is particularly of utmost importance that each partner and the managers behave and perform
in ways that create an open and positive atmosphere for the project. They are to spot
controversies as early as possible, so that a reasonable discussion and agreement can be
achieved promptly. In case of conflicts at the activity level, the coordinator will mediate and find
solutions. In case no solution is found the matter will be brought to the Management Board.
Risk Assessment can take place at any time during the project. It allows the project
management to:
1. Take whatever actions are needed in advance to avoid, reduce or mitigate the effects of
risks. It is better to spend money on mitigation than to include contingency in the plan
(See Table 1.1.) for contingency plan on technological risks).
2. For all those risks which are deemed to be significant a contingency plan is established.
3. Measure, monitor and track the effects of the risks identified and handle these to a
successful end.
As tools in the process, two report tables will be used:
1. A Risk Assessment Report, in order to manage correct risk evaluation and identification.
2. A Risk Action Report, in order to manage a proper risk mitigation action plan.
Kinds of Risk: Initially, the following three kinds of risks will be monitored in VINEROBOT:
1. Management risks.
2. Technological risks, already addressed in previous section B 1.3.1. (Table B1.1.).
3. Financial and administrative risks.
As the project develops, the management board decides over other risks that need to be
managed such as, for example, market-dependent risks. In the tables below risks in the above
categories are identified.
B 2.2 Beneficiaries
Televitis Group
Project coordinator leads WP6, and carries out the following major tasks:
Provides technical and agronomical requirements including the multi-sensor robot, fusion
and modelling and data management requirements (WP1).
Participates in the design of the on-the-go multisensory robot (WP2).
Contributes to data fusion and modelling (WP3).
Supports partners in the data management process (WP4).
Partakes in the platform validation with the end user (WP4).
Participates in the in-field demonstrations (WP2 & WP4)
Contributes to the dissemination of the results (WP5).
Key people behind the Televitis group are the following:
Dr. Javier Tardaguila (20% FTE )
Coordinator, head of the Televitis Group, is an Associate Professor of Viticulture at the
University of La Rioja. He holds a Bachelor in Agricultural Engineering, a Master’s degree
in Viticulture and Oenology (Madrid 1989) and a PhD in Viticulture (University of Padua,
Italy 1994). He was visiting researcher at the University of California (Davis, USA) in 1994
and in DPI (Mildura, Australia) in 2005. Research activities include canopy management
and precision viticulture. He has coordinated viticulture research projects at national and
regional levels. Currently, he participates in FP7 projects (MoDeM_IVM, INNOVIVE,
VineMan.Org), and in several others at national and regional levels.
EDMANS Group
The EDMANS (Engineering Data Mining And Numerical Simulations) group focuses on data
mining, machine learning, predictive modelling, finite element analysis (FEM), and
computational fluid dynamics (CFD). It is one of the most active groups at University of La Rioja.
Over the last 10 years, 9 EC, 11 national government projects, and more than 50 other R&D
ones with Spanish and European companies have been tackled. The work has yielded more than
50 papers, 25 PhDs and 15 patents (7 being exploited).
Key people involved behind the EDMANS group are depicted below:
Dr. Francisco Javier Martínez de Pison Ascacíbar (15% FTE)
The group leader is an Associate Professor of Data Mining and Project Management at
University of La Rioja. He holds a Bachelor Degree in Electronics Engineering, a Master’s
degree in Industrial Engineering and a PhD in Data Mining and Machine Learning.
Research activities concern the use of data mining and machine learning methods in
several different fields such as industry, agriculture and environment. He has coordinated
many research projects and has an extensive experience with EC projects.
Some national grapevine-related programs (VINNEO, DEFISTIM) and other EC projects (CROPS,
INNOVINE) are being currently addressed. Previous experiences with designing sensors
integrated on robots for mapping on-the-go agronomical parameters in vineyards, establishing
calibration curves for anthocyanin data on grape clusters, and setting up calibration curves for
chlorophyll data on leaves are added to the project.
The main tasks to be carried out by the WALL partner are as follow:
Lead the design of the multisensory robot operational requirements (leader of WP2).
Participate in data fusion and modelling (WP3).
Support the partners in the data management process (WP4).
Participate in the platform validation (WP4).
Participates in the in-field demonstrations (WP2 & WP4)
Dissemination activities and management (WP5 and WP6).
Key personnel involved in WALL partner’s participation in the project are given below:
Christophe Millot (15% FTE)
Computer programmer with expertise in Delphi, Assembleur, WINDEV, NT, MYSQL, he
has worked for companies such as Alsthom, GAE and DRTEFP. Developer of mapping
software for vinegrowers, the Association Viticole d’Alsace, Bureau Interprofessionnel des
Vins de Bourgogne, Institut National de l’Origine et de la Qualité, Syndicats Viticoles,
Fédération Viticole, Organismes de Défense et Gestion and Cooperatives wineries.
Manu Goncalves
He has a technical background including electronics, robots design and communications
and software of control
Guy Dorido
He has technical background including electronics, robots design and communications
and software of control.
Previous experience with remote surveillance and territory supervision systems for the Military
and the Civil Protection Agencies is sufficient to cover all the production-related aspects related
to the partner’s role in the following main tasks:
Contribute to the definition of data display and management requirements (WP1)
Participate in the design of the on-the-go multisensory robot (WP2).
Contribute to data fusion and modelling, RBG acquisition, and integrating the design, the
disposal and the connection of signals (WP3).
Provide technical support with the connective parts of the sensor and the robot (WP4).
Participates in the in-field demonstrations (WP2 & WP4)
Partake in dissemination activities and management (WP5 and WP6).
