Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

IET Intelligent Transport Systems

Review Article

Crowd intelligence for sustainable futuristic ISSN 1751-956X


Received on 23rd May 2019
Revised 2nd January 2020
intelligent transportation system: a review Accepted on 13th March 2020
E-First on 8th April 2020
doi: 10.1049/iet-its.2019.0321
www.ietdl.org

Rathin Chandra Shit1


1Department of Computer Science & Engineering, International Institute of Information Technology, Bhubaneswar, India
E-mail: rathin088@gmail.com

Abstract: Connected vehicles and fully automated driving systems are the main objectives of the future transportation system.
A safe interactive system that interacts with people and things is essential to achieve these objectives. In this context, a crowd
intelligence system plays a key role in interactive system development. Crowd intelligence is a combined method of data
collection, integration and analysis from devices such as the smartphones, wearables, vehicles and a wide range of Internet of
Things applications to use them as sensors. This collective feedback-driven interactive method is opportunistic for the
development of the future transportation system. In this study, a survey is conducted considering crowd-intelligence techniques
for the transportation system. From this survey, various challenges of the intelligent transportation system have been outlined
and crowd-intelligent solutions have been discussed. A layered structure of transportation system architecture is suggested
considering various problems in each layer and its crowd-intelligent solutions. The crowd-intelligence-based mobility, traffic
control, traffic prediction, parking solutions have been discussed in this survey. Moreover, the importance of crowd-intelligent
techniques and its applicability is discussed for sustainable development of futuristic transport infrastructure.

1 Introduction is adopted by the transport system with the help of collaboration,


communication and transport state of the vehicle which make it
The transportation system evolves into an intelligent, efficient, cost-effective and sustainable. The transport state information of a
reliable and sustainable system by using smart sensing, vehicle is enhanced with precise sensor and crowdsourcing
communication and computing techniques. The traits that specify a techniques. Further, the information mining from this vehicle state
transportation system as an intelligent system are: on-demand is conducted with artificial intelligence techniques to form a crowd
services (i.e. lower cost per transportation), sharing of intelligent system. This crowd intelligence-based transportation
infrastructure (vehicle sharing, public transport), autonomy system helps in solving the underused resources problems of the
(automatic traffic/vehicle control, smart parking system), transportation infrastructure [4]. The crowd intelligence-based
environment friendly (electric vehicle, CO2 emission management transportation system has been deployed in modern goods delivery,
and fuel management). On-demand transportation model is also urban logistics, car sharing [5]. This approach is sustainable to
called ride-hailing or ride-sourcing which is a new mobility model resource consumption, transport safety and accident prevention [6].
where private individuals may sell their rides for eager customers. The sharing economy uses efficient management of the
This new pooled service model is launched by transportation vehicular resources. The vehicles process huge computing and
services such as Uber, Ola, Lyft. These services claim that over communication tasks to use physical resources efficiently. These
millions of vehicles can be reduced from the city streets by using computing tasks are further improved by sharing the computation
this model. This sharing method can reduce the combined task with the proximity of the other vehicles. Redondi et al. [7]
transportation length and cost [1]. proposed a task-sharing method for camera-based visual sensor
The transportation map represents the graphical view of the network. This method shares the image feature computation task
mobility data pattern of a particular region [2]. These data are for faster computation of real-time monitoring which is useful for
collected from crowdsourcing vehicles. These maps guide the smart city applications. Few task-sharing-based traffic management
individuals to plan and share the trip, which reduces the cost per systems have been proposed in the literature [8]. These are useful
ride and congestion. The public and private vehicles data-sourcing for the development of safe, secure and sustainable urban transport
provides services like bus presence detection, bus scheduling, easy system. This computational task-sharing method forms edge
payment method and booking of seats. These services can help computing-based architecture for faster processing where the
passengers to design their rides with more assurance. The visibility selected edge device processes the computation tasks. However,
of the entire path and expenditure prior to an actual journey helps the computational limitations of the system can be improved with
in better planning. These methods reduce the congestion on the fog infrastructure by using the whole system processing [9]. The
road and the cost per ride of the passengers. Integration and sharing main advantage of the fog infrastructure is its location awareness.
of information from different transport methods are necessary for This location awareness facilitates the task sharing to the nearby
this approach. The Society of Automotive Engineers developed an devices/vehicles by reducing communication delay. Several
architecture description language to represent essential components intelligent vehicular networks are suggested based on this fog
for modelling the transport map in the cyber-physical system [3]. infrastructure for a monitoring system in transport and smart city
This architecture uses the concepts of the state machine, security [10, 11].
and privacy. Further, this architecture evolves using fifth- The level of improvement of autonomy is important for
generation vehicle connectivity and artificial intelligence to form a transportation services. The artificial intelligence and machine
cooperative intelligent transportation system. learning techniques when combined with crowdsourced data
The technological advancement helps in connecting the demand produce services such as real-time traffic monitoring [12, 13],
and supply sides of business models. This allows the users to traffic prediction, travel time prediction [14, 15], travel activity
access the goods and services of other user if available. This tracking [16, 17]. These services combinedly form another level of
phenomenon is called the sharing economy. The sharing economy the autonomy in the transportation system. Realising autonomy the

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 480
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
major issue is to schedule the service requests from the subsystems. inadequate for the current transportation system. Implementing
These requests are further optimised for the maximum profit of all sensor techniques ensures to address these challenges by tracking
the users. A point-to-point-based public transport system is the anomalous events and by reducing the serious issues of the
proposed in [18] for the autonomy of services which serves system [37]. In this context, the Internet of Things (IoT) can
maximum people within the minimum time frame. influence on vehicle maintenance and fault detection followed by
The driver assistance system represents another level of the risk analysis of the vehicle. Recent crowdsourcing-based
autonomy in the transportation system. This assistive system uses emergency and crisis management systems ensure the safety of the
features from the environment data and previous driver behaviour systems. The emerging software-based transport systems are prone
to predict real-time driver assistance [19]. The time, travel mode to accidents [38]. These accidents damage automaker's reputations
and service availability are the various features used by the system and consumer trust for this new platform [39] embedded with
to recommend assistance to the driver. Pedestrian walking large-scale computer system [40]. These accidents show that the
behaviour is also an important feature to improve the assistive safety of the transport system is expensive. These safety measures
system. In a smart city scenario, certain places like cultural further reduce the performance, increase the weight and complexity
buildings, malls and nearby areas have more pedestrian than other in the existing system. Hence, the safety measure should be
scenarios. There are numerous factors like weather, land area, time, designed in the early stages of development of the system [41].
pedestrian volume can affect pedestrian behaviour. The driver The above-mentioned future transport system challenges can be
assistance system becomes more robust and autonomous by overcome by using improved data perception methods. The
addition of these features. Rathod and Khot [20] proposed an crowdsensing is a technique which passively collects data through
Radio-Frequency IDentification (RFID) tag-based driver assistance sensing hardware and autonomously supplies the data through
system. The computer vision-based techniques for driver assistance networking to achieve crowdsourcing task. The crowdsensing is
are proposed by Leone et al. [21]. The authors claimed that the the second generation crowdsourcing method which is also called
visual analytic tools have more information to build the mobility social sensing or participatory sensing [42]. This method not only
pattern of the city [22] which is essential for the driver assistance collects and analyses the data but also mines the hidden
system. These assistive systems are improved with smart traffic information from the data such as group behaviour and network
control [12], congestion control [13] and computer vision-based evolution. The crowdsensing method improves the environmental
background subtraction [23, 24] to form a complete route guidance sensing by using IoT and intelligent networking involving anyone
system. The computer vision-based approaches are more complex in the sensing process. The process of crowdsensing generates
than RFID-based approach. Although several driver assistance and huge data and creates information overload problem. Finding the
route guidance systems have been proposed in the literature, yet relationship of sensed data using artificial intelligence techniques
unexpected events occur such as accidents and disruptions. Hence, reveals the hidden information also called crowd intelligence
an automated recovery system is essential to stabilise the transport sensing, which is a potential candidate for solution of intelligent
system for autonomy. Rescheduling and disruption management transport system problems.
services can act as an automated recovery to maintain stability and This paper aims to outline the current development of the
autonomy of the system. transport system, its challenges and future research scope. In this
Parking is a challenge in the transport system. The vehicles survey, the recent development along with its applications has been
searching for parking space add congestion to the transport system. discussed from the recent literature. A crowd-intelligence-based
Hence, to reduce congestion and improve mobility a smart parking perspective for future transport system is discussed. This provides
system is essential. There are two approaches to smart parking a solid overview of the researchers and practitioners in this field.
system design based upon the use of sensor and camera. The The contributions from this survey are summarised as follows:
sensor-based approach searches the proximity of the vehicle and is
mainly used in parking guidance system [25, 26]. The main (i) The paper explores recent transportation system and its
demerit of this system is high installation and maintenance cost. challenges from the literature.
The camera-based system uses image processing method [27] (ii) The concept of crowd intelligence and its applications in
which implements either object detection method or occupancy different transport problems have explained.
status detection [28] with image background subtraction. Machine (iii) The role of crowd intelligence for smart transport system
learning algorithms for occupancy detection [29] are used by architecture development is presented.
feature extraction followed by classification. Classification (iv) Traffic/intersection management and mobility pattern analysis
methods like support vector machine (SVM) and deep learning using crowd-intelligence techniques from the literature are
(DL) are also reported in the literature [30]. Another approach is discussed.
based on the crowdsourcing [31] which relies on location service (v) Mobility as a service (MaaS) and impact of crowd intelligence
from a smartphone that estimates the parking status by activity is surveyed for future transport infrastructure.
recognition such as arrival and departure detection [32].
Fuel efficiency and environmentally friendly vehicle designing The remaining of the paper is organised as follows: Section 2 and
is a challenge for the transportation system. In this context, the its subsection explain the crowd intelligence and its use in the
electric vehicle supports urban pollution control and power smart transport system. Section 3 delves into the importance of
management [33, 34]. These electric vehicles cooperatively crowd intelligence in the transportation system model. Sections 4
communicate and form vehicle to grid structure for optimising and 5 survey the crowd intelligence in the service/application layer
power requirement along with the transportation functionalities. and vehicular cloud/fog layer, respectively. Section 6 presents
The electrical vehicular infrastructure can be possible with the methods and future direction for crowd-intelligence-based
integration of electrical charging facilities in the transportation transportation system. Finally, a conclusion is presented in Section
system. The current electric vehicle system takes enormous 7.
charging time to travel a very short distance [35]. These issues
challenge for further research for designing better charging
techniques and charging point deployment in the transportation 2 Crowd intelligence and properties of crowd-
infrastructure. In this context, Al Junaibi and Farid [36] proposed intelligent system
the sensed vehicular mobility data-based charging point Crowd intelligence is the collective intelligence that arises from
deployment method which can be further improved by artificial cooperation, coordination and cognition for decision making to
intelligence techniques. The charging infrastructure for electric improve the functionality of the transport services. This concept is
vehicles and their management can make the electric transportation applied in crowdsourcing applications. This technique also
system attainable. involves social network intelligence for mass activities. Crowd
Autonomy and safety are the major challenges for the future intelligence uses artificial intelligence techniques to extract
transportation system. The transport system is getting more knowledge from crowd-sensed data for better decision making and
complex to address these challenges. The system safety is recommendation. It also benefits the knowledge from the crowd

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 481
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 1 Crowd intelligence and its properties

instead of the individual. The modern information and topologies. However, to solve data inconsistency and redundancy
communication technology enhances the collective knowledge by in the vehicular network with social interaction has trust issues for
using both human and sensor information. The concept of crowd sharing of important information. Hence, a trust model that uses
intelligence is presented diagrammatically in Fig. 1. past interactions of the vehicle to deduce social trust among
Crowd intelligence follows three stages: first, the data is vehicles [52] for these networks is necessary. A trustworthy
acquired by various heterogeneous sensors from vehicles, devices, parking space recommendation for a community of vehicles is
buildings and human activities; second, the huge acquired data is proposed by Timpner et al. [53] and further a trustworthy social
converted into a standard form; and finally, the analysis of these vehicular framework model is presented in [54] which defines the
data is carried out by using artificial intelligence techniques for a principles of trust and control.
specific application. The crowd-intelligence architecture comprises
three components such as cooperation, coordination and prediction 2.2 Missing data maintenance
as shown in Fig. 1. Furthermore, the properties and objectives of
crowd intelligence are outlined in Fig. 1.The challenges and issues The traffic data collection methods are developed gradually over
of the crowd-intelligence methods are discussed as follows. time. Initially, the loop detectors are used as sensors but with time
other sensing methods have been used such as GPS, smartphones,
2.1 Sensing data quality maintenance vehicle probes. The recent sensor technologies have changed the
data collection methods [55, 56]. These data collection methods
Crowdsource information can address the road navigation problem provide missing data which is a major issue in data collection. The
in cities [43]. Although the route from the map-based services such missing data pattern in crowdsourced data is categorised as missing
as Google maps can provide the best route but this cannot provide at random (MAR) or not MAR (NMAR) [57]. In the MAR model,
actual road condition, weather and traffic parameters. Hence, the missing data is independent of all other unobserved data
crowd-intelligence-based navigation by using the actual driver's whereas in the NMAR model the missing data is related to some
feedback can make a better navigation system design. Crowd unobserved information. Hence, NMAR models can represent
intelligence uses crowdsourced data from the users with social more challenging scenarios than MAR models. The minimisation
network information to build a vehicular social network (VSN) of potential bias is observed by incorporation of information to
[44]. Cisco reported that 300 million passenger vehicles generate describe the difference between missing and observed data to form
400 million GB of data for wireless communication [45]. To a missing data pattern [57]. In real-time traffic scenarios, the
extract the important information from this huge amount of data is missing data observation probability depends upon traffic states
a challenging task. Hence, the use of VSN information such as the such as traffic volume or speed [58].
relationship between the vehicles and interest of passengers can The data missing problem is addressed with regression
further extract the important information in the vehicular cloud approaches such as linear regression, SVM [59], K-Nearest
from such huge data. Wan et al. [46] developed a context-aware Neighbors (KNN) like parametric methods [60] and other methods
dynamic parking method by using these social vehicular interaction [61] but these methods face the challenge of complex data dynamic
information. modelling. The expectation maximisation methods [62, 63] are also
Growing smartphones usage can help to get the crowd sensed suggested in the literature. Matrix and tensor completion methods
data for a better recommendation of the vehicular network such as [64–66] are also suggested in the literature which promise to give
traffic management [47], dynamic route selection [48]. However, computationally efficient and accurate results than regression
this approach of crowdsensing faces challenges like data methods. Furthermore, the dimensional reduction method [67] for
redundancy and inconsistency [49]. The moving vehicles have traffic data is applicable but faces challenges of temporal
different speed and it changes the topology of the vehicular dependencies modelling [67]. Neural network and DL-based
network. Vehicular communication and social interaction-based methods [68–71] are recommended in the literature for high-
systems have been proposed to extract information from these dimensional complex traffic data. The Gaussian process [72]
dynamic topologies. Distributed incentive [50] and game-theoretic method is used to model complex traffic patterns such as traffic
models [51] have shown promising results for such dynamic
482 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 2 Smart transportation layers and crowd intelligence