With previous experience in monitoring vineyards, the Buzet partner provides 80 acres of great
testing ground, all the vine work to be done, and actively carries out the following tasks:
Participates in the selection of sensing needs (WP1 and WP3).
Participates in the platform validation, and prototype testing on its premises (WP4).
Participates in the in-field demonstrations (WP2 & WP4)
Contributes to dissemination activities and management (WP5 and WP6).
Key people from the partner participating in the project are the following:
Mr. Sébastien Labails (14% FTE)
He holds an Agricultural Engineering degree, and is the vineyard manager in the “Buzet”
denomination. He counsels its vine growers from planting to harvest. He has a huge
experience in vine grapes logistics and a high sensibility in vineyard observation (soil,
physiology and maturation) over 2,000 ha. He is the manager of the sustainable
development in vineyards with all winegrowers’ tenants, and the responsible for training
programs.
Dagieux Sophie
Laboratory technician of Vignerons de Buzet
Carine Magot
Vineyard manager of Vignerons de Buzet, engineer
Damien Charbonnier
Financial and legal coordinator for the project, financial controller at Vignerons de Buzet
Key people participating in the project from the Geisenheim University partner are as follow:
Dr. Manfred Stoll (5% FTE)
He holds two degrees (Engineering and Biology), and a PhD. He leads the department of
Viticulture. As a biologist and viticulturist, his research aims at improving the
ecophysiological understanding on grapevines, and implementing such results to applied
management practices. His specific areas of research focus on thermal imagery and proxy
sensing, vineyard adaptation and variability to environmental stress as well as the impact
of climate change to current and future production.
application of global grids to quantize the spatial variability of orchards. In 2011, the Edmund
Optics Research & Innovation Award was awarded to the lab as well as the Valencia Idea Award.
One of its current members received in 2012 the CIGR Armand Blanc Prize for a conference
paper and in 2013 the Farming by Satellite Prize organized by the European Commission.
Partner Avanzare has an extensive experience in different national and European projects
related to the development of sensors for instance, SECUREVIN and DEMETER. In addition, the
partner is well-known in the European scenario for projects such as DIBBIOPACK, BUONAPART-
E, NANOMASTER, PHOENIX, etc. Avanzare has developed new sensors for selecting parameters
and nanotechnology miniaturization. In addition the industrial point of view brought into the
project by Avanzare provides technical and agronomic requirements through existing and
potential customers.
Avanzare carries out the following main tasks within the project:
Establishes the agronomical and expected sensor parameters requirements (WP1).
Supports data fusion and modelling tasks (WP3).
Supports partners in the data management process (WP4).
Participates in the in-field demonstrations (WP2 & WP4)
Takes part in the platform validation (WP4).
Leads WP5 (dissemination of the results).
Support in the management of the project (WP6).
Key people on behalf of the AVAN partner are as follow:
Dr. Julio Gomez (7% FTE)
As mentioned earlier, he is Avanzare’s founder and CEO and responsible for IPR and
disseminations issues with the project. He has been an active investigator in more than
120 research projects, and the author of 37 papers and 4 books. Julio has strong
experience in nanoparticles preparation and self-assembly.
Lucía Badaya
Diego Ruiz
Sara Ortega
Dr. Miriam Puelles
In terms of partners’ geographical localization, UDLR, UPV and AVAN, are sited in Spain, which is
the first country in the worldwide ranking of vineyard surface and produces one third of the
total EU27 grape production. All three partners are located in two economically and socially
important wine producing regions in Spain—la Rioja and Valencia—which account for 131,490
ha of vineyard surface (12.7 % of Spain vineyard surface and 3.7 % of EU vineyard surface) and
produce 4.3 million of litters of wine (12.2 % of Spain’s production and 2.6 % of EU production)
(ICEX report 2011). GEIS is localized in Germany, where the wine industry is focused not only on
grape and wine production (9.4 millions of hl) but also in commercialization and trade. Germany
was the second major export country with a value of 334.1 million € (OIV 2012 Report). Italy,
host country of one of the SMEs (SIVI), is a remarkable example of the application of the
techniques in agriculture, and ranks second as producer. Undoubtedly, France is one of the
major role players in the worldwide and European wine industry (802 thousand ha of vineyard,
accounting for a 22.9 % of total EU vineyard surface, and 49.6 million L of wine, representing the
31.6 % of the EU wine production). Three partners—two technology-developers and a winery—
are located in France, FORC, WALL and VIBU, respectively.
Figure B2.2. Geographical localization of the VINEROBOT project partners in the EU27
All these countries (France, Spain, Italy and Germany) are well-known wine producers with
peculiarities in each country. This diversity represents a synergic approach to the project. Other
participants are involved in an Advisory Board including experts in the tasks of the project, and
representative persons for the end-users of the project results (farming associations, wineries,
and consumer’s associations).
Simon Blackmore. Professor and Head of Engineering at Harper Adams University (UK).
e-mail: simon.blackmore@harper-adams.ac.uk.
His field of research and expertise deal with: Precision Farming; Agricultural robots and
smart machines; Instrumentation systems; Farm management information systems and
decision support.
Additionally, a careful revision of the past and on-going projects on multi-sensory and robotics
in agriculture has yielded the following projects, which may be contacted in order to exchange
expertise, experience and knowledge.
CROPS (http://www.crops-robots.eu/)
Intelligent sensing and manipulation for sustainable production and harvesting of high value
crops, clever robots for crops. FP7-NMP-2009-3.4-1. Reference: 246252
Objective
CROPS will develop scientific know-how for a highly configurable, modular and clever carrier
platform that includes modular parallel manipulators and intelligent tools (sensors, algorithms,
sprayers, grippers) that can be easily installed onto the carrier and are capable of adapting to
new tasks and conditions. Several technological demonstrators will be developed for high value
crops like greenhouse vegetables, fruits in orchards, and grapes for premium wines.