congestion [73], travel time [74], transport flow, traffic volume of the individuals. Related to this issue several privacy measures
[75] as the stochastic process to address traffic data missing. have been suggested in the literature. Differential privacy is a
standard method where the data owner controls the sharable part of
2.3 Crowdsource data authentication the data [83]. Hence, the individual participant's dataset is
independent of the rest. The privacy and accuracy are trade-offs in
The major issue of the crowdsourced data is its authenticity. There this differential privacy mechanisms where privacy is maintained at
are spoofed data attacks in the internet cloud and vehicular cloud the cost of accuracy. A privacy method that is improved with
[76]. The malicious agents force the traffic flow structure by constant accuracy often called scalable privacy [84]. In this
creating phantom vehicles which increase congestion. Hence both approach, each data owner is hidden by using uncertain data
the cloud should work together for these kinds of spoofed message distribution. Allocation of incentive services in crowdsource-based
attacks. The real issue is to detect these data in the safety-critical applications is another issue in maintaining privacy with uncertain
message within the smallest time frame. Collaborative validation data distribution [85]. The participants who contribute to crowd
and local consensus achievement within the cluster are the solution data pool by using phone may expose the path they travelled.
for this problem [77]. In this approach first, the sensor data is Using anonymity can be a solution, then it is difficult to provide
collected and processed to generate keys. The computed keys are the reward for their contribution. Hence, it is an open issue for
shared with neighbours and the signed data is uploaded to edge privacy-preserving mechanisms for these crowdsourcing
cloud. The surrounding vehicles then broadcast the signature to participants.
verify and download signed data from edge cloud. Finally, the
neighbour vehicle's data is validated against local records. In this
approach, the vehicular cloud satisfies the resource for computing
3 Transportation system model and crowd
needs [78]. intelligence
Public transport system faces huge loss because of free-riding The transport system is enhanced by the recent development of
problem. Millions of financial losses are estimated over the world communication, computing and sensor technologies. The smart
in public transport with serious damage in terms of revenue [79, transportation system comprises a smart vehicle, vehicular cloud
80]. Further, this issue undermines honest passengers who pay the platform and application layer are shown in Fig. 2. The smart
high increasing fare. One solution to this problem is to increase the vehicle has the important characteristics of a certain level of
fine for free riding [79]. Hence, efficient strategies are required for autonomy and electrification. The goal of autonomy is to make
the combat of free rider's problem. Manual method [81] by using driverless vehicle and an automatic parking system. The higher
the manual implementation of team members for checking this level of autonomy is still under an open research problem. The
invasion spends huge expense and lacks proper control. These vehicular electrification with electric cars and magnetic trains is the
problems of the public transport system can be addressed by future of fuel-efficient transport system. Although this is
crowdsourced data authentication followed by crowd-intelligence environment friendly but inefficient for a long-distance journey.
techniques [82]. Hence, the electric outlet system is essential for recharging the
vehicles in the city for supporting long-distance journey. The
2.4 Scalable privacy vehicular cloud accesses the vehicles, processes the sensed data
and builds the application for users. This layer provides the
The privacy of individuals in the cooperative transportation system framework for vehicular crowd-intelligence services. The
is an important issue. Crowd-intelligence approach should use application layer uses vehicular cloud the framework and provides
crowdsourcing data by restricting the user's details for the privacy on-demand mobility and shared economy-based services. The on-

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 483
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
demand mobility is the main objective of MaaS platform. This options by using the preferences such as time, convenience, cost
approach provides demand-based costing services. The shared and payment in the real time. This paradigm shifts from centralised
economy-based applications are the bike-share, car-share, peer-to- to user-oriented mobility model. The fuzzy nature of the MaaS
peer rental of the vehicles. Security is a concern in all the layers of paradigm makes it difficult to address the problems such mode's
this transportation framework. The futuristic transport system uses identification for service, the real-time information, trip planning,
crowd-intelligence techniques in vehicular cloud and application ticketing, booking by diverse technologies such as GPS, E-ticket,
layers of the transportation system. The main objective of the E-payment. Trip disruption warning and synchronisation are
crowd-intelligence system in transportation is to provide intelligent different emergency services. Jittrapirom et al. [91] reviewed the
vehicle routing and intersection/traffic management. MaaS transport concept and defined the attributes to describe the
The algorithms like Dijkstra, Bellman–Ford are the first selected scheme and application. Further, the authors analyse the
algorithms that are used to address route queries and optimal route service considering travel demand modelling and a supply-side
determination problems. Other meta-heuristic and greedy-based analysis.
approaches are also present for the successful prediction of the Most of the existing transportation systems are unstructured
optimal path. However, these algorithms are slow in query which affect different transport operators and transportation
processing to be applicable in the real-time road network. In the market. There are several proposals for integrating multimodal
case of the real-time road network, the information about road transport and development of MaaS. The MaaS model is analogous
conditions can help to pre-process the optimum path to cloud computing which follows everything-is-a-service
recommendation. Besides road conditions the topological paradigm. A similar idea such as the system of systems to that of
properties of the network can refine the route estimation. The MaaS model is present in the literature [92, 93]. However, the
public transportation is different and time dependant as compared concept of MaaS is materialised by Callegati et al. [94]. In the
to normal road transport i.e. the time table of the transportation is MaaS model there are mobility resources to discover, plan, book
used for better planning and routing. The problem of time and guide for rides by any mode of transportation. Callegati et al.
dependency can be solved with speed-up methods such as [95] pointed out that a single mobility operator is unreliable and
multimodal and multi-criteria planning and routing. The public also suggested a platform for MaaS operator which automatises
transportation can be enhanced by considering weather, service mobility service among many providers. The authors validate their
delay, breakdown, uncertainties and interruptions. This should be proposal with a prototype and various micro-services. A
considered in designing time and modelling a multimodal network. comparative discussion of crowd-intelligence-based algorithms is
Traffic prediction plays an important role in urban mobility. shown in Table 1.
There are two types of traffic prediction such as short term and It is important to optimise the transportation and traffic for
long term. The short term predicts the upcoming precise future of saving the travel time of individuals. The real-time traffic
the vehicle. However, the long term predicts the traffic states of the management systems [102] have been suggested for optimising
city for a long time period. The time-series prediction algorithms traffic of different scenarios. The variables that are considered for
consider features such as traffic flow in vehicles per hour and optimised path recommendations are traffic volume, speed, traffic
boundary conditions of the road. Other features such as average flow, travel time and stochastic shortest route. To improve path
speed and occupancy for prediction are also suggested in the recommendation various data mining approaches have also been
literature [86]. This modelling works successfully in the normal suggested in the literature [103]. The crowdsourced data promise to
condition of the transport system. When there is a change in system recommend the best path since the data is collected by the
parameters such as a change in the weather, these time-series satisfaction and necessity levels of the individuals. Senanayake and
algorithms fail. For such scenarios, other seasonal algorithms have Wijayanayake [96] developed an algorithm for transportation path
been proposed by Kumar and Vanajakshi [87]. Artificial prediction by using crowdsourced data. This method predicts an
intelligence and machine learning-based approaches have also been optimised transportation path and time slot while an user plans a
suggested to model the traffic states. However, these are journey. The authors validated the proposed algorithm in the traffic
computationally complex but give a faster prediction. Most of the scenarios of Sri Lankan cities. Vehicular cloud and IoT-based
algorithms are presented for short-term prediction and do not methods for optimised path prediction have been suggested for
consider the other variables like traffic accidents, weather issues smart city context. Carnevale et al. [97] analysed the different
[88]. Hence, for proper traffic analysis, the long-term prediction performance indicators for design and development of mobility
and data analysis are inevitable. service in smart transportation context. The authors have studied
The data collected from the sensors of urban vehicular the mobility pattern of the University of Messina personnel. They
infrastructures such as vehicles, devices, humans and buildings are have suggested that the traffic congestion, fuel consumption and
heterogeneous. Crowdsourcing is the method of acquisition of this CO2 emission can be regulated by controlling the travelling habits
huge data followed by integration and analysis of the data for a of the individuals.
certain application. The urban planning is more dynamic and The mobile phones can be used as a crowd-sensing device since
participatory for the transportation system. The accurate mobility it is embedded with different navigation sensors. The traffic
prediction in urban scenarios requires detailed data which is a behaviour and navigation are easily detected with the various
major challenge. Most of the existing crowdsourcing approaches crowd-sensing devices. Nickerson et al. [104] categorised the
use the internet for the central collection of data, so crowdsourcing navigation services into trip planning, car sharing, navigation,
without internet is not possible. The data ownership and privacy logistics, parking and location-based services. Scherieck et al.
are the major concerns for digital crowdsourcing data. Data [101] developed an integrated mobility service with an open and
extraction and mining the proper semantics from the noisy modular platform. Further, they designed an architecture by taking
crowdsource data are the fundamental issues [89]. The proper mobility data to standardise the service. Andrea et al. [94]
business model for inter-modal mobility may address privacy classified the potential threats of MaaS model and proposed a
challenges. solution of its mitigation in the transport system scenario. The
authors follow a structure from individual operator to federated
4 Crowd intelligence in service/application layer MaaS providers. Further, they aimed at harmonising data flow and
invocation by developing a federated service. They found that
The main aim of the service/application layer is to develop these threats are similar and also present in the fields such as cloud
techniques concerning the user's mobility without using personal computing, supply chain management, trusted business partnership
vehicles. Goodall et al. [90] analyses the service layer paradigm, and service architecture. However, the effectiveness of the
core elements, concepts, evolution and suggested MaaS model for proposed architecture is not validated experimentally. In a real-
transportation. The MaaS paradigm changes the traditional method world scenario, the validation and future deployment may make
of locating, booking and paying for different modes of transport to MaaS solution immune to threats.
let users plan and book door-to-door trips by using a single Xia et al. [99] investigated a large-scale mobility scenario for
application. In the MaaS model, the users exploit all the network an energy-efficient transportation system that also called green

484 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
transportation. The authors generated trips with unconstrained mobility, green transport, smart parking and smart delivery system.
gravity model and weighted Dijkstra algorithm by taking traffic Suciu et al. [98] presented a shared economy-based platform for
and road information. This trip generation algorithm is based on monitoring smart city mobility by using crowdsourcing to choose
weighted segments of the road structure. The authors used the open traffic data and collaborative development of smart transport
different mobility patterns such as human mobility model along applications. In this approach, the data is collected from
with the origin–destination data [105] for different scenarios by heterogeneous and distributed devices such as sensors and
their unique probability distribution. The macroscopic attributes of actuators. Further, this data is analysed, controlled and managed by
the mobility traces are used further for VSNs. The green distributed services.
transportation requires expensive infrastructure change of the smart
cities. Andersen et al. [100] proposed an alternative solution of 5 Crowd intelligence in vehicular cloud/fog layer
smart nudging for the green transportation system. The concept of
nudging is to influence the decision or behaviour by suggestion and The vehicular cloud layer is the central control of a transport
reinforcements. In smart transportation scenarios, it is used to system. Mobility and traffic management are the problem area of
select the energy-efficient mode for green transport. The authors this layer. Using crowd intelligence for these problems is addressed
proposed the method as sense, analyse, inform and nudge. With in the literature. In Tables 2 and 3 the mobility and traffic
sense, analysis and information the system can make smart nudge. management-based crowd-intelligence algorithms are categorically
This approach helps people in making decisions, which is best for presented, respectively.
green transport by using public vehicle, real-time traffic and
pollution control [106]. This nudge's primary aim is to help 5.1 Crowd intelligence for smart mobility
people's behaviour towards a better society by healthy, green
transport and long-term interest of society. Crowdsourcing is a potential solution for urban mobility. The
Sharing of transport resources can solve transport mobility crowdsourcing application creates personalised mobility maps
problems. In this context, Carnevale et al. [97] developed a shared from the crowdsensing data. This mobility map is a guidance
economy to address problems such as transport monitoring, shared system for traffic and congestion control. Agarwal et al. [42]
proposed a crowd sense-based urban mobility model and tested it