The CROPS robotic platform will be capable of site-specific spraying (targets spray only towards
foliage and selective targets) and selective harvesting of fruit (detects the fruit, determines its
ripeness, moves towards the fruit, grasps it and softly detaches it).
Another objective of CROPS is to develop techniques for reliable detection and classification of
obstacles and other objects to enable successful autonomous navigation and operation in
plantations and forests. The agricultural and forestry applications share many research areas,
primarily regarding sensing and learning capabilities.
Project coordinator: Jan Bontsema. Wageningen University & Research Centre, Applied Plant
Research – WUR (NL)
Partners:
Wageningen UR (NL)
University of Leuven (BE)
Ben-Gurion University (IL)
University of Ljubljana (SI)
UMEA University (SE)
Università degli Studi di Milano (IT)
CSIC, Inst. de Automática Industrial (ES)
Technical University Munich (DE)
Case New Holland NV (BE)
INIA PROGAP (CL)
Force-A (FR)
Festo (DE)
Swedish Univ. of Agricultural Sciences (SE)
Jentjens Machinetechniek (NL)
RHEA (http://www.rhea-project.eu/)
Robot fleets for Highly Effective Agriculture and forestry management. FP7-NMP-2-LA-2010.
Reference: 245986.
Objective
In the last two decades, a precise management of agricultural land has been made possible due
to the availability of new technologies, including global positioning systems (GPS), geographic
information systems (GIS), sensors, automation of agricultural machinery, and high resolution
image sensing. As a result, the concept of Precision Agriculture has emerged as the
management strategy that uses information technologies to collect and process data from
multiple sources in order to facilitate decisions associated with crop production. Moreover, the
EUs sixth environmental action programme addresses the need to encourage farmers to change
their use of plant protection products.
RHEA is focused on the design, development, and testing of a new generation of automatic and
robotic systems for both chemical and physical mechanical and thermal effective weed
management focused on both agriculture and forestry, and covering a large variety of European
products including agriculture wide row crops (processing tomato, maize, strawberry, sunflower
and cotton), close row crops (winter wheat and winter barley) and forestry woody perennials
(walnut trees, almond trees, olive groves and multipurpose open woodland).
RHEA aims at diminishing the use of agricultural chemical inputs in a 75%, improving crop
quality, health and safety for humans, and reducing production costs by means of sustainable
crop management using a fleet of small, heterogeneous robots ground and aerial equipped with
advanced sensors, enhanced end-effectors and improved decision control algorithms. RHEA can
be considered as a cooperative robotic system, falling within an emerging area of research and
technology with a large number of applications as reported by the FP6 Network of Excellence
EURON, Special Interest Group on Cooperative Robotics, funded by the European Commission.
RHEA will be a unique opportunity to gather a very large number of multidisciplinary research
groups with adequate funds to accomplish an authentic step forward in applying precision
agriculture techniques in a massive way. This consortium joints a number of multidisciplinary,
experienced researchers capable of improving individual scientific knowledge, but a large
cooperation project is demanded to sum up the individual efforts in a holistic manner. The
success of RHEA could bring a new means of applying automatic systems to agriculture and
forestry crops with an important impact in improving the economy and environment as well as
in maintaining the sustainability of rural areas by launching new technological jobs.
Project coordinator: Pablo González-de-Santos. Center for Automation and Robotics (UPM-
CSIC). Madrid (ES)
Partners:
CogVis GmbH (Austria)
Forschungszentrum Telekommunikation Wien Ltd (Austria)
Cyberbotics Ltd (Switzerland)
Università di Pisa (Italy)
Universidad Complutense de Madrid (Spain)
Tropical (Greece)
Soluciones Agrícolas de Precisión S.L. (Spain)
Universidad Politécnica de Madrid (UPM) (Spain)
AirRobot GmbH & Co. KG (Germany)
Università degli Studi di Firenze (Italy)
Centre National du Machinisme Agricole, du Génie Rural, des Eaux et des Forêts -
CEMAGREF (France)
Case New Holland Belgium N.V. (Belgium)
Bluebotics S.A. (Switzerland)
CM Srl (Italy)
FUTUREFARM (http://www.futurefarm.eu/)
Objective
In the future European farmers will have to effectively manage information on and off their
farms to improve economic viability and to reduce environmental impact. All three levels, in
In addition, the communication between agriculture and other sectors needs improvement.
Crop products for the value added chains must show their provenance through a transparent
and certified management strategy and farmers receiving subsidies are requested to respect the
environment through compliance of standards. To this end, an integration of information
systems is needed to advise managers of formal instructions, recommended guidelines and
implications resulting from different scenarios at the point of decision making during the crop
cycle. This will help directly with making better decisions as the manager will be helped to be
compliant at the point and time of decision making.
Partners:
3D-MOSAIC (http://www2.atb-potsdam.de/3d-mosaic/)
Advanced Monitoring of Tree Crops for Optimized Management - How to Cope With
Variability in Soil and Plant Properties? ERA-Net ICT-Agri transnational project
Objective
In the context of current global changes, assuring the supply of fresh produce and increasing
economic viability are priority targets within plantation management. In the cultivation of tree
crops, water is a critical input factor and irrigation is necessary in all European countries. As a
matter of fact, input requirements in an orchard vary in space and time due to the variability of
climate, soil and plant growth.