Table 1 Used cases of crowd intelligence in service/application layer


Research Proposed concept Crowdsource Transportation Service Application Remark
applicability mode category
[90] reviewed integration crowdsource is used integration of ride/car future autonomous public private partnership and
of modes for data collection services sharing drive and dynamic integration of transportation
routing service can form MaaS
platforms
[96] transportation route crowdsource-based mode selection from location-based location aware crowdsourcing and Data-
recommendation mobility data collection all mode information transportation mining is used for best
mode selection transportation mode
recommendation
[95] MaaS stack uses crowdsource all modes trip planner SMAll (smart a micro-service paradigm is
operator. mobility for all) presented which is under
gradual development and
integration according to
dynamic demands
[97] GE (generic crowd data of home sharing of modes smart mobility minimise traffic complement of private and
enablers) smart office mobility data is services congestion, fuel public transport for
crowdsourcing used to design mobility consumption and multimodal sharing or pooling
services CO2 emission transport can improve the
transport
[91] supply side analysis, use of crowdsource sharing of mode and demand MaaS architecture different attribute for MaaS
travel demand data for development services responsive on- design are defined and challenges
modelling and of MaaS model demand are explained
business model services
design
[98] mobility simulation crowdsourced data different modes of living lab of integrated, transportation model is
platforms from city and maps are communication of city data to test durable, efficient developed for mobility plan of
collected city is simulated different and safe the city ecosystem
services transportation
system.
[99] VSN model extensive observation trip generation from urban mobility urban green the author generated mobility
and correction of mobility transportation model for context aware
OpenStreetMap with system application of VSN
crowdsensing data
[100] green transportation sensor and crowd- environmental situational smart nudging the combination of sense
choice with IoT based sensing is friendly aware analyse and nudging help to
suitable transportation option information choose the best
transportation option
available
[101] sustainable digital open mobility using integrated mobility open digital modular mobility integrate all mobility services
mobility service crowdsource sensing service mobility service cover all to form open and modular
apps mobility model
[94] threats in MaaS crowdsourcing attacks sabotage mobility privacy and financial fraud to the author discuss the insider
concepts services security service manager threats as MaaS

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 485
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 2 Comparative analysis of crowd-intelligence-based smart mobility approaches
Work Environment Type of sensor and Crowdsourcing method Mobility application Remark
data
[107] smart city open ended suggestion crowdsourcing + content traffic forecasting application of crowdsourced solution
from users analysis system for transport application in Indian
context is discussed
[108] Florianopolis, Brazil GPS, accelerometer, mobility data analysis city resource collaborative public transportation
City (ParticipActBrazil cellphone sensors allocation using mobility habits of people
Project)
[109] Gratwein-Strabengel smart phone feedback volunteered geographic demand responsive instant feedback on problems and
(Province of Styria, application information analysis transport system incidents for decision makers and
Austria) (emergency reporting, transport manager
citizen science)
[110] Austin, Texas, New user provided data coproductive planning megaregion planning author developed a context-
York, Chicago dependent knowledge of public
engagement for transportation
technologies
[111] smart city Google direction data path safety score analysis safe navigation crime data of the path is analysed for
with crime data prediction of the safest path in city
[112] Shenzhen City mobile phone data, taxi data fusion of multi-source human mobility model combining mobile phone signal data
GPS data, subway data with transportation data to form a
smartcard data human mobility model
[113] smart city business collection of data from VAS sustainable policy visual analytic help business strategy
ecosystem internet sources making and policy making in smart city
[114] city subway route collection of passenger crowdsourcing + dynamic subway route planning passenger data is used to predict
flow, time, week, programming algorithm travel requirement and dynamic route
location and bus plan
information
[115] Mumbai city public smartphone GPS crowdsourcing + filtering real-time public data collected from 14 different bus
transport algorithm transport vehicle route of Mumbai, India and utilised in
positioning public transport
[116] bicycle trips of Sydney smartphone sensor crowdsourcing + validation supporting city suitable use of bicycling mobility
City GPS using geographic planning crowdsourced data for city planning is
coverage, origin- discussed
destination match,
demographic match, route
match
[117] Krakow city web-based crowdsourcing + Data individual mobility the needs of the individuals having
crowdsourcing tool analysis analysis reduced mobility like senior or
physical disability is analysed by the
authors
[118] smart city smartphone sensors crowdsourcing of route route recommendation traffic situation of road as well as the
with driver's personal information + system driver's subjective behaviour is used
information crowdsourcing of driver for better route recommendation
social information
[119] Chandigarh city smartphone sensors crowdsource + dynamic road surface smooth transportation infrastructure
time warping (DTW) based monitoring development by monitoring roads with
time series analysis crowdsource
[120] smart city user check in data such crowdsourcing + data urban mobility the analysis of crowd data for smart
as traffic sensor and analysis using map reduce city application
bike station data model
[121] Kolkata, Bhubaneswar, smartphone data crowdsourcing + Data city traffic map data processing and generation of
Durgapur Cities of Processing travel trajectory map for city traffic
India mapping

among the Brazil citizens. This model further helps in planning for identifying, assessing and selecting internet data source categories
resource allocation and mobility needs of the city. Verma. [121] for visualising business ecosystem of the city also called business
developed a CrowdMap, which consists of intelligent data logging ecosystem explorer. This VAS-based crowdsourced framework
module, smartphone app and a server-side processing to extract bus stimulates knowledge flows in the ecosystem for different city
route information followed by the annotation in the city map. This policies [131, 132]. The authors validate this model with the case
crowdsource-based approach is useful for identification of user, study of the smart city and conclude that crowdsourced-based data
route identification and total trajectory travelled by the user. This collection from internet sources showcase significant part of the
work is an extension of the work [129] with the improvement of smart city mobility model.
city mobility map. Tomaras [120] analysed the heterogeneous The travel demand of a city is modelled with human mobility
urban data to identify the key points for the operation of pattern. Several methods of human mobility models are presented
crowdsourcing technique for sustainable smart city application. in the literature [133, 134]. Mobile phone-based crowdsourcing for
The time efficiency [130] for travel time estimation of the real human mobility pattern is proposed in [135, 136]. Further, a lot of
time bus network is analysed with crowdsourcing. Faber et al. models based on principal component analysis (PCA) [137],
[113] developed a visual analytic system (VAS)-based explore return model [138], diverse scale model [139] and
crowdsourced smart city mobility model. This is based on radiation models [140, 141] are suggested in the literature.
486 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
However, for most of the cities, the data from automobile or destination match, distance–duration distribution and route match.
passengers is not available. Huanga et al. [112] proposed the They prepared the crowd-sensed data by using different
human mobility model by using a single-day mobile phone representations as per the mobility information. This validation
signalling data, 3 months subway smartcard data and taxi GPS process is experimentally demonstrated by the authors in the city of
data. He used these data to estimate the real-time mobility of Sydney, Australia by using bicycling data. They further discussed
humans or the crowding events. crowdsourced data limitation and its future approach towards the
The on-demand shared vehicular infrastructure is still in entire system. Griffin et al. [110] explored the crowdsourced geo-
progress and regulatory initiatives are planned for the cities. information system by using public generated information to form
However, the cities are still facing overcrowding and bad driving a mega-region planning system. The authors validated their
which lead to accidents. Kumar et al. [107] identified various planning system by using the data from the Austin City, Texas and
safety and security components of the cities. They used the bike share system of New York and Chicago.
crowdsourcing techniques for sustainable and efficient The digital information plays an important role in the
transportation systems by using a planned mobility demand model. knowledge-based economy. The authors discussed the major
The authors validated the proposed crowdsourced framework in information gaps in crowdsourcing-based sensing approaches.
Indian cities. de Almeida Buosi et al. [108] developed a Crowd-sensing approach such as the ride report of the individual is
crowdsource-based urban mobility pattern determination method. used for bicycles and vehicles. The authors provided evidence of
In the work, the authors showed how urban mobility analytic tools using crowdsourcing for improved planning of mega-region. This
help in determining the visiting pattern in the urban area. A bike- technology shows promising results, although there is the
sharing system under large city situation context such as requirement of development of biased free participatory
marathons, large social events, closure of the road, the general technology. Dienstl and Scholz [109] proposed a volunteered geo-
condition for deep analysis and understanding is presented in information system-based crowdsourcing approach where direct
[142]. Further, they have developed the relocation algorithm for feedback from the user is taken from the demand–response
sudden changes in urban dynamics. transportation system. The consumer gives data about the
The city resident's mobility and quality of life depend upon the experience of travelling. From that consumer's location, the
sustainable mobility of the city. The process of development of transport route incidents are traced. This feedback method is
sustainable mobility requires crowdsourcing tools [117] for all efficient with the position and time period. The authors
group of individual participation [143–145]. Hoy et al. [117] implemented this scenario in the Province of Styria, Austria. As
proposed the web-based crowdsourced solution for data collection this model is based on smartphone feedback application, there is no
and travel behaviour prediction. The authors analysed the extra hardware implementation cost. Dienstl et al. further claimed
significance of different solutions for sustainable mobility in the that the demand–responsive transportation model can track
city. This web-based solution also covers the mobility of incidents more accurately in the city. Mukheja et al. [115]
individuals with disabilities [146]. The recent improvement of developed a real-time positioning method by using crowdsourced
crowdsourced devices [147] enhances the potential of data positioning data from smartphone GPS. The authors claimed that
collection for a dedicated solution [117] to improve sustainable this location information by using the crowdsourced method can
mobilities [146, 148, 149]. replace the traditional location systems. However, crowdsourced-
The crowdsource-based data can help in the planning and based location information is erroneous, so it can only be used as
development of the city. The mobile phone has sensor capability an improving technique for localisation. The authors experimented
which captures the mobility data of the individual or the intended the algorithm by using 14 different bus routes of Mumbai city,
vehicle in terms of trip pattern. The crowd-sensed data lack in India and presented that this method reduces the average error up
accuracy and representation. Leao et al. [116] developed a to 21.3%. This location information further can be utilised for
validation process of the crowd-sensed data with five principles sharing locations in public transport and telematics application
such as geographic coverage, demographic match, origin–

Table 3 Comparative analysis of crowd-intelligence-based traffic management approaches


Research Environment Types of sensor and Crowdsourcing Traffic Remark
data method management
application
[122] smart smartphone GPS and mobile app based traffic violation information retrieval for decision
environments camera is used to record crowdsensing reporting making. Real-time map of traffic
data violation
[123] Palo Alto City magnetometer data crowdsource + traffic calming moderating or calming of traffic speed
correlation based on psychological or practical
effect of lane speed, lane narrowing,
restricted access and other
interventions
[124] smart city Wi-Fi radio wave crowdsource for real-time traffic collaborative information gathering by
simulation sensors information gathering + condition (traffic crowdsource and reliable traffic
Blockchain for incentive accident/jam) information is generated with
sharing blockchain
[125] New York City GPS traces of city buses crowdsource + DL traffic congestion traffic congestion prediction and
for 3 months architecture mapping mapping with limited data
[126] Chittagong /Dhaka road and vehicle sensor crowdsource + Bayesian road junction traffic real-time traffic condition for road
City Bangladesh data model forecasting intersection is predicted with
crowdsource data with probabilistic
model
[127] Hangzhou City GPS data from the bus crowdsource + anomaly traffic anomaly long-term traffic anomaly is detected for
index algorithm detection Hangzhou city which is used for traffic
planning
[128] prototype smartphone application crowdsource analysis traffic condition Crowdsourcing-based traffic reporting
data reporting is shown better than automatic road
condition detection