Spatial patterns of soil and plant properties can be regarded as a 3D-Mosaic. Thus, an optimum
orchard management has to address seasonal and spatial variability in soil and micro-climate
that affect plant growth and fruit development. However, irrigation today is commonly
managed at the orchard level and not at the tree level. Uniform irrigation frequently creates
sub-optimum conditions with some parts of the orchard having insufficient water while other
parts suffer from water logging and subsequent oxygen shortage. Consequently, water is
wasted and/or yield and fruit quality are reduced. Tree-scale leaf area index and fruit load and
individual fruit-scale size, water content, and maturity-related pigmentation represent
information vital for orchard management resp. irrigation decision making. Until now, these
parameters have been underutilized due to lack of automation of their monitoring.
Project coordinator: Manuela Zude. Leibniz Institute for Agricultural Engineering Potsdam-
Bornim (DE).
Partners:
Finally, the presence of SMEs, RTD performers, grapevine growers, wineries, and other end-
users guarantees the interest for the topic along the entire grapevine production chain from
farm to fork. Table B2.4 shows a summary of the skills and contribution to the planned work of
the VINEROBOT partners.
Table B2.4. Matrix of the skills required, and partners expertise and contribution to develop the VINEROBOT project.
3. Fusion and modeling - Good ability to analyze and WALL: It has developed an advanced Lead the design of the
combine data from multiple robot for use in agriculture and, multisensory robot operational
sources and type. particularly, in viticulture. Specialized in requirements (leader of WP2).
-Data mining, algorithm building distribution of mapping (GisMap) and Participate in data fusion and
expertise and modeling traceability software (GisMapTrace),
Partners are convinced that the overall financial plan is adequate for the successful project
execution.
Table B2.6. Percent allocation of total budget and staff effort to the different activities
RTD activities receive the most part of resources in terms of budget (91.0%) and staff efforts
(98.5%). The budget assigned to UDLR is higher than that for the other RTD performers; it
covers several activities and has more personnel as well as project management efforts of the
RTDs. No costs are assigned for durable equipments.
Dissemination activities account for a small part of the project resources (3.8 % of total budget,
Figure B2.3.a). This does not mean that they are considered of negligible relevance. Efforts will
be devoted by all partners in dissemination of the results using their contact network with
technicians, farmers, journalists of the specialized press, etc. An important role is given to the
website. Publication in scientific journals is considered free of costs (except for some specific
journals), a small budget will be used to organise a one-day international congress within a
relevant event for the vine-wine chain and to print technical leaflets and proceedings for the
congress (see WP5 for details). Demonstration activities are considered part of the
dissemination action and included under the RTD budget. These activities are key steps for the
project. These resources are committed to SMEs; the three RTD performers are indirectly
involved because in-vineyard demonstrations will be organised by the SMEs in the vineyards
closed to WALL-YE and in VIBU premises (see WP2 and WP4 for details).
Management activities receive 5.2 % of the total budget. Due to the duration of the RTD work
packages, WP2 to WP4 last until M31 in some of the tasks, this percentage is a bit higher than
other projects. These resources are mainly allocated to manage the project activities,
knowledge circulation and IPR issues, to participate in meetings planned within the project and
to cover travel costs for the management board (see WP6 for more details).
19.5% UDLR
RTD FORC
WALL
DISS SIVI
8.5% VIBU
MGM
GEIS
7.2%
UPV
4.2% 15.2% AVAN
91.0% 9.6%
Fig B2.3.a Percent of total budget allocation Fig B2.3.b Percent distribution of the EC
to the different activities. contribution among partners.
Total Direct Costs for each partner are summarized in Table B2.7. Allocation of Subcontracting
and Other Costs by partner (in €), and detailed in Tables B2.8.
Cost 1. UDLR 2. FORC 3. WALL 4. SIVI 5. VIBU 6. GEIS 7. UPV 8. AVAN Total
Subcontracting 2,500 18,000 59,000 2,500 7,857 89,875
Travels 19,000 3,000 16,400 11,000 3,000 3,000 36,000 6,000 97,400
Equipment 1,000 8,200 20,000 500 3,000 51,000 83,700
Consumibles 4,000 4,947 2,400 2,000 500 2,000 5,000 37,250 58,097
O. Direct Cost 3,000 3,000
Total 26,500 25,947 86,000 33,000 4,000 8,000 97,500 51,107 332,072
EC contribution
Requested EC contribution represents 76.9% of the whole cost of the project. The consortium
will distribute the contribution among partners as shown in Figure B2.3.b. No other enterprises
or end-users will financially contribute to the consortium. RTD performers do not retain
ownership of foregrounds and thus they do not co-invest in the project.
UDLR
FORC
WALL
SIVI
VIBU
GEIS
UPV
AVAN
B3. Impact
B 3.1 Strategic impact
This project aims at delivering an innovative use-case autonomous agricultural robot endowed
with advanced mobility behaviours and equipped with several non-invasive sensing
technologies to provide a complete, non-destructive, and site-specific precise assessment of
vineyard status.
The European market. The market targeted for the new scouting robot is the EU27 wine
industry, including European wine producers and their supplier services. As indicated in the last
OIV report (OIV, 2012), Europe retains almost half of the world vineyard surface (47.1%),
accounting for 44% of the total grape production in the world, including wine table and wine
grapes (Figure B3.1a). The EU27 is the world’s leading wine producer, making around 165
million hl every year, and accounting for 66.5% of the world’s wine production (Figure B3.1b),
57% of global consumption, and 70% of exports in global terms (OIV, 2012).
A 5.0% B
3.0%
5.9% 4.0%
Europe Europe
21.0% 19.0%
44.0% Asia Asia
America America
Africa Africa
5.5%
Oceania 66.5% Oceania
26.7%
Figure B3.1- A. Grape production worldwide (OIV, 2012); B. Wine production worldwide (OIV,
2012)
The powerful wine industry in the EU27 is composed of 1.5 million holdings (16% of all
agricultural farms), utilising a vine area of 3.4 million ha (2% of the total agricultural area), and
employing 1.5 million annual working units (15% of agricultural work). In 2006, wine production
totalled € 16.0 billion, which represented 5.4% of the agricultural output (Europe Press Release,
EU wine reform, 2006). Wine production plays a primary role in the agricultural activity of most
wine-producing EU Member States. It represents around 10% of the value of agricultural
production in France, Italy, Austria, Portugal, Luxembourg and Slovenia, and a little less in Spain.
The importance of wine production as economic activity is even greater at regional and local
level: for certain regions the value of wine production exceeds 20%, or even 30%, of total
agricultural production, and the percentages can even get higher when considering more
restricted geographical areas as southern Italy and Sicily. With regard to wine producers in 2011
for the EU27 Member States, 35.000 wineries were active and about 500 companies supplied
auxiliary services (Zoomvino, OIV, Eurostat) (Table B3.1).
In terms of individual countries, Spain ranks in first place with regard to vineyard surface (Table
B3.2) whilst France leads in wine production and number of wine producers (Table B3.3),
followed by Italy, Spain and Germany, respectively. These four countries (i.e. the countries
involved in the VINEROBOT Consortium) play a key role in the EU27 wine industry, hosting more
than 80 % of wine producers in the EU27, and accounting for more than the 80% of EU27 wine
production, covering the 82% of the total area cropped with grapes in the EU27 (OIV, 2012)
(Table B3.3).
Table B3.1. Characteristics of the European (EU 27) wine market and wineries’ suppliers
Economic indicator Unit Parameters Relevance % References/ data
of total source
Table B3.2. Vineyard surface and wine production in EU27 countries and worldwide.
As far as export figures are concerned, the worldwide market share (expressed in hl of wine) is
leaded by the 5 largest EU exporters (Italy, Spain, France, Germany and Portugal, respectively),
which account for 65.5 % of the total (OIV 2012) (Figure B3.2). In terms of value, according to
the Global Trade Atlas (GTA), which compiles data from the customs of various countries, the
global amount of wine and grape juice exports in 2011 raised to 23.264 million euros (€).
France is by far the world export leader (3,568 million €; first quarter of 2012), followed by Italy
(1,786 million €, first quarter of 2012) and Spain (1,144 million €, first quarter of 2012).
25.2% 23.5%
Italy
Spain
France
2.8% Germany
3.9% Portugal
21.5%
Southern hemisphere +USA
13.6%
Figure B3.2- World wine exports (expressed in wine volume) in 2011 (OIV, 2012)
The scientific and technological impact is mainly driven by the innovative nature of the vine
scouting robot, as no other similar product exists in the market, in spite of the need for
technology in the highly competitive world of wine making. The two main novelties are: i) the
robot will be able to navigate in any standard vineyard for at least four hours without human
intervention (at present there are no commercial robots that can be left alone in outdoors fields
for a timeframe longer than a demo duration of several minutes); and ii) at the end of the
monitoring mission, the robot will provide a set of crop maps of agronomical parameters in a
compatible format to be easily displayed in regular computers and smartphone-like devices.
Innovative nature of the vine scouting robot. The overwhelming difficulties for vineyard
managers to access and interpret detailed crop information from remote sensing providers
practically leaves the deployment of a medium-size ground vehicle as the only solution to
gather high-level (high resolution + high update rate + high precision) information. While most
of the technology to build and commercialize a scouting robot for agricultural applications is
readily available, its actual construction is never coming, probably due to reliability and safety
issues. In this regard, the reliability of the robot to navigate by itself without provoking any risky
or dangerous situation in the field to plants, vehicles, structures, and principally field workers is
a crucial technological goal and expected outcome of the project.
Those countries more interested in robotic applications as Japan and the USA are releasing
commercial robots of comparable size and aptitudes. So, for example, the Boston-based
company iRobot has been successfully selling the robotic vacuum cleaner RoombaTM, and farm
machinery manufacturer John Deere just released (2012) an autonomous mower for the
European market. The current pressure to release a robot for agricultural tasks and farmer
assistance is rapidly growing, and the first prototype that gets close enough to a working
product will make a tremendous impact in agriculture at a global scale. In fact, one of the
partners of the VINEROBOT Consortium, WALL-YE, has already attempted to build a robot
prototype for vineyard inspection. The success of the new vine robot in the global market will
be strongly favoured by its innovative nature as described in section B1.2.
A comparison between the projected vine robot and possible competitors already in the market
is extremely favourable to the multi-sensor scouting robot, as none of the current commercial
vineyard-sensor products provides autonomous and multi-data integrated information on
grapevine agronomical, physiological, and grape quality features altogether, neither on-the-go
nor for large vineyard areas.
The main potential impact for end-users also involves several aspects: i) end-users will draw
clear economic and environmental advantages from using the scouting robot as they will
optimize vineyard inputs and management costs; ii) the fast and versatile display of the crop
maps makes the robot a flexible and powerful decision-support tool for making educated
decisions based on measurable data.
Economic and environmental benefits for final users. The new product is generally targeted at
vineyard managers, i.e., the people who make crucial decisions about the vineyard’s
management or provide consultancy to grape growers and winery managers. Therefore,
potential end-users can be any of the following beneficiaries: i) technicians of grower’s
consortia, cooperatives or large farm enterprises; and ii) external consultants. The use of the
scouting robot and the immediate display of valuable information on the own robot and
alternatively on smartphone-like devices will promote advanced business operations for any
member of these categories. In general, the robot is intended to provide final users with the
following benefits: i) reasonably accurate yield forecast and consistent grape yield estimations
on one hand, and quality improvements on the other hand, thanks to an optimized crop
monitoring and advanced vineyard management; ii) continuous and accurate vineyard
agronomical, physiological, and grape quality information to better cope with the impact of
climate change and global warming on grape growing; iii) optimization of the labour demand
and costs for vineyard management; iv) more sustainable vineyard management with
optimization of chemical inputs (i.e. fertilizers, fungicides and pesticides for disease and pest
control) and improved use of natural resources (water); and v) meet industry and community
expectations, as well as government regulations on environmental vineyard management and
long-term sustainability.