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 487
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[150, 151]. Hence, the crowdsourcing approach is said to be The crowdsourcing and mobility trace methods are combined to
helpful for the improvement of the location of the vehicle [152]. build a traffic speed model for well suited robust prediction. Inter-
The sharing economy plays an important role in the paradigm vehicle traffic monitoring and traffic condition are predicted in
shift in transportation. Ride-sharing services such as bicycle [167]. Techniques such as fuzzy logic and neural network are used
sharing, car sharing and shared buses [153] implement this sharing for traffic prediction of smart transport [168, 169]. Khan et al.
economy to make the low cost per ride. Hence, transportation seeks [126] proposed a crowdsourcing-based traffic estimator which uses
the use of intelligence techniques [154]. Public transportation plays the Bayesian inference method. This inference technique is used to
a major role in urban travel [155]. Therefore, the dynamic bus predict the traffic in one junction based on the traffic data in the
routing of public transport affects the entire city traffic condition neighbouring junction. Further, the authors developed and
[127]. Kong et al. [114] proposed a dynamic routing based on the validated the proposal with the web-based prototype.
crowdsourcing model of shared bus data. They further claimed that Traffic management and regulation are becoming the major
this model predicts travel requirements. This method first analyses challenges for smart cities and impact the whole transportation
the resident's travelling behaviours and extracts five features such system. Improper traffic management brings accidents and
as flow, time, week, location followed by the prediction of the congestion in the transport system. One approach is to implement a
travel by using machine learning. Then this method uses a dynamic surveillance system in the traffic area to address this problem but it
programming algorithm to generate optimal dynamic routes. The is expensive for developing countries. Jameela et al. [122]
authors developed a subway shuttle bus system by using the proposed a framework based on the crowd-sensing model to
crowdsourced data and showed that it outperforms other methods overcome this challenge. The authors used crowdsource-based
in terms of dynamic bus routing. There are transport applications violation report for traffic regulation. The crowd-sensing-based
such as Google Maps, Waze which provide live traffic and reports are further evaluated in multiple phases in this system.
navigation information for a path. However, these applications do Firstly, spam is eliminated followed by traffic law expert
not provide information about the physical condition of the road. evaluation of crowd-sensed data. Finally, the user is rewarded with
Singh et al. [119] proposed a smartphone-based crowdsourcing an incentive ensuring anonymity. The blockchain technologies
approach for road surface detection and optimum route have the potential for promising crowdsourcing application
recommendation. In this work, the motion sensors such as development. FixMyStreet [170] is a blockchain-based
accelerometer, GPS, compass are used for sensing the road surface crowdsourcing application used for city road management. Here
condition of potholes and cracks. The authors claimed that the the citizen reports about the broken road and gets incentives. The
method handles speed variation and time deformation associated government finds out the road condition such as unauthorised
with the time-series data. The authors experimentally compared bump and dumping with this approach. In crowdsourcing, the
this method with other methods such as SVM, hidden Markov blockchain-based approach retains the privacy of the user and gives
model and neural network. The authors called it a smart patrolling an incentive with the help of cryptocurrency, making it a good
service based on crowdsourcing in a real-time environment and candidate for the crowdsourcing approach. The crowdsourcing
experimented it in Chandigarh, India. Alabbasi et al. [156] based sharing economy with blockchain expands toward self-
proposed a deep Q network for learning optimal dispatch policies driving and connected vehicle to form a future efficient and safe
for ride-sharing services to reduce congestion of vehicles. transportation system. Fujihara [124] developed a collaborative
Wang et al. [118] introduced a crowdsource-based social aware crowdsourcing approach incentivised by blockchain technique. The
path recommendation system. This system forecasts the optimised author further claimed that this system could work in real-time
route to diminish the driver's fatigue and mood. The authors traffic prediction with a live traffic accident and road segment jam
claimed that the method integrates multiple sensing data from detection.
different sensors of the smartphone, driver's real-time mood with The vehicular traffic problem is moderated with traffic
fatigue and driving history for the recommendation [157]. Further, management strategies such as lane narrowing, deflection, speed
they analysed the effectiveness of this method. There are many management, restricted access. This approach not only moderates
problems addressed in the literature with crowdsourcing such as the traffic but also reduces the accident, noise and waiting time.
the road accidents and traffic congestion [158, 159]. The crowd- The traffic is reduced by applying technologies such as cooperative
sensed data is used to analyse the social internet of vehicle with communication and regulation also called traffic calming. Low-
cooperative quality assurance services and optimisation [160]. A power wide area network such as LoRaWAN [171] is used for low-
music delivery service using crowdsourced data is also formulated cost wireless communication extensively used by these traffic
in [161] where the recommendation system plays situation aware calming approaches. Further, these techniques can be used in smart
music to reduce driver's fatigue and negative emotions. Further, an city scenarios [172, 173]. Iannucci et al. [123] developed a
emotion aware cognitive system [162] is developed based on crowdsensing-based traffic calming approach. The authors claimed
crowdsourced data to deploy different emotion aware mobile that their design is roadworthy, uses low power and supports
applications. Traditional navigation systems find the shortest path economical traffic sensing for a smart city. Further, the authors
between source and destination. However, there is often the conducted the data analysis and communication performance. Dai
requirement of the safest path. Goel et al. [111] developed a robust et al. [174] developed a cooperative game-theoretical model for
safety model by using crime data from the government sources and travel planning to reduce bus clustering or jam.
crowdsource data from the travellers. In this work, they proposed a Traffic anomaly detection is a method of finding an unexpected
safety index for each route and optimise the best route by higher traffic pattern which helps to determine traffic problems. Big traffic
safety index. Further, the authors claimed that this safety measure data made it possible to mine and predict anomalous behaviour of
adapts with time according to the user's feedback scores from traffic in the urban scenario. The PCA method is used for anomaly
crowdsourced data gathered online. The criminal risk and other detection from traffic data [175, 176]. Anomaly detection is further
factors that are affecting the safety on the road are analysed by the improved with the addition of wavelet transformation techniques.
authors. Methods such as the likelihood ratio test [177] and segment-based
outlier detection [178] are also used for traffic anomaly detection.
5.2 Crowd intelligence for traffic management Kong et al. [127] proposed a traffic condition model for calculating
anomaly index by using temporal and special segment of the
The traffic-related problems can be divided into sub-problems such crowdsourced data. The authors experimentally validated the
as traffic speed estimation [163], the traffic jam location and proposed method by using crowdsourced data from Hangzhou city.
prediction [164], the traffic jam minimisation [165] and traffic flow Although several applications have been developed for automatic
modelling [166]. Iyer et al. [125] proposed an approach of road navigational purpose but these do not take feedback from the users.
congestion prediction by using DL architecture. The authors Using huge user base can help to collect traffic condition by user
achieved the goal of predicting the error in the traffic segment feedbacks such as animals on the road or wrong road sign. Hence,
speed of 1 m/s on average. However, with technologies such as the user feedback-based system facilitates the traffic condition
crowdsourcing, the above problems can be addressed effectively. reporting. Simple software-based solution for crowdsourcing can

488 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
help in gathering data from a wide user base. Rantanen et al. [128] crowd-intelligence system. The forward and reverse flow of
proposed a reporting prototype for the road condition and data product delivery is an important issue for reliable services [218].
collection. They examine the challenges of these prototype The sustainable delivery and collection network is an issue in the
developments. transport system. Chena et al. [219] proposed a citywide e-
commerce application based on crowdsourcing. E-commerce can
6 Future of crowd intelligence in transportation improve its services with the help of delivery systems such as
drone delivery, last-mile delivery, autonomous delivery. These
system services can be possible with a crowd-intelligence system with
The crowd-intelligence methods for different layers of application autonomous mobility [220].
have been surveyed in this paper. From the layer structure of the
transportation system, it is learned that crowd intelligence plays a 6.1.2 Emergency vehicle notification: There are several
central role for coordination and control of the transportation algorithms for activity detection [221] of human and vehicle
system. The problem area and the crowdsource-based solution suggested in the literature. However, these algorithms fail most of
methods have been found out for different layers of the transport the time because of GPS failure or GPS error. Stitzel et al. [222]
system. In this section a broader classification of transport problem developed an automatic crash notification system which uses
and crowd-intelligence-based solutions are classified followed by telemetry measurement to predict the risk of the vehicle crash. The
future direction from the literature is outlined. authors considered response time and intensity of the crash as the
matrices for notification. Further, the authors incorporated this
6.1 Classification of crowd-intelligence methods algorithm into the vehicle for crash intensity detection to inform
emergency decision. The driver's behaviour prediction [223–225]
The crowd-intelligence methods have been described for different algorithms based on visual sensing are developed to detect drowsy
layers of the transportation system. The explained methods can be and distraction of driver for safety in road. Advance driver
categorised based upon different problem areas such as travel detection and collision avoidance systems have been developed by
demand prediction, optimum route recommendation, traffic using the crowdsourced approach. Hadiwardoyo et al. [226] and
management, traffic state prediction, missing data modelling for Tornell et al. [227] developed safety navigation applications in
crowdsensing, dynamic topology of vehicular network and user vehicular network scenarios. These methods use a network
mobility modelling. The intelligent approaches to process these infrastructure of the vehicle to notify other driver about the
crowdsensed data are reinforcement learning, game theory, emergency. These methods fully rely on crowd-sensed data without
dimensionality reduction, clustering, artificial neural network, any assistance from the vehicular infrastructure. The social media
SVM, optimisation, fuzzy logic and probabilistic inference feeds are considered as a huge source of information and novel
methods as shown in Table 4. platform for crowdsourcing. These informations are used to detect
The future transportation system relies on crowd-intelligence emergencies such as the traffic jam, storms and fire. Xu et al. [228]
techniques. The futuristic transportation system needs to solve the proposed a 5W (what, where, when, who and why) model for
above-mentioned problems. There are various methods present in urban emergency detection using social crowd-sensed data. The
the literature to tackle these problems. Mostafaeipour et al. [184] authors use spatial and temporal information from social media
proposed a meta-heuristic method such as bat and firefly algorithm information to detect real-time events. Further, they used GIS-
to calculate travel demand of a transportation network. A based annotation of these events for intelligent prediction.
spatiotemporal fuzzy neural network for passenger demand is
proposed by Liang et al. [185]. The authors used feature fusion
with the convolution for the demand prediction. A probabilistic 6.1.3 Dynamic traffic light: The crowd-sensing-based traffic light
inference model based on day-to-day traffic data is proposed by Ma management method has wide applications in the smart transport
and Qian [186] for travel demand prediction. system. The static and dynamic map generation from the
crowdsource data is presented several times in the literature [229–
231]. The maps from the crowdsourced data give information such
6.1.1 E-commerce with crowdsource: The delivery system of as lane geometry, lane boundary, speed limit, road stability status,
future cities faces challenges such as high demand, less delivery road topology and road intersection. Ruhhammer et al. [196]
time and complex traffic congestion. Crowd-sense-based delivery proposed a method to extract road intersection and traffic light
system guarantees to address these issues. The crowdsource-based information from the crowd-sensed data. The authors have
transport system helps to understand the behaviour patterns of e- identified driving directions in the entry point of intersection for a
commerce systems. The development of crowdsourcing-based set of traffic light scenarios. They suggested that the traffic lights
delivery and return of goods system is a real-time application of the

Table 4 Problem area versus crowd-intelligent solutions in transport system


Area versus Reinforcement Game Dimensionality Clustering Artificial SVM Search and Logic / Probabilistic
methods learning theory reduction (CNN, neural optimisation fuzzy inference
DBM, network logic
Kmeans) (supervised)
travel demand [156] [174] [179, 180] [181] [182] [183] [184] [185] [186]
prediction
optimum path/route [187] [188] [189] [96] [162] [190] [111, 114, 118] [191] [119]
recommendation
traffic management [192] [193] [175, 176] [127, 178] [177] [194] [47, 123, 165] [195] [196]
traffic state [197] [198] [199] [200] [86, 125, 168] [201] [53, 167] [169] [126]
prediction
missing traffic data [68] [202] [67] [203] [60, 69–71] [59] [62–66] [61] [68, 72]
modelling
dynamic parking [204] [205] [206] [207] [208] [208] [209] [210] [211]
method
dynamic topologies [212] [51] [67] [178] [125] [201] [160] [169] [141]
of vehicular network
user mobility [212] [213] [137] [214] [215] [215] [216] [217] [112, 138,
modelling 139, 141]

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 489
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
can be improved with this adaptive traffic light signalling method. control [192], traffic flow management [193] and crash prediction
Further, the authors experimented the proposed method by taking [194] show a path to automated traffic prediction. Further, the
344 intersections from a city and proposed a promising solution. automated traffic management needs traffic state prediction which
is possible with crowd-intelligence techniques such as the traffic
6.1.4 Collision avoidance system: The growth of crowd-sensing signal prediction [197], lane changing detection [198], traffic stat
and intelligent processing techniques makes the vehicle navigate prediction [199, 201] and congestion detection [200].
autonomously through the unknown environment. The existing The crowdsensing techniques often miss sensing some portion
approaches [232, 233] use scanning or point cloud [234] methods of traffic information. The future transport system should model
to form the grids of the vehicle or robots [235]. These grids help these missing traffic data with intelligent techniques. There are
the evaluation of the motion pattern without collision [236]. These crowd-intelligence techniques which use Gaussian process [68],
grids use mapping solutions for the development of collision game theory [202], clustering [203] and reinforcement learning
avoidance system. The computational and hardware burden are [68] methods for these missing data modelling. Smart parking
addressed by crowdsourcing-based mapping for a collision space searching and their management are an essential part of the
avoidance system. Lopez et al. [237] developed a perception and futuristic transportation system. This is achieved by crowd-
planning-based collision avoidance system using crowd intelligence techniques such as optimisation [209], fuzzy logic
information. The authors used intermediate perception for efficient [210], SVM [208], clustering [207], game theory [205],
computation, which estimates from terminal states until a collision- dimensionality reduction [206] and reinforcement learning [204].
free path is generated. Further, the authors experimented and The development of future transport system depends upon the user
concluded its ability to plan aggressive collision avoidance in high mobility model. The crowdsensing data by using intelligent data
cluttered scenarios. processing techniques such as reinforcement learning [212], game
theory [213], cluster traffic model [214], SVM [215], fuzzy-based
6.1.5 Automatic road enforcement: Road enforcement is an [217] and optimised [216] models can form the basis of future
official procedure of allowing or penalising the lawbreakers of the mobility modelling of transportation systems. A comparative table
road. The road safety is related to inconsistencies of the assessment of these crowd-intelligence methods is shown in Table 4.
of road enforcement laws. Proper road enforcement can reduce
injuries and death in road transportation. The road enforcement 7 Conclusion
comprises traffic law, alcohol enforcement, speed, seat belt and In this paper, the crowd-intelligence-based techniques for future
advanced traffic systems. Current road enforcement is based on intelligent transport system are reviewed. The crowd intelligence is
road policing to comply with traffic laws [238, 239]. Al-Adhoobi described as crowdsensing followed by intelligent processing of
et al. [240] conducted a study for evaluation of enforcements and these sensing data to solve the transport problems. This paper first
their measurements. Further, they examined means of maximising defines crowd intelligence and its properties. Then different layers
enforcement strategies by automation based on crowd-intelligence of the transport system are discussed and problems of each layer
approaches. This automatic approach provides road safety to are outlined. Further, the crowd-intelligence-based solutions for
several vehicles. each layer are presented. Finally, different problems of the
transport system layers are extensively analysed and the algorithms
6.1.6 Transport demand modelling: The development of long- are classified based on the problem area and intelligent solution
term infrastructure, sustainable urban mobility, congestion method. Moreover, a comparative table is formulated comparing all
management and control requires the knowledge of transport the discussed crowd-intelligent algorithms. This outline and
behaviour of the city. These transport behaviours can be modelled classification give an insight for further research direction of novel
mathematically for prediction of travel demand [181, 182]. This crowd-intelligent algorithms. This survey portrays the algorithms
model considers the proper selection of the route, choice of route and their evolution for smart mobility and its impact on future
delay, mobility and traffic flow as parameters for modelling. In the human mobility with crowd intelligence for individual and
literature parameters like speed, delay and queues have been collective planning.
considered for modelling the urban mobility [241]. The use of
crowd-sensed data from the mobile phone for mobility prediction
of the vehicles is present in the literature. A dimensionality 8 References
reduction-based travel demand model is presented by Duan et al. [1] Santi, P., Resta, G., Szell, M., et al.: ‘Quantifying the benefits of vehicle
[179] Chu et al. [180]. Several models have been developed in the pooling with shareability networks’, Proc. Natl. Acad. Sci., 2014, 111, (37),
pp. 13290–13294
literature correlating the movement of the mobile phone with the [2] Zegras, P.C., Eros, E., Butts, K., et al.: ‘Tracing a path to knowledge?
vehicle displacement [242]. These displacement matrix scaling and Indicative user impacts of introducing a public transport map in Dhaka,
optimisation are big challenges for real-time traffic scenarios. Bangladesh’, Camb. J. Regions Econ. Soc., 2014, 8, (1), pp. 113–129
Hence, correction factors for scaling and implementing in real [3] Feiler, P.H., Gluch, D.P.: ‘Model-based engineering with AADL: an
introduction to the SAE architecture analysis & design language’ (Addison-
traffic scenarios are required. The transportation data from round Wesley, Boston, MA, USA., 2012)
trips and chain trips is important for prediction of travel behaviour. [4] Gobble, M.M.: ‘Defining the sharing economy’, Res.-Technol. Manage.,
Hence, to extract and model the public transportation, the mining 2017, 60, (2), pp. 59–63
of the chain trip from the data is essential which further depends [5] Uteng, T.P., Julsrud, T.E., George, C.: ‘The role of life events and context in
type of car share uptake: comparing users of peer-to-peer and cooperative
upon city policies. Another approach of mobility pattern modelling programs in Oslo, Norway’, Transp. Res. D, Transp. Environ., 2019, 71, pp.
uses the GPS data and extracts mobility models from the data by 186–206
using geometric and probabilistic models [243]. Hafezi et al. [244] [6] Cafiso, S., Di Graziano, A., Pappalardo, G.: ‘Road safety issues for bus
proposed an activity pattern recognition approach for travel transport management’, Accident Anal. Prev., 2013, 60, pp. 324–333
[7] Redondi, A., Cesana, M., Tagliasacchi, M., et al.: ‘Cooperative image
demand modelling. Jiang et al. [183] developed the least-square analysis in visual sensor networks’, Ad Hoc Netw., 2015, 28, pp. 38–51
SVM for transport demand modelling. [8] Ferreira, J., Alam, M., Fernandes, B., et al.: ‘Cooperative sensing for
improved traffic efficiency: the highway field trial’, Comput. Netw., 2018,
143, pp. 82–97
6.2 Future direction of crowd intelligence for transport system [9] Osanaiye, O., Chen, S., Yan, Z., et al.: ‘From cloud to fog computing: a
review and a conceptual live vm migration framework’, IEEE Access, 2017,
The future direction of the intelligent transportation system is to 5, pp. 8284–8300
develop a smart route recommendation system. The route [10] Banerjee, S., Choudekar, P., Muju, M.: ‘Real time car parking system using
recommendation system needs crowdsensing followed by image processing’. 2011 3rd Int. Conf. on Electronics Computer Technology,
intelligent data processing algorithms such as reinforcement Kanyakumari, India, 2011, vol. 2, pp. 99–103
[11] Al Kharusi, H., Al Bahadly, I.: ‘Intelligent parking management system based
learning [187], game theory [188], dimensionality reduction [189], on image processing’, World J. Eng. Technol., 2014, 2, (2), pp. 55–67
support vector regression [190] and neuro-fuzzy approaches [191]. [12] Feng, Y., Head, K.L., Khoshmagham, S., et al.: ‘A real-time adaptive signal
The automated traffic management is essential for the future control in a connected vehicle environment’, Transp. Res. C, Emerg. Technol.,
transportation system and recent techniques such as the traffic light 2015, 55, pp. 460–473