The quantification of all these economic advantages is not an easy task. At least three major
issues should be taken into consideration for the EU27 members. Firstly, the irrigation
scheduling of vineyards, as water stress conditions certainly do have the largest impact on grape
yield and quality; nevertheless, it may only count for some countries or for very hot growing
seasons. Secondly, disease forecasting, which is also a crucial matter since it affects spray
requirements, environmental friendly production, and crop yield and quality. Thirdly, the price
disparity due to the spatial variation of the fruit quality within a vineyard. So far, this issue is
way underestimated, and it certainly has a huge impact on the entire production chain since it
may result in either different money returns or significant variations in grape quality, which in
turns leads to higher efforts and investments during the subsequent winemaking processes.
In addition to the aforementioned advantages, field technicians and vineyard advisors will take
advantage of the new opportunity to better assist their customers and members of growers
associations by acquiring accurate and fast information for advisory purposes. They may also be
able to reduce their visits to growing sites, focusing on the ones really needed of prompt
corrections, which will be received anytime, anywhere in the world through an efficient
communication framework including field laptops, tablets, and smartphone-like devices.
Conventional wisdom assumes that technical assistance to a winery or grape growing
exploitation involves, on average, 15 working days per year, including desk work and field visits.
The efficient use of the vine robot may reduce this time by about 50% to 70%, increasing the
working capacity significantly, which will eventually be translated into reduced costs for
technical assistance and advisory activities. More importantly, the advent of the robot will
provide European professionals with objective and rigorous criteria for: i) justifying actions
intended to meet yield regulations present in European grape-growing regions of proven high-
quality reputation (i.e. Appellation d’Origine in France, DOC and DOCG in Italy, DOC in Spain,
Qualitätsweinprüfstellen in Germany, etc); ii) outlining vineyard pest and disease management;
and iii) efficiently using vital natural resources such as water, contributing to improve the
sustainability and competitiveness of vineyards. Regarding the use of pesticides, the European
Commission has recently adopted a new legislative framework (European Parliament, 2009)
with a Directive on the Sustainable Use of Pesticides (European Commission, 2009).
In the context of climate change and global warming, the VINEROBOT project will have
significant environmental as well as economical benefits. Irrigation in agriculture already
accounts for about 70% of the total water used worldwide, and the irrigated surface area has
increased linearly since 1960 (Shultz, 2000). A general increase in water demand on a global
scale (urban, industrial, and environmental sectors) has been predicted, and limitations for
agricultural water usage are expected (Fereres and Soriano, 2007). To address this issue, the
European Commission has been long working on preserving and promoting the rational use of
water as a natural resource by issuing the Water Framework Directive (2000/60/EC; COM (2007)
414- final), which identified an initial set of policy options to be taken at European, regional, and
national levels in order to address water scarcity within the EU. This set of proposed policies
intends to move the EU towards a water-efficient and water-saving economy.
Impact for the European Community. The robot is intended to optimize and modernize the
management of European vineyards, thus contributing to their sustainability and to the increase
of competitiveness of grape growers and wine producers (in front of New World wine-
producing countries), which represents a strategic point for Europe that benefits the whole
economy of the Union. The robotic system developed in this project can give a significant
contribution to this objective.
Besides the positive economic impact for the European wine industry, the optimization of such
an essential natural resource as water, specifically in the context of climate change and global
warming, will have a significant environmental benefit for the European community, as a more
rational use of water for irrigation in viticulture will be favoured by the use of the multi-sensor
scouting robot developed through the VINEROBOT project. Furthermore, the reduction of
chemical inputs will also mean a key ecological advantage for the European community. In the
vineyard management attempted with the VINEROBOT platform, the use and quantity of
chemicals are optimised and limited to strictly necessary situations, providing a higher
protection of the environment while maintaining the same productivity levels. Up to date,
69,819 tons of pesticides and fertilizers are employed each year in European vineyards
(Eurostat, 2007). Considering 30% as an average saving in chemical treatments due to the use of
the multi-sensor scouting robot, the reduction in chemical inputs may reach 20,646 tons per
year for the whole Europe. The combined effect of reducing water and chemical inputs
represents an indirect, but not negligible, impact for: i) the environment and ecology of
European ecosystems; ii) the quality of life, health, natural resources, and working conditions of
vine growers (about 2.4 million people over EU27); iii) food safety and health for the consumers.
For all these reasons, this project is in full agreement with the EC document “Assessing
economic impacts of the specific measures to be part of the Thematic Strategy on the
Sustainable Use of Pesticides” ENV.C.4/ETU/2003/0094R and the Water Framework Directive
(2000/60/EC).
The robotic system designed and developed throughout this project will provide the five
participating SMEs with a competitive and unique product that they could not have been able to
develop individually due to its inherent complexity and necessary multidisciplinary approach.
This project will allow the SME involved distinguish themselves from competitors by integrating
their proprietary technology in an original and innovative product with a large potential market,
and consequently, with a clear economic impact.
Regarding the particular SMEs participating in the project, WALL, FORC and SIVI will benefit
from the integration of their products into the new robot, providing their sensors and
equipment to a complete tool package. The French enterprise FORC already offers a range of
fluorescence-based sensors for agricultural purposes, but is still not capable of assuring a
dominant position in the wine-sensing market. WALL and SIVI will integrate the mapping
algorithms, navigation engines, and multi-sensor package into the new robot prototype,
departing from some of their current technology-based products. AVAN, which is already
present in the wine market, will contribute to the prototype commercialization and
dissemination of results, taking advantage of its sales network and customer portfolio. This
project will allow the SME participants to make significant progress by integrating their
technology and expertise in an innovative, original, advanced product geared to satisfy a wide
potential market, bestowing a clear positive economic impact to the participating SMEs. VIBU
will be the first and unique final user to test and validate the robotic product with the
advantages of intensive training, free use, and novelty-driven profits.
The economic impact for the SMEs was calculated on the basis of the financial plan of Table
B3.4. Initial plans intend to start the production and commercialization of the platform on the
first year after the end of the project (project ends in 2016, with the presentation of the
prototype at the VINITECH exhibition in Bordeaux, France). During this year, about a 0.12% of
the market will be targeted and doubled to 0.25% in 2017, topping 0.5% in 2018. The estimated
price for the first-generation robot will round 40.000 €, but a great effort will be made along the
project to decrease this price as much as possible, even from the very first prototype. The
robotic system will be initially introduced in the large producing groups, and immediately in
medium and smaller companies. AVAN has several distributors outside Europe, namely in Chile,
Australia/New Zeeland, the USA, and Mexico.
The proposed research plan can be economically justified by the massive commercialization of
the new platform, as indicated by Table B3.4 and Figure B3.3. Due to the SMEs’ dimension and
their diverse know-how, they can only reach this goal in a reasonable time by setting up a joint
effort through a Consortium like this and accessing to EU funding.
Table B3.3. Time table and financial plan for exploiting the project results (x1000 €)
Economic justification for the research. Considering the five operative years from 2016 to 2020,
the following returns can be estimated according to Table B3.4 and graphically represented in
Figure B3.3, which fully justify the SME´s investment for the proposed project:
A positive EBITDA (net sales-operative expenses) may be achieved in the first year of
commercialization, first semester 2017, and first semester 2018 for cumulated EBITDA.
The net sales and UE support can increase up to 5,9 million € in 2020.
The EBITDA breakeven point will likely be in 2016, the first year of commercialization.
The EBITDA is satisfactory and will allow to sustain the amortization and depreciation.
The ratio EBIT/Net sales & EU support (cumulated data) is 47%, adequate to
demonstrate the profitability of the investments.
7000
€ x 1000
5000
Net sales and UE support
3000 EBITDA
1000 Accumulated EBITDA
-1000
Year
Figure B3.3. Financial trends and breakeven point.
Human resources and employment. New personnel will have to be employed in production,
promotion, and technical assistance to final users and customers. These new employees may be
initially selected among the people involved in the project by the RTD Universities as PhD
students or postdoc associates. This procedure will lead to a double beneficial effect: first, the
new hired personnel will have already been trained and worked with the robot prototype
during the project; and second, employment opportunities for young people entering the labour
market will be created. The participating SMEs will still work on the robot after the end of the
project, given that the scouting robot will need additional upgrades due to its dynamic nature
and technological progress involved.
The dissemination activities will be orchestrated by the IPR & Dissemination Manager, the
Technical Manager, and AVAN (leader of WP5). The leader of WP5 jointly with the IPR and
Results will be disseminated from the very first moment they become available, and the idea is
to attend major annual conferences, even if only preliminary results are obtained. However, the
protection of patentable material may delay the publication of interesting data. In that respect,
the IPR & Dissemination Manager will have to deal with SMEs and find the optimum trade-off.
The website will be the main vehicle for sharing and disseminating information on the
VINEROBOT project, starting from M1. Additional dissemination will be carried out by some
social network accounts devoted to the VINEROBOT project (Twitter and/or Facebook). These
contents will also be accessible from the project website which will also have a private section
only for partners and a public section where papers and news will be posted in a timely manner.
Demos will be announced beforehand and reports/videos of past demos will be accessible.
After the end of the project, the VINEROBOT website will be operative at least for three years
and repository of open source code will be available.
The plan for disseminating the project results will follow two stages. The first stage will be to
initially achieve broad coverage of the project scope to a wide range of European vine growers
and wineries, viticulture professionals, researchers, and users, with activities likely to include:
• Leaflets about the VINEROBOT project in English and the other four languages of the
VINEROBOT partners (Spanish, German, Italian and French)
• Publication of non-confidential information in learned journals, such as: Australian
Journal of Grape and Wine Research, Vitis, Journal International des Sciences de la
Vigne et du Vin, American Journal of Enology and Viticulture, Sensors, Computers and
Electronics in Agriculture, etc… These are targeted to scientific community worldwide.
Expectation 1-2/year, and led by RTDs.
• Publication of the broad results of the work in the trade press of different EU countries.
• Publication of the broad results of the work in the electronic press, such as Vignevini.
• Presentations at conferences, including the most important meetings in Viticulture and
Agricultural machinery. These are targeted to scientific community worldwide.
Expectation 1-2/year, and led by RTDs.
• Demonstrations in the field (see WP2, Task 2.6 and WP4, Task 4.6) and prototypes at
exhibitions including VINITECH, ENOMAQ, VINEXPO. These are targeted to scientific and
potential industry end-users. Expectations 1-2/year, with strong participation of SMEs.
• Transfer of the non-confidential results to students at universities.
Each partner spreads preliminary results and information about the project in
national conferences, divulgation (non-technical) journals, and tradeshows,
generally in the local language to reach growers, industry, and academic
institutions.
Large-scope dissemination in Europe. This will be through international
conferences EuroAgEng (biannually) and the ECPA (European Conference of
Precision Agriculture). Both conferences require English. When possible and
appropriate, a special session will be requested to disseminate the results of
this project (typically sharing the session with other European projects).