490 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[13] Ramezani, M., de Lamberterie, N., Skabardonis, A., et al.: ‘A link [46] Wan, J., Zhang, D., Zhao, S., et al.: ‘Context-aware vehicular cyber-physical
partitioning approach for real-time control of queue spillbacks on congested systems with cloud support: architecture, challenges, and solutions’, IEEE
arterials’, Transportmetrica B: Transp. Dyn., 2017, 5, (2), pp. 177–190 Commun. Mag., 2014, 52, (8), pp. 106–113
[14] Lu, C., Dong, J.: ‘Estimating freeway travel time and its reliability using radar [47] Baguena, M., Calafate, C.T., Cano, J.C., et al.: ‘An adaptive anycasting
sensor data’, Transportmetrica B: Transp. Dyn., 2018, 6, (2), pp. 97–114 solution for crowd sensing in vehicular environments’, IEEE Trans. Ind.
[15] Li, Z., Jiang, S., Li, L., et al.: ‘Building sparse models for traffic flow Electron., 2015, 62, (12), pp. 7911–7919
prediction: an empirical comparison between statistical heuristics and [48] Wan, J., Liu, J., Shao, Z., et al.: ‘Mobile crowd sensing for traffic prediction
geometric heuristics for Bayesian network approaches’, Transportmetrica B: in internet of vehicles’, Sensors, 2016, 16, (1), p. 88
Transp. Dyn., 2019, 7, (1), pp. 107–123 [49] Guo, B., Wang, Z., Yu, Z., et al.: ‘Mobile crowd sensing and computing: the
[16] Legara, E.F.T., Monterola, C.P.: ‘Inferring passenger types from commuter review of an emerging human-powered sensing paradigm’, ACM Comput.
eigentravel matrices’, Transp. B: Transp. Dyn., 2018, 6, (3), pp. 230–250 Surv. (CSUR), 2015, 48, (1), p. pp. 1–31
[17] Medina, S.A.O.: ‘Inferring weekly primary activity patterns using public [50] Haddadou, N., Rachedi, A., Ghamri Doudane, Y.: ‘A job market signaling
transport smart card data and a household travel survey’, Travel Behaviour scheme for incentive and trust management in vehicular ad hoc networks’,
Soc., 2018, 12, pp. 93–101 IEEE Trans. Veh. Technol., 2015, 64, (8), pp. 3657–3674
[18] Lam, A.Y., Leung, Y.W., Chu, X.: ‘Autonomous-vehicle public transportation [51] Zhang, J., Gauthier, V., Labiod, H., et al.: ‘Information dissemination in
system: scheduling and admission control’, IEEE Trans. Intell. Transp. Syst., vehicular networks via evolutionary game theory’. 2014 IEEE Int. Conf. on
2016, 17, (5), pp. 1210–1226 Communications (ICC), 2014, pp. 124–129
[19] Foell, S., Phithakkitnukoon, S., Kortuem, G., et al.: ‘Predictability of public [52] Yang, Q., Wang, H.: ‘Towards trustworthy vehicular social network’, IEEE
transport usage: a study of bus rides in Lisbon, Portugal’, IEEE Trans. Intell. Commun. Mag., 2015, 53, (8), pp. 42–47
Transp. Syst., 2015, 16, (5), pp. 2955–2960 [53] Timpner, J., Schürmann, D., Wolf, L.: ‘Trustworthy parking communities:
[20] Rathod, R., Khot, S.: ‘Smart assistance for public transport system’. 2016 Int. helping your neighbor to find a space’, IEEE Trans. Dependable Secur.
Conf. on Inventive Computation Technologies (ICICT), Coimbatore, India, Comput., 2016, 13, (1), pp. 120–132
2016, vol. 3, pp. 1–5 [54] Abbani, N., Jomaa, M., Tarhini, T., et al.: ‘Managing social networks in
[21] Leone, G., Moroni, D., Pieri, G., et al.: ‘An intelligent cooperative visual vehicular networks using trust rules’. 2011 IEEE Symp. on Wireless
sensor network for urban mobility’, Sensors, 2017, 17, (11), p. 2588 Technology and Applications (ISWTA), 2011, pp. 168–173
[22] Buch, N., Velastin, S.A., Orwell, J.: ‘A review of computer vision techniques [55] Gong, L., Fan, W.: ‘Applying travel-time reliability measures in identifying
for the analysis of urban traffic’, IEEE Trans. Intell. Transp. Syst., 2011, 12, and ranking recurrent freeway bottlenecks at the network level’, J. Transp.
(3), pp. 920–939 Eng., A Syst., 2017, 143, (8), p. 04017042
[23] Chen, Y.L., Wu, B.F., Huang, H.Y., et al.: ‘A real-time vision system for [56] Mauch, M., Skabardonis, A., Davies, L.: ‘Validating the cost-effectiveness
nighttime vehicle detection and traffic surveillance’, IEEE Trans. Ind. model for California's freeway incident management program’. Celebrating
Electron., 2011, 58, (5), pp. 2030–2044 50 Years of Traffic Flow Theory: A Symp., 2014, pp. 179–191
[24] Loce, R.P., Bernal, E.A., Wu, W., et al.: ‘Computer vision in roadway [57] Little, R.J., Rubin, D.B.: ‘Statistical analysis with missing data’ (Wiley,
transportation systems: a survey’, J. Electron. Imaging, 2013, 22, (4), p. 2019), vol. 793
041121 [58] Rodrigues, F., Henrickson, K., Pereira, F.C.: ‘Multi-output Gaussian processes
[25] Mainetti, L., Patrono, L., Stefanizzi, M.L., et al.: ‘A smart parking system for crowdsourced traffic data imputation’, IEEE Trans. Intell. Transp. Syst.,
based on iot protocols and emerging enabling technologies’. 2015 IEEE 2nd 2019, 20, (99), pp. 594–603
World Forum on Internet of Things (WF-IoT) IEEE, Milan, Italy, 2015, pp. [59] Zhang, Y., Liu, Y.: ‘Data imputation using least squares support vector
764–769 machines in urban arterial streets’, IEEE Signal Process. Lett., 2009, 16, (5),
[26] Bielsa, A.: ‘Smart city project in santander to monitor parking free slots’ pp. 414–417
(Libelium World, Hoboken, NJ, USA., 2013) [60] Chang, G., Zhang, Y., Yao, D.: ‘Missing data imputation for traffic flow based
[27] Huang, C.C., Wang, S.J.: ‘A hierarchical Bayesian generation framework for on improved local least squares’, Tsinghua Sci. Technol., 2012, 17, (3), pp.
vacant parking space detection’, IEEE Trans. Circuits Syst. Video Technol., 304–309
2010, 20, (12), pp. 1770–1785 [61] Tang, J., Zhang, G., Wang, Y., et al.: ‘A hybrid approach to integrate fuzzy c-
[28] Magrini, M., Moroni, D., Nastasi, C., et al.: ‘Visual sensor networks for means based imputation method with genetic algorithm for missing traffic
infomobility’, Pattern Recognit. Image Anal., 2011, 21, (1), pp. 20–29 volume data estimation’, Transp. Res. C, Emerg. Technol., 2015, 51, pp. 29–
[29] De Almeida, P.R., Oliveira, L.S., Britto, A.SJr, et al.: ‘Pklot – a robust dataset 40
for parking lot classification’, Expert Syst. Appl., 2015, 42, (11), pp. 4937– [62] Asif, M.T., Mitrovic, N., Garg, L., et al.: ‘Low-dimensional models for
4949 missing data imputation in road networks’. 2013 IEEE Int. Conf. on
[30] Amato, G., Carrara, F., Falchi, F., et al.: ‘Car parking occupancy detection Acoustics, Speech and Signal Processing, 2013, pp. 3527–3531
using smart camera networks and deep learning’. 2016 IEEE Symp. on [63] Tak, S., Woo, S., Yeo, H.: ‘Data-driven imputation method for traffic data in
Computers and Communication (ISCC) (IEEE, 2016, pp. 1212–1217 sectional units of road links’, IEEE Trans. Intell. Transp. Syst., 2016, 17, (6),
[31] Wang, X., Zheng, X., Zhang, Q., et al.: ‘Crowdsourcing in its: the state of the pp. 1762–1771
work and the networking’, IEEE Trans. Intell. Transp. Syst., 2016, 17, (6), pp. [64] Mardani, M., Mateos, G., Giannakis, G.B.: ‘Subspace learning and imputation
1596–1605 for streaming big data matrices and tensors’, IEEE Trans. Signal Process.,
[32] Grazioli, A., Picone, M., Zanichelli, F., et al.: ‘Collaborative mobile 2015, 63, (10), pp. 2663–2677
application and advanced services for smart parking’. 2013 IEEE 14th Int. [65] Ran, B., Song, L., Zhang, J., et al.: ‘Using tensor completion method to
Conf. on Mobile Data Management, Milan, Italy, 2013, vol. 2, pp. 39–44 achieving better coverage of traffic state estimation from sparse floating car
[33] Zhang, S., Xiong, R.: ‘Adaptive energy management of a plug-in hybrid data’, PloS One, 2016, 11, (7), p. e0157420
electric vehicle based on driving pattern recognition and dynamic [66] Chiou, J.M., Zhang, Y.C., Chen, W.H., et al.: ‘A functional data approach to
programming’, Appl. Energy, 2015, 155, pp. 68–78 missing value imputation and outlier detection for traffic flow data’,
[34] Jian, L., Zheng, Y., Xiao, X., et al.: ‘Optimal scheduling for vehicle-to-grid Transportmetrica B Transp. Dyn., 2014, 2, (2), pp. 106–129
operation with stochastic connection of plug-in electric vehicles to smart [67] Li, L., Li, Y., Li, Z.: ‘Efficient missing data imputing for traffic flow by
grid’, Appl. Energy, 2015, 146, pp. 150–161 considering temporal and spatial dependence’, Transp. Res. C, Emerg.
[35] Pointon, J.: ‘The multi-unit dwelling vehicle charging challenge’, Electr. Veh. Technol., 2013, 34, pp. 108–120
Virtual Summit, 2012, 69, pp. 1–33 [68] Duan, Y., Lv, Y., Liu, Y.L., et al.: ‘An efficient realization of deep learning for
[36] Al Junaibi, R., Farid, A.M.: ‘A method for the technical feasibility assessment traffic data imputation’, Transp. Res. C, Emerg. Technol., 2016, 72, pp. 168–
of electrical vehicle penetration’. 2013 IEEE Int. Systems Conf (SysCon), 181
2013, pp. 606–611 [69] Elhenawy, M., Rakha, H.: ‘Spatiotemporal traffic state prediction based on
[37] Ryley, T., Burchell, J., Davison, L.: ‘Research in transportation business & discriminatively pre-trained deep neural networks’, Adv. Sci. Technol. Eng.
management’, 2013 Syst. J., 2017, 2, pp. 678–686
[38] Areddy, J.T., Jie, Y.: ‘How china's train tragedy unfolded’, China Realtime [70] Ma, X., Dai, Z., He, Z., et al.: ‘Learning traffic as images: a deep
Report, The Wall Street Journal, China, Available at: http://blogs wsj com/ convolutional neural network for large-scale transportation network speed
chinarealtime/2011/12/29/wenzhou% E2, 2011, 80 prediction’, Sensors, 2017, 17, (4), p. 818
[39] Fleming, C.H., Leveson, N.G.: ‘Early concept development and safety [71] Yu, H., Wu, Z., Wang, S., et al.: ‘Spatiotemporal recurrent convolutional
analysis of future transportation systems’, IEEE Trans. Intell. Transp. Syst., networks for traffic prediction in transportation networks’, Sensors, 2017, 17,
2016, 17, (12), pp. 3512–3523 (7), p. 1501
[40] Scott, A., Menn, J.: ‘Exclusive: air traffic system failure caused by computer [72] Álvarez, M.A., Lawrence, N.D.: ‘Computationally efficient convolved
memory shortage’ (Reuters, London, UK, 2014) multiple output Gaussian processes’, J. Mach. Learn. Res., 2011, 12, (May),
[41] Jiménez, F., Naranjo, J.E., Anaya, J.J., et al.: ‘Advanced driver assistance pp. 1459–1500
system for road environments to improve safety and efficiency’, Transp. Res. [73] Liu, S., Yue, Y., Krishnan, R.: ‘Adaptive collective routing using Gaussian
Procedia, 2016, 14, pp. 2245–2254 process dynamic congestion models’. Proc. of the 19th ACM SIGKDD Int.
[42] Agarwal, V., Banerjee, N., Chakraborty, D., et al.: ‘Usense – a smartphone Conf. on Knowledge Discovery and Data Mining ACM, Chicago, IL, USA.,
middleware for community sensing’. 2013 IEEE 14th Int. Conf. on Mobile 2013, pp. 704–712
Data Management, Milan, Italy, 2013, vol. 1, pp. 56–65 [74] Rodrigues, F., Borysov, S.S., Ribeiro, B., et al.: ‘A Bayesian additive model
[43] Fan, X., Liu, J., Wang, Z., et al.: ‘Crowdsourced road navigation: concept, for understanding public transport usage in special events’, IEEE Trans.
design, and implementation’, IEEE Commun. Mag., 2017, 55, (6), pp. 126– Pattern Anal. Mach. Intell., 2017, 39, (11), pp. 2113–2126
128 [75] Xie, Y., Zhao, K., Sun, Y., et al.: ‘Gaussian processes for short-term traffic
[44] Zheng, Y.: ‘Trajectory data mining: an overview’, ACM Trans. Intell. Syst. volume forecasting’, Transp. Res. Rec., 2010, 2165, (1), pp. 69–78
Technol. (TIST), 2015, 6, (3), p. pp. 1–41 [76] Garip, M.T., Reiher, P., Gerla, M.: ‘Ghost: concealing vehicular botnet
[45] Yang, Q., Zhu, B., Wu, S.: ‘An architecture of cloud-assisted information communication in the vanet control channel’. 2016 Int. Wireless
dissemination in vehicular networks’, IEEE Access, 2016, 4, pp. 2764–2770 Communications and Mobile Computing Conf. (IWCMC), 2016, pp. 1–6