Large-scope dissemination activities in America and Japan. This will be carried
out either in the ASABE International Annual Meeting or in specific precision
agriculture conferences organized by ISPA or ASABE.
*
ASABE stands for American Society of Agricultural and Biological Engineers
(http://www.asabe.org/).
Trips to ASABE will constitute a key dissemination activity to present the VINEROBOT Project
and prototype worldwide. This may have an impact on the final number of robots sold to
universities and research centers. An estimated number of increased sells of units is 10 units.
Considering a revenue of 2000€ per unit, this would make 20000 €.
Publication efforts will normally be initiated and guided by the leaders of WPs. The IPR and
Dissemination Manager will assure consistency in all publications, from highly scientific
communications to press releases. General articles on the progress or results of the project may
be motivated and led by one of the Operational Managers. All project publications, including
the web site, will contain the following statement: "The work leading to these results has
received funding from the European Union under grant agreement n° 610953". This will be
reinforced and routinely checked by the Management Board. A list of project publications
searchable under various criteria of interest will be maintained as well as a list of papers written
outside the consortium that cite the VINEROBOT project. Moreover, web statistics, including
number of hits and in-links over time will be kept. These will be assured by the leader of WP5,
AVAN.
The IPR & Dissemination Manager will keep in contact with other FP7 projects working either in
the field of robotics or viticulture. Courtesy visits will be organized for other FP7 teams
interested in this project and potential collaborations will be established if deemed appropriate
for both projects.
Annex 1 – Description of Work – Part B Page 66 of 70
Version N°3 agreed with the EC services on 28-06-2013
FP7-ICT-2013-10-610953
"The VINEROBOT project commits to participate to meetings and events as requested by the
Commission and assigns a budget of 10.000 EUR for this purpose."
The anticipated results of this project are: i) novel sensing procedures and methods; ii) an
autonomous ground wheel-based robot; and iii) smart and optimized IT-based vineyard
management procedures.
Once the detailed benefits and limitations of the autonomous vine robot equipped with several
non-invasive sensing technologies to monitor a vineyard have become clear, a market study will
be undertaken to identify target applications and a detailed exploitation plan will be drawn up.
It will likely follow the format of the Exploitation Strategy Seminars offered by the European
Commission in order to assist consortia in optimizing the exploitation of results. The
Coordinator (and some other partners) already has a positive experience on this mechanism,
where an independent facilitator leads discussions using a procedure especially developed for
collaborative research projects. This process identifies, reviews, and characterizes the project
results (expected and achieved), analyses the Intellectual Property Rights situation, assesses the
non-technical risks, and prepares an action plan to enhance exploitation while reducing risks.
We have found that these seminars offer a clear approach to dealing with exploitation issues,
and we have seen that they enhance improvements to exploitation strategies and future
commercial plans.
The plan proposed will promote dissemination and exploitation, creating a framework for
establishing and developing a co-operation agreement. In particular, it will consider:
The partners will exploit the results of the project by selling the scouting robot including all the
acquisition devices. The partners also have a strong interest in developing incomes from
licensing and royalties from companies outside the current partnership. The main route to
market the scouting robot is through each of the industrial project partners according to their
sales departments, distributors, commercial agents, etc.
To succeed with the commercial plans, we will undertake a programme of market analysis &
stimulation to get ready for a direct business development after the project, including:
As initially proposed, AVAN will lead the exploitation activities in the project. They will mediate
with the General Assembly and individual partners’ commercial departments to identify and
notify the partners of possible opportunities and difficulties in exploiting the project findings.
With their guidance, the General Assembly will:
• Develop a use-case agricultural robot under the scope of Unmanned Ground Vehicles.
It is envisaged that the exploitation activities of the General Assembly (as well as those of the
individual partners) will continue beyond completion of the project.
The Consortium collaboration agreement, which will be based on the DESCA FP7 Model
Consortium Agreement template, will be used to define the mechanism for protecting the
results of the work and defining intellectual property ownership, use of background
information, confidentiality, patents, and the terms and conditions under which licensing can
take place.
At this proposal stage, it is anticipated that each partner will own the intellectual property
generated by it within the project, and will grant the other partners free use of that intellectual
property for purposes of the project and for exploitation. It is anticipated that each Partner will
be entitled to exploit all the foreground information and that they will also grant other partners
access rights, on favourable terms, to any background necessary for the exploitation of
foreground.
We will use the DESCA model agreement to identify a positive list of Background IP to which
other partners have access rights, along with a negative list of Background IP which is
specifically excluded from the obligation to grant access rights. At this stage it is anticipated that
access rights to foreground, if needed for use of a partner's own foreground, shall be granted on
fair and reasonable conditions, whilst access rights for internal research activities shall be
granted on a royalty-free basis. It expected that the partners will negotiate any additional
access rights to foreground on financial conditions to be agreed.
In addition, we expect that the main innovations in the project will be protected by means of
patents, as it is our full intention to offer the technology for licensing to ensure that it has the
broadest possible uptake and to satisfy the demand from companies currently outside the
partnership.
References cited
Fereres, E; Soriano, MA. 2007. Deficit irrigation for reducing agricultural water use. Journal of
Experimental Botany, 58 (2): 147-159.
OIV 2012. Global economic vitiviniculture data for 2011: stable wine production despite the fall
in the world vineyard.
Schultz, H.R. 2000. Climate change and viticulture: A European perspective on climatology,
carbon dioxide and UV-B effects. Australian Journal of Grape and Wine Research, 6, 2-12.
Zoomvino 2012.Website www.zoomvino.es provided by Nabu Media Ltd.