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 491
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[77] Joy, J.: ‘Vehicular blocktrees’. 2017 IEEE Vehicular Networking Conf. [108] de Almeida Buosi, M., Cilloni, M., Corradi, A., et al.: ‘A crowdsensing
(VNC), Turin, Italy, 2017, pp. 147–150 campaign and data analytics for assisting urban mobility pattern
[78] Joy, J., Cusack, G., Gerla, M.: ‘Poster: time analysis of the feasibility of determination’. 2018 IEEE Symp. on Computers and Communications
vehicular blocktrees’. Proc. of the 3rd Workshop on Experiences with the (ISCC), 2018, pp. 00224–00229
Design and Implementation of Smart Objects, Snowbird, UT, USA., 2017, pp. [109] Dienstl, B., Scholz, J.: ‘A concept for smart transportation user-feedback
25–26 utilizing volunteered geoinformation approaches’. The 4th Conf. on
[79] Kulkarni, V., Chapuis, B., Garbinato, B., et al.: ‘Addressing the free-rider Sustainable Urban Mobility, 2018, pp. 538–545
problem in public transport systems’, arXiv preprint, arXiv:180304389, 2018 [110] Griffin, G.P., Jiao, J., et al.: ‘Can crowdsourcing support co-productive
[80] Fürst, E.: ‘Free riders and ticket fraud in public transport: a delphi analysis’. transportation planning in megaregion? evidence from local practice’
European Transport Conf. 2012 Association for European Transport (AET) (University of Texas at Austin. Cooperative Mobility for Competitive
Transportation Research Board, Glasgow, UK., 2012 Megaregions, (CM2-2017-2), 2018, pp. 1–37
[81] Edwards, F.L., Goodrich, D.C.: ‘Introduction to transportation security’ [111] Goel, N., Sharma, R., Nikhil, N., et al.: ‘A crowd-sourced adaptive safe
(CRC Press, Boca Raton, FL, USA., 2016) navigation for smart cities’. 2017 IEEE Int. Symp. on Multimedia (ISM),
[82] Goodfellow, I., Pouget Abadie, J., Mirza, M., et al.: ‘Generative adversarial 2017, pp. 382–387
nets’. Advances in Neural Information Processing Systems 27 (NIPS 2014), [112] Huang, Z., Ling, X., Wang, P., et al.: ‘Modeling real-time human mobility
2014, pp. 2672–2680 based on mobile phone and transportation data fusion’, Transp. Res. C,
[83] Dwork, C., McSherry, F., Nissim, K., et al.: ‘Calibrating noise to sensitivity in Emerg. Technol., 2018, 96, pp. 251–269
private data analysis’, J. Privacy Confidentiality, 2016, 7, (3), pp. 17–51 [113] Faber, A., Rehm, S.V., Hernandez Mendez, A., et al.: ‘Modeling and
[84] Joy, J., Gray, D., McGoldrick, C., et al.: ‘Scalable privacy: poster’. Proc. of visualizing smart city mobility business ecosystems: insights from a case
the 10th ACM Conf. on Security and Privacy in Wireless and Mobile study’, Information, 2018, 9, (11), p. 270
Networks, Boston, MA, USA., 2017, pp. 271–272 [114] Kong, X., Li, M., Tang, T., et al.: ‘Shared subway shuttle bus route planning
[85] Wang, X., Wu, W., Qi, D.: ‘Mobility-aware participant recruitment for based on transport data analytics’, IEEE Trans. Autom. Sci. Eng., 2018, 15,
vehicle-based mobile crowdsensing’, IEEE Trans. Veh. Technol., 2018, 67, (4), pp. 1507–1520
(5), pp. 4415–4426 [115] Mukheja, P., Velaga, N.R., Sharmila, R., et al.: ‘Smartphone-based
[86] Mai, T., Ghosh, B., Wilson, S.: ‘Multivariate short-term traffic flow crowdsourcing for position estimation of public transport vehicles’, IET Intell.
forecasting using Bayesian vector autoregressive moving average model’, Transp. Syst., 2017, 11, (9), pp. 588–595
Transportation Research Board 91st Annual Meeting, Washington DC, USA, [116] Leao, S.Z., Lieske, S.N., Pettit, C.J.: ‘Validating crowdsourced bicycling
2012 mobility data for supporting city planning’, Transp. Lett., 2019, 11, pp. 486–
[87] Kumar, S.V., Vanajakshi, L.: ‘Short-term traffic flow prediction using 497
seasonal arima model with limited input data’, Eur. Transp. Res. Rev., 2015, 7, [117] Hoy, K.N., Rogala, S.: ‘Using a crowdsourcing tool to collect data on the
(3), p. 21 travel behaviour and needs of individuals with reduced mobility’. Scientific
[88] Zarei, N., Ghayour, M.A., Hashemi, S.: ‘Road traffic prediction using and Technical Conf. Transport Systems Theory and Practice, 2018, pp. 75–84
context-aware random forest based on volatility nature of traffic flows’. Asian [118] Wang, J., Lu, Y., Wang, X., et al.: ‘Sar: a social-aware route recommendation
Conf. on Intelligent Information and Database Systems, 2013, pp. 196–205 system for intelligent transportation’, Comput. J., 2018, 61, (7), pp. 987–997
[89] Widhalm, P., Yang, Y., Ulm, M., et al.: ‘Discovering urban activity patterns in [119] Singh, G., Bansal, D., Sofat, S., et al.: ‘Smart patrolling: an efficient road
cell phone data’, Transportation. (Amst), 2015, 42, (4), pp. 597–623 surface monitoring using smartphone sensors and crowdsourcing’, Pervasive
[90] Goodall, W., Dovey, T., Bornstein, J., et al.: ‘The rise of mobility as a Mob. Comput., 2017, 40, pp. 71–88
service’, Deloitte Rev., 2017, 20, pp. 112–129 [120] Tomaras, D.: ‘Crowdsourcing techniques for smart urban mobility’. 2018
[91] Jittrapirom, P., Caiati, V., Feneri, A.M., et al.: ‘Mobility as a service: a critical IEEE Int. Conf. on Pervasive Computing and Communications Workshops
review of definitions, assessments of schemes, and key challenges’, Urban (PerCom Workshops), 2018, pp. 460–461
Plan., 2017, 2, p. 13 [121] Verma, R.: ‘Crowdmap: crowdsourcing based city traffic map generation’.
[92] Parker, J.M.: ‘Applying a system of systems approach for improved 2017 9th Int. Conf. on Communication Systems and Networks
transportation’, SAPI EN S Surv. Perspectives Integrating Environ. Soc., (COMSNETS), 2017, pp. 582–583
2010, 3, (2), pp. 1–5 [122] Jameela, M., Afzal, H., Khurshid, K., et al.: ‘Crowdsourced system to report
[93] Giuli, D., Paganelli, F., Cuomo, S., et al.: ‘Toward a cooperative approach for traffic violations’. Proc. of the 4th Int. Conf. on Vehicle Technology and
continuous innovation of mobility information services’, IEEE Syst. J., 2013, Intelligent Transport Systems (VEHITS-2018), Funchal, Madeira, Portugal,
7, (4), pp. 669–680 vol. 1, 2018
[94] Callegati, F., Giallorenzo, S., Melis, A., et al.: ‘Insider threats in emerging [123] Iannucci, B., Ramirez, P.C.G., Teng, E., et al.: ‘Crowdsourced traffic
mobility-as-a-service scenarios’. Proc. of the 50th Hawaii Int. Conf. on calming’, 2018
System Sciences (2017), 2017, Available: http://dx.doi.org/10.24251/ [124] Fujihara, A.: ‘Proposing a system for collaborative traffic information
HICSS.2017.321 gathering and sharing incentivized by blockchain technology’. Int. Conf. on
[95] Callegati, F., Gabbrielli, M., Giallorenzo, S., et al.: ‘Smart mobility for all: a Intelligent Networking and Collaborative Systems, 2018, pp. 170–182
global federated market for mobility-as-a-service operators’. 2017 IEEE 20th [125] Iyer, S.R., Boxer, K., Subramanian, L.: ‘Urban traffic congestion mapping
Int. Conf. on Intelligent Transportation Systems (ITSC), 2017, pp. 1–8 using bus mobility data’. CEUR Workshop Proc., 2018, vol. 2227, pp. 7–13
[96] Senanayake, J., Wijayanayake, W.: ‘Applicability of crowd sourcing to [126] Khan, S.Z., Rahuman, W.A., Dey, S., et al.: ‘Roadcrowd: an approach to road
determine the best transportation method by analysing user mobility’, traffic forecasting at junctions using crowd-sourcing and Bayesian model’.
Available at SSRN 3272072, 2018 2017 Int. Conf. on Research and Innovation in Information Systems (ICRIIS),
[97] Carnevale, L., Celesti, A., Di Pietro, M., et al.: ‘How to conceive future Langkawi, Malaysia, 2017, pp. 1–6
mobility services in smart cities according to the fiware frontiercities [127] Kong, X., Song, X., Xia, F., et al.: ‘Lotad: long-term traffic anomaly detection
experience’, IEEE Cloud Comput., 2018, 5, (5), pp. 25–36 based on crowdsourced bus trajectory data’, World Wide Web, 2018, 21, (3),
[98] Suciu, G., Butca, C., Dobre, C., et al.: ‘Smart city mobility simulation and pp. 825–847
monitoring platform’. 2017 21st Int. Conf. on Control Systems and Computer [128] Rantanen, P., Sillberg, P., Soini, J.: ‘Towards the utilization of crowdsourcing
Science (CSCS), 2017, pp. 685–689 in traffic condition reporting’. 2017 40th Int. Convention on Information and
[99] Xia, F., Rahim, A., Kong, X., et al.: ‘Modeling and analysis of large-scale Communication Technology, Electronics and Microelectronics (MIPRO),
urban mobility for green transportation’, IEEE Trans. Ind. Inf., 2018, 14, (4), 2017, pp. 985–990
pp. 1469–1481 [129] Verma, R., Ghosh, S., Shrivastava, A., et al.: ‘Unsupervised annotated city
[100] Andersen, A., Karlsen, R., Yu, W.: ‘Green transportation choices with iot and traffic map generation’. Proc. of the 24th ACM SIGSPATIAL Int. Conf. on
smart nudging’, in: ‘Handbook of smart cities’ (Springer, Cham, Switzerland, Advances in Geographic Information Systems, Burlingame, CA, USA., 2016,
2018), pp. 331–354 p. 59
[101] Schreieck, M., Pflügler, C., Setzke, D.S., et al.: ‘Improving urban [130] Tomaras, D., Boutsis, I., Kalogeraki, V.: ‘Travel time estimation in real-time
transportation: an open plat-form for digital mobility services’, in: ‘Digital using buses as speed probes’. 2015 IEEE Int. Conf. on Pervasive Computing
marketplaces unleashed’ (Springer, Berlin, Germany, 2018), pp. 479–489 and Communication Workshops (PerCom Workshops), 2015, pp. 63–68
[102] Biem, A., Bouillet, E., Feng, H., et al.: ‘Real-time traffic information [131] Basole, R.C., Russell, M.G., Huhtamäki, J., et al.: ‘Understanding business
management using stream computing’, IEEE Data Eng. Bull., 2010, 33, (2), ecosystem dynamics: a data-driven approach’, ACM Trans. Manage. Inf. Syst.
pp. 64–68 (TMIS), 2015, 6, (2), pp. 1–32
[103] Ehmke, J.F., Mattfeld, D.C.: ‘Data allocation and application for time- [132] Rehm, S.V., Goel, L.: ‘Using information systems to achieve complementarity
dependent vehicle routing in city logistics’, Europ. Transp., 2010, 46, pp. 24– in sme innovation networks’, Inf. Manage., 2017, 54, (4), pp. 438–451
35 [133] Lu, X., Bengtsson, L., Holme, P.: ‘Predictability of population displacement
[104] Nickerson, R.C., Varshney, U., Muntermann, J.: ‘A method for taxonomy after the 2010 Haiti earthquake’, Proc. Natl. Acad. Sci., 2012, 109, (29), pp.
development and its application in information systems’, Eur. J. Inf. Syst., 11576–11581
2013, 22, (3), pp. 336–359 [134] Song, C., Qu, Z., Blumm, N., et al.: ‘Limits of predictability in human
[105] Harri, J., Filali, F., Bonnet, C.: ‘Mobility models for vehicular ad hoc mobility’, Science, 2010, 327, (5968), pp. 1018–1021
networks: a survey and taxonomy’, IEEE Commun. Surv. Tutorials, 2009, 11, [135] Ahas, R., Silm, S., Järv, O., et al.: ‘Using mobile positioning data to model
(4), pp. 19–41 locations meaningful to users of mobile phones’, J. Urban Technol., 2010, 17,
[106] Samaranayake, S., Glaser, S., Holstius, D., et al.: ‘Real-time estimation of (1), pp. 3–27
pollution emissions and dispersion from highway traffic’, Comput.-Aided Civ. [136] Järv, O., Tenkanen, H., Toivonen, T.: ‘Enhancing spatial accuracy of mobile
Infrastruct. Eng., 2014, 29, (7), pp. 546–558 phone data using multi-temporal dasymetric interpolation’, Int. J. Geogr. Inf.
[107] Kumar, H., Singh, M.K., Gupta, M.: ‘Smart mobility: crowdsourcing Sci., 2017, 31, (8), pp. 1630–1651
solutions for smart transport system in smart cities context’. Proc. of the 11th [137] Eagle, N., Pentland, A.S.: ‘Eigenbehaviors: identifying structure in routine’,
Int. Conf. on Theory and Practice of Electronic Governance, Galway, Ireland, Behav. Ecol. Sociobiol., 2009, 63, (7), pp. 1057–1066
2018, pp. 482–488 [138] Pappalardo, L., Simini, F., Rinzivillo, S., et al.: ‘Returners and explorers
dichotomy in human mobility’, Nat. Commun., 2015, 6, p. 8166

492 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[139] Yan, X.Y., Wang, W.X., Gao, Z.Y., et al.: ‘Universal model of individual and [170] Baykurt, B.: ‘Redefining citizenship and civic engagement: political values
population mobility on diverse spatial scales’, Nat. Commun., 2017, 8, (1), p. embodied in fixmystreet.com’, AoIR Selected Papers of Internet Research,
1639 2011, 1, pp. 1–18
[140] Simini, F., González, M.C., Maritan, A., et al.: ‘A universal model for [171] Alliance, L.: ‘A technical overview of lora and lorawan’. White paper, 2015
mobility and migration patterns’, Nature, 2012, 484, (7392), p. 96 [172] Noreen, U., Bounceur, A., Clavier, L.: ‘A study of lora low power and wide
[141] Ren, Y., Ercsey Ravasz, M., Wang, P., et al.: ‘Predicting commuter flows in area network technology’. 2017 Int. Conf. on Advanced Technologies for
spatial networks using a radiation model based on temporal ranges’, Nat. Signal and Image Processing (ATSIP), Fez, Morocco, 2017, pp. 1–6
Commun., 2014, 5, p. 5347 [173] Wixted, A.J., Kinnaird, P., Larijani, H., et al.: ‘Evaluation of lora and lorawan
[142] Tomaras, D., Boutsis, I., Kalogeraki, V.: ‘Modeling and predicting bike for wireless sensor networks’. 2016 IEEE SENSORS, Orlando, FL, USA,
demand in large city situations’. 2018 IEEE Int. Conf. on Pervasive 2016, pp. 1–3
Computing and Communications (PerCom), Athens, Greece, 2018, pp. 1–10 [174] Dai, Z., Liu, X.C., Chen, Z., et al.: ‘A predictive headway-based bus-holding
[143] Starzynska, B., Kujawinska, A., Grabowska, M., et al.: ‘Requirements strategy with dynamic control point selection: A cooperative game theory
elicitation of passengers with reduced mobility for the design of high quality, approach’, Transp. Res. B, Methodol., 2019, 125, pp. 29–51
accessible and inclusive public transport services’, Manage. Production Eng. [175] Chawla, S., Zheng, Y., Hu, J.: ‘Inferring the root cause in road traffic
Rev., 2015, 6, (3), pp. 70–76 anomalies’. 2012 IEEE 12th Int. Conf. on Data Mining, 2012, pp. 141–150
[144] Ferrari, L., Berlingerio, M., Calabrese, F., et al.: ‘Improving the accessibility [176] Kuang, W., An, S., Jiang, H.: ‘Detecting traffic anomalies in urban areas using
of urban transportation networks for people with disabilities’, Transp. Res. C, taxi gps data’, Math. Probl. Eng., 2015, 2015 pp. 1–13doi:
Emerg. Technol., 2014, 45, pp. 27–40 [177] Pang, L.X., Chawla, S., Liu, W., et al.: ‘On detection of emerging anomalous
[145] Shrestha, B., Millonig, A., Hounsell, N., et al.: ‘Review of public transport traffic patterns using gps data’, Data Knowl. Eng., 2013, 87, pp. 357–373
needs of older people in european context’, J. Popul. Ageing., 2017, 10, (4), [178] Huang, C., Wu, X.: ‘Discovering road segment-based outliers in urban traffic
pp. 343–361 network’. 2013 IEEE Globecom Workshops (GC Wkshps), Atlanta, GA,
[146] May, A., Parker, C.J., Taylor, N., et al.: ‘Evaluating a concept design of a USA., 2013, pp. 1350–1354
crowd-sourced ‘mashup'providing ease-of-access information for people with [179] Duan, Z., Lei, Z., Zhang, M., et al.: ‘Understanding multiple days metro
limited mobility’, Transp. Res. C, Emerg. Technol., 2014, 49, pp. 103–113 travel demand at aggregate level’, IET Intell. Transp. Syst., 2018, 13, (5), pp.
[147] Hammon, L., Hippner, H.: ‘Crowdsourcing’, Business Inf. Syst. Eng., 2012, 4, 756–763
(3), pp. 163–166 [180] Chu, K.F., Lam, A.Y., Li, V.O.: ‘Deep multi-scale convolutional lstm network
[148] Mirri, S., Prandi, C., Salomoni, P., et al.: ‘On combining crowdsourcing, for travel demand and origin-destination predictions’, IEEE Trans. Intell.
sensing and open data for an accessible smart city’. 2014 Eighth Int. Conf. on Transp. Syst., 2019, pp. 1–14
Next Generation Mobile Apps, Services and Technologies, Oxford, UK., [181] Li, W., Cao, J., Guan, J., et al.: ‘A general framework for unmet demand
2014, pp. 294–299 prediction in on-demand transport services’, IEEE Trans. Intell. Transp. Syst.,
[149] Mobasheri, A., Deister, J., Dieterich, H.: ‘Wheelmap: the wheelchair 2019, 20, pp. 2820–2830
accessibility crowdsourcing platform’, Open Geospatial Data Softw. [182] Kerkman, K., Martens, K., Meurs, H.: ‘Predicting travel flows with spatially
Standards, 2017, 2, (1), p. 27 explicit aggregate models: on the benefits of including spatial dependence in
[150] Corsar, D., Edwards, P., Nelson, J., et al.: ‘Linking open data and the crowd travel demand modeling’, Transp. Res. A, Policy Pract., 2018, 118, pp. 68–88
for real-time passenger information’, J. Web Semantics, 2017, 43, pp. 18–24 [183] Jiang, S., Chen, W., Li, Z., et al.: ‘Short-term demand prediction method for
[151] Chaudhary, M., Bansal, A., Bansal, D., et al.: ‘Finding occupancy in buses online car-hailing services based on a least squares support vector machine’,
using crowdsourced data from smartphones’. Proc. of the 17th Int. Conf. on IEEE Access, 2019, 7, pp. 11882–11891
Distributed Computing and Networking, Singapore, 2016, p. 35 [184] Mostafaeipour, A., Goli, A., Qolipour, M.: ‘Prediction of air travel demand
[152] Lekshmy, V.V., Corsar, D., Joshi, S., et al.: ‘Utilising the crowd to support a using a hybrid artificial neural network (ann) with bat and firefly algorithms:
smart mobility system’, Transportation Research Board 96th Annual Meeting, a case study’, J. Supercomput., 2018, 74, (10), pp. 5461–5484
Washington DC, USA., 2017 [185] Liang, X., Wang, G., Min, M.R., et al.: ‘A deep spatio-temporal fuzzy neural
[153] Cohen, B., Kietzmann, J.: ‘Ride on! mobility business models for the sharing network for passenger demand prediction’. Proc. of the 2019 SIAM Int. Conf.
economy’, Organ. Environ., 2014, 27, (3), pp. 279–296 on Data Mining, 2019, pp. 100–108
[154] Wirtz, J., Tang, C.: ‘Uber: competing as market leader in the us versus being a [186] Ma, W., Qian, Z.S.: ‘Statistical inference of probabilistic origin-destination
distant second in China’. Services Marketing: People Technology Strategy, demand using day-to-day traffic data’, Transp. Res. C, Emerg. Technol., 2018,
Singapore, 2016, pp. 626–632 88, pp. 227–256
[155] Hong, L., Yan, Y., Ouyang, M., et al.: ‘Vulnerability effects of passengers’ [187] Gao, Y., Jiang, D., Xu, Y.: ‘Optimize taxi driving strategies based on
intermodal transfer distance preference and subway expansion on reinforcement learning’, Int. J. Geogr. Inf. Sci., 2018, 32, (8), pp. 1677–1696
complementary urban public transportation systems’, Reliab. Eng. Syst. Saf., [188] Lo, C.L., Chen, C.H., Hu, J.L., et al.: ‘A fuel-efficient route plan method
2017, 158, pp. 58–72 based on game theory’, J. Internet Technol., 2019, 20, (3), pp. 925–932
[156] Alabbasi, A., Ghosh, A., Aggarwal, V.: ‘Deeppool: distributed model-free [189] Qi, X., Wu, G., Boriboonsomsin, K., et al.: ‘Data-driven decomposition
algorithm for ride-sharing using deep reinforcement learning’, arXiv preprint, analysis and estimation of link-level electric vehicle energy consumption
arXiv:190303882, 2019 under real-world traffic conditions’, Transp. Res. D, Transp. Environ., 2018,
[157] Nehls, K.: ‘Crowdsourcing’ (Taylor & Francis, Philadelphia, PA, USA., 2015) 64, pp. 36–52
[158] Hu, X., Li, X., Ngai, E.C.H., et al.: ‘Multidimensional context-aware social [190] Philip, A.M., Ramadurai, G., Vanajakshi, L.: ‘Urban arterial travel time
network architecture for mobile crowdsensing’, IEEE Commun. Mag., 2014, prediction using support vector regression’, Transp. Dev. Econ., 2018, 4, (1),
52, (6), pp. 78–87 p. 7
[159] Ning, Z., Xia, F., Ullah, N., et al.: ‘Vehicular social networks: enabling smart [191] Pamucar, D., Cirovic, G.: ‘Vehicle route selection with an adaptive neuro
mobility’, IEEE Commun. Mag., 2017, 55, (5), pp. 16–55 fuzzy inference system in uncertainty conditions’, Decision Making: Appl.
[160] Ning, Z., Hu, X., Chen, Z., et al.: ‘A cooperative quality-aware service access Manage. Eng., 2018, 1, (1), pp. 13–37
system for social internet of vehicles’, IEEE Internet Things J., 2018, 5, (4), [192] Liang, X., Du, X., Wang, G., et al.: ‘Deep reinforcement learning for traffic
pp. 2506–2517 light control in vehicular networks’, arXiv preprint, arXiv:180311115, 2018
[161] Hu, X., Deng, J., Zhao, J., et al.: ‘Safedj: a crowd-cloud codesign approach to [193] Bui, K.H.N., Jung, J.J.: ‘Cooperative game-theoretic approach to traffic flow
situation-aware music delivery for drivers’, ACM Trans. Multimed. Comput., optimization for multiple intersections’, Comput. Electr. Eng., 2018, 71, pp.
Commun. Appl. (TOMM), 2015, 12, (1s), p. pp. 1–24 1012–1024
[162] Hu, X., Cheng, J., Zhou, M., et al.: ‘Emotion-aware cognitive system in [194] Dong, N., Huang, H., Zheng, L.: ‘Support vector machine in crash prediction
multi-channel cognitive radio ad hoc networks’, IEEE Commun. Mag., 2018, at the level of traffic analysis zones: assessing the spatial proximity effects’,
56, (4), pp. 180–187 Accident Anal. Prev., 2015, 82, pp. 192–198
[163] Cui, Z., Henrickson, K., Ke, R., et al.: ‘Traffic graph convolutional recurrent [195] Quek, C., Pasquier, M., Lim, B.B.S.: ‘Pop-traffic: A novel fuzzy neural
neural network: a deep learning framework for network-scale traffic learning approach to road traffic analysis and prediction’, IEEE Trans. Intell. Transp.
and forecasting’, arXiv preprint, arXiv:180207007, 2018 Syst., 2006, 7, (2), pp. 133–146
[164] Jain, V., Sharma, A., Subramanian, L.: ‘Road traffic congestion in the [196] Ruhhammer, C., Baumann, M., Protschky, V., et al.: ‘Automated intersection
developing world’. Proc. of the 2nd ACM Symp. on Computing for mapping from crowd trajectory data’, IEEE Trans. Intell. Transp. Syst., 2017,
Development. (ACM, Atlanta, GA, USA., 2012, p. 11 18, (3), pp. 666–677
[165] Dangi, V., Parab, A., Pawar, K., et al.: ‘Image processing based intelligent [197] Li, L., Lv, Y., Wang, F.Y.: ‘Traffic signal timing via deep reinforcement
traffic controller’, Undergraduate Acad. Res. J. (UARJ), 2012, 1, (1), pp. 13– learning’, IEEE/CAA J. Autom. Sin., 2016, 3, (3), pp. 247–254
17 [198] Wang, M., Hoogendoorn, S.P., Daamen, W., et al.: ‘Game theoretic approach
[166] Liu, W., Zheng, Y., Chawla, S., et al.: ‘Discovering spatio-temporal causal for predictive lane-changing and car-following control’, Transp. Res. C,
interactions in traffic data streams’. Proc. of the 17th ACM SIGKDD Int. Emerg. Technol., 2015, 58, pp. 73–92
Conf. on Knowledge Discovery and Data Mining, San Diego, CA, USA., [199] Krol, J., Anvari, B., Lot, R.: ‘Low-dimensional estimation and prediction
2011, pp. 1010–1018 framework for description of the oscillatory traffic dynamics’, EPiC Series
[167] Rahman, H.A., Mart, J.R., Srivastava, K.D.: ‘Road traffic forecasting through Comput., 2019, 62, pp. 78–79
simulation and live gps-feed from intervehicle networks’. 2012 IEEE Global [200] Pattanaik, V., Singh, M., Gupta, P., et al.: ‘Smart real-time traffic congestion
Humanitarian Technology Conf., 2012, pp. 36–40 estimation and clustering technique for urban vehicular roads’. 2016 IEEE
[168] Halawa, K., Bazan, M., Ciskowski, P., et al.: ‘Road traffic predictions across Region 10 Conf. (TENCON), 2016, pp. 3420–3423
major city intersections using multilayer perceptrons and data from multiple [201] Cheng, A., Jiang, X., Li, Y., et al.: ‘Multiple sources and multiple measures
intersections located in various places’, IET Intell. Transp. Syst., 2016, 10, based traffic flow prediction using the chaos theory and support vector
(7), pp. 469–475 regression method’, Physica A, 2017, 466, pp. 422–434
[169] Chan, K.Y., Dillon, T.S.: ‘On-road sensor configuration design for traffic flow [202] Li, Z., Zheng, H., Feng, X.: ‘3d convolutional generative adversarial networks
prediction using fuzzy neural networks and taguchi method’, IEEE Trans. for missing traffic data completion’. 2018 10th Int. Conf. on Wireless
Instrum. Meas., 2013, 62, (1), pp. 50–59 Communications and Signal Processing (WCSP), Hangzhou, People's
Republic of China, 2018, pp. 1–6

IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494 493
© The Institution of Engineering and Technology 2020
17519578, 2020, 6, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2019.0321 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [18/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[203] Laña, I., Olabarrieta, I.I., Vélez, M., et al.: ‘On the imputation of missing data Annual Int. Conf. on Mobile Systems, Applications, and Services, Taipei,
for road traffic forecasting: new insights and novel techniques’, Transp. Res. Taiwan, 2013, pp. 13–26
C, Emerg. Technol., 2018, 90, pp. 18–33 [225] Patra, S., Arnanz, J.H., Calafate, C.T., et al.: ‘Eyes: a novel overtaking
[204] Zhang, P., Xiong, L., Yu, Z., et al.: ‘Reinforcement learning-based end-to-end assistance system for vehicular networks’. Int. Conf. on Ad-Hoc Networks
parking for automatic parking system’, Sensors, 2019, 19, (18), p. 3996 and Wireless, 2015, pp. 375–389
[205] Zhang, L., Li, Y.: ‘A game-theoretic approach to optimal scheduling of [226] Hadiwardoyo, S.A., Patra, S., Calafate, C.T., et al.: ‘An android its driving
parking-lot electric vehicle charging’, IEEE Trans. Veh. Technol., 2015, 65, safety application based on vehicle-to-vehicle (v2v) communications’. 2017
(6), pp. 4068–4078 26th Int. Conf. on Computer Communication and Networks (ICCCN),
[206] Wu, Y.F., Tran, V.H., Huang, C.C.: ‘Semi-supervised and multi-task learning Vancouver, BC, Canada, 2017, pp. 1–6
for on-street parking space status inference’. 2019 Int. Conf. on Multimedia [227] Tornell, S.M., Calafate, C.T., Cano, J.C., et al.: ‘Evaluating the feasibility of
Analysis and Pattern Recognition (MAPR), 2019, pp. 1–6 using smartphones for its safety applications’. 013 IEEE 77th Vehicular
[207] Vu, H.T., Huang, C.C.: ‘Parking space status inference upon a deep cnn and Technology Conf. (VTC Spring), 2013, pp. 1–5
multi-task contrastive network with spatial transform’, IEEE Trans. Circuits [228] Xu, Z., Liu, Y., Yen, N., et al.: ‘Crowdsourcing based description of urban
Syst. Video Technol., 2018, 29, (4), pp. 1194–1208 emergency events using social media big data’, IEEE Trans. Cloud Comput.,
[208] Zheng, Y., Rajasegarar, S., Leckie, C.: ‘Parking availability prediction for 2016, pp. 1–1
sensor-enabled car parks in smart cities’. 2015 IEEE Tenth Int. Conf. on [229] Carisi, R., Giordano, E., Pau, G., et al.: ‘Enhancing in vehicle digital maps via
Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), gps crowdsourcing’. 2011 Eighth Int. Conf. on Wireless On-Demand Network
Singapore, 2015, pp. 1–6 Systems and Services, 2011, pp. 27–34
[209] Leclercq, L., Sénécat, A., Mariotte, G.: ‘Dynamic macroscopic simulation of [230] Hu, S., Su, L., Liu, H., et al.: ‘Smartroad: a crowd-sourced traffic regulator
on-street parking search: a trip-based approach’, Transp. Res. B, Methodol., detection and identification system’. Proc. of the 12th Int. Conf. on
2017, 101, pp. 268–282 Information Processing in Sensor Networks, Philadelphia, PA, USA., 2013,
[210] Nakrani, N., Joshi, M.: ‘An intelligent fuzzy based hybrid approach for pp. 331–332
parallel parking in dynamic environment’, Procedia Comput. Sci., 2018, 133, [231] Aly, H., Basalamah, A., Youssef, M.: ‘Map + +: a crowd-sensing system for
pp. 82–91 automatic map semantics identification’. 2014 Eleventh Annual IEEE Int.
[211] Li, R., Wang, W., Chen, Y., et al.: ‘An end-to-end fully automatic bay parking Conf. on Sensing, Communication, and Networking (SECON), Singapore,
approach for autonomous vehicles’. ASME 2018 Dynamic Systems and 2014, pp. 546–554
Control Conf. (American Society of Mechanical Engineers Digital Collection, [232] Liu, S., Watterson, M., Tang, S., et al.: ‘High speed navigation for quadrotors
Atlanta, GA, USA, 2018 with limited onboard sensing’. 2016 IEEE Int. Conf. on Robotics and
[212] Wen, J., Zhao, J., Jaillet, P.: ‘Rebalancing shared mobility-on-demand Automation (ICRA), 2016, pp. 1484–1491
systems: A reinforcement learning approach’. 2017 IEEE 20th Int. Conf. on [233] Chen, J., Liu, T., Shen, S.: ‘Online generation of collision-free trajectories for
Intelligent Transportation Systems (ITSC), 2017, pp. 220–225 quadrotor flight in unknown cluttered environments’. 2016 IEEE Int. Conf. on
[213] Zhou, Z., Yu, H., Xu, C., et al.: ‘Dependable content distribution in d2d-based Robotics and Automation (ICRA), 2016, pp. 1476–1483
cooperative vehicular networks: A big data-integrated coalition game [234] Rusu, R.B., Cousins, S.: ‘Point cloud library (pcl)’. 2011 IEEE Int. Conf. on
approach’, IEEE Trans. Intell. Transp. Syst., 2018, 19, (3), pp. 953–964 Robotics and Automation, 2011, pp. 1–4
[214] Wang, H., Liu, R.P., Ni, W., et al.: ‘Vanet modeling and clustering design [235] Hornung, A., Wurm, K.M., Bennewitz, M., et al.: ‘Octomap: an efficient
under practical traffic, channel and mobility conditions’, IEEE Trans. probabilistic 3d mapping framework based on octrees’, Auton. Robots, 2013,
Commun., 2015, 63, (3), pp. 870–881 34, (3), pp. 189–206
[215] Ghouti, L.: ‘Mobility prediction in mobile ad hoc networks using neural [236] Dey, D., Shankar, K.S., Zeng, S., et al.: ‘Vision and learning for deliberative
learning machines’, Simul. Modelling Pract. Theory, 2016, 66, pp. 104–121 monocular cluttered flight’. in: ‘Field and service robotics’ (Springer, Cham,
[216] Salazar, M., Rossi, F., Schiffer, M., et al.: ‘On the interaction between Switzerland, 2016), pp. 391–409
autonomous mobility-on-demand and public transportation systems’. 2018 [237] Lopez, B.T., How, J.P.: ‘Aggressive 3-d collision avoidance for high-speed
21st Int. Conf. on Intelligent Transportation Systems (ITSC), 2018, pp. 2262– navigation’. 2017 IEEE Int. Conf. on Robotics and Automation (ICRA),
2269 Singapore, 2017, pp. 5759–5765
[217] Soleymani, S.A., Abdullah, A.H., Zareei, M., et al.: ‘A secure trust model [238] Bates, L.: ‘Procedural justice and road policing: is it important’. Australasian
based on fuzzy logic in vehicular ad hoc networks with fog computing’, IEEE Road Safety Research, Melbourne, Australia: Policing and Education Conf.,
Access, 2017, 5, pp. 15619–15629 Melbourne, Australia, Available at https://research-repository.griffith.edu.au/
[218] Li, R.C., Tee, T.J.C.: ‘A reverse logistics model for recovery options of e- bitstream/handle/10072/67711/99597_1.pdf, 2014
waste considering the integration of the formal and informal waste sectors’, [239] Stanojevic, P., Jovanovic, D., Lajunen, T.: ‘Influence of traffic enforcement on
Procedia – Soc. Behav. Sci., 2012, 40, pp. 788–816 the attitudes and behavior of drivers’, Accident Anal. Prev., 2013, 52, pp. 29–
[219] Chen, C., Pan, S., Wang, Z., et al.: ‘Using taxis to collect citywide e- 38
commerce reverse flows: a crowdsourcing solution’, Int. J. Prod. Res., 2017, [240] Al Adhoobi, M.S., Jose, M., Singh, A.V.: ‘A review of different road safety
55, (7), pp. 1833–1844 measures and enforcements’. 2017 6th Int. Conf. on Reliability, Infocom
[220] Giret, A., Carrascosa, C., Julian, V., et al.: ‘A crowdsourcing approach for Technologies and Optimization (Trends and Future Directions)(ICRITO),
sustainable last mile delivery’, Sustainability, 2018, 10, (12), p. 4563 Noida, India, 2017, pp. 691–697
[221] Kim, Y., Ghorpade, A., Zhao, F., et al.: ‘Activity recognition for a smartphone [241] Anda, C., Erath, A., Fourie, P.J.: ‘Transport modelling in the age of big data’,
and web-based human mobility sensing system’, IEEE Intell. Syst., 2018, 33, Int. J. Urban Sci., 2017, 21, (Supp1), pp. 19–42
(4), pp. 5–23 [242] Devillaine, F., Munizaga, M., Trépanier, M.: ‘Detection of activities of public
[222] Stitzel, J.D., Weaver, A.A., Talton, J.W., et al.: ‘An injury severity-, time transport users by analyzing smart card data’, Transp. Res. Rec., 2012, 2276,
sensitivity-, and predictability-based advanced automatic crash notification (1), pp. 48–55
algorithm improves motor vehicle crash occupant triage’, J. Am. Coll. Surg., [243] Huang, L., Li, Q., Yue, Y.: ‘Activity identification from gps trajectories using
2016, 222, (6), pp. 1211–1219 spatial temporal pois’ attractiveness’. Proc. of the 2nd ACM SIGSPATIAL
[223] Meseguer, J.E., Calafate, C.T., Cano, J.C., et al.: ‘Drivingstyles: a smartphone Int. Workshop on Location Based Social Networks, San Jose, CA, USA.,
application to assess driver behavior’. 2013 IEEE Symp. on Computers and 2010, pp. 27–30
Communications (ISCC), 2013, pp. 000535–000540 [244] Hafezi, M.H., Liu, L., Millward, H.: ‘A time-use activity-pattern recognition
[224] You, C.W., Lane, N.D., Chen, F., et al.: ‘Carsafe app: alerting drowsy and model for activity-based travel demand modeling’, Transportation. (Amst),
distracted drivers using dual cameras on smartphones’. Proc. of the 11th 2019, 46, (4), pp. 1369–1394

494 IET Intell. Transp. Syst., 2020, Vol. 14 Iss. 6, pp. 480-494
© The Institution of Engineering and Technology 2020

You might also like