A Semantic Agent-Based Safety Model For Alzheimer'sPatients For Driving

Innovative Computing Review (ICR)
Volume 1 Issue 1, Summer 2021

Homepage: https://journals.umt.edu.pk/index.php/icr
Article: A Semantic Agent-Based Safety Model for Alzheimer’s Journal QR

Patients for Driving
Author(s): Yasir Sultan
The School of Computing and Built Environment Birmingham City University Article QR
Affiliation: Birmingham United Kingdom
Received: April 13, 2021

Article History: Revised: May 26, 2021
Accepted: June 30, 2021
Available Online: June 30, 2021
S. Yasir, “A Semantic Agent-based safety model for Alzheimer’s

Citation: patients for driving”, Innov Comput Rev, vol. 1, no. 1, pp. 60–74,
2021.
Copyright
Information: This article is open access and is distributed under the terms of
Creative Commons Attribution 4.0 International License
A publication of the
School of Systems and Technology
University of Management and Technology, Lahore, Pakistan

A Semantic Agent-Based Safety Model for Alzheimer’s
Patients for Driving
Yasir Sultan1
ABSTRACT: Alzheimer’s disease I. INTRODUCTION

drastically affects the human Alzheimer’s disease, commonly
neurological system, causing the death known as AD, is an exceptionally
of neurons and synapses in cerebral hazardous neurological illness. Early
cortex and other regions. Consequently, indications of AD include a momentary
patients with Alzheimer’s disease loss of memory, confusion, critical
require 24/7 care and assistance with issues with language, time
daily life activities. This study uses an disorientation, withdrawal from work
ontology-based semantic model to and from different social exercises,
assist the Alzheimer’s patients through enthusiastic disinterest, and various
activity recognition in the focused other conditions. Alzheimer’s disease
scenario of driving. The safety model affects the neurological framework as it
presented in this paper alerts the user if is a neurodegenerative disease.
anything dangerous is about to happen Moreover, it generally affects the
when they enter a car or during driving. lifestyle of the patients as well. This
It should be noted that the model is
paper distinguishes and assesses the
specifically designed for the patients impact of Alzheimer’s disease on other
suffering from a mild form of life function [1].
Alzheimer’s disease. We have used
semantics to define the objects and Alzheimer’s disease exists even in
environment for the patients. After the 21st century because the assessments
proper monitoring and evaluation, our of the scientists concerning the
model suggests a proper course of fundamental drivers of this disease
action based on semantic reasoning. fluctuates undeniably [2]. For example,
most of the scientists recommend that
INDEX TERMS: activities of daily living, hereditary qualities pose a significant
cognitive impairment, domain risk factor that invigorates the turn of
ontologies, semantic reasoning events and movement of Alzheimer’s
disease. Moreover, according to
1
Yasir Sultan is with The School of Computing and Built Environment Birmingham City
University Birmingham United Kingdom. Yasir Sultan is the corresponding author and available
at yasir.sultan2@mail.bcu.ac.uk

61
A Semantic Agent-Based Safety Model for…
previous research, 80% of factors driving ability of these patients. Data

leading to Alzheimer’s are characterized was collected from 28 drivers with
as hereditary up to 80% [3]. dementia and 28 healthy elderly
patients. After proper research and
Alzheimer’s disease drastically
evaluation, the researchers concluded
affects the neurological system of
that there is a strong correlation between
human beings, causing the death of
driving performance and the
neurons and synapses in cerebral cortex
performance based on specific
and other regions. The patients of this
cognitive tests [4].
disease require 24/7 care, so that they
may live their life properly. For them, A common trait of Alzheimer’s or
living their life independently becomes intellectually impaired patients is their
a very difficult task. Even inclination to experience spatial and
accomplishing a simple task like driving fleeting confusion, which causes them
also becomes a challenging ordeal for to get lost. To avoid this problem,
these patients. caregivers should utilize GPS gadgets to
locate and keep tabs on the patients in a
For this reason, we aim to propose
discrete manner.
an ontology-based semantic model that
assists patients before and during Based on the data taken from
driving. Through this model, the administrative healthcare projects and
patients’ activities are recognized and organizations that observe Alzheimer’s
based on standard danger levels scales patients, and also from geo-area
be assessed and if any the patient will be specialist co-ops, the current research
notified. encourages the use of GPS gadgets by
specialized caregivers to monitor
II. RELATED WORK
Alzheimer’s and intellectually impaired
Alzheimer’s disease remains an area patients. Thus, this paper suggests the
of interest for researchers because of the use of GPS tracker for Alzheimer’s
number of people affected by it over the patients 24/7, so that their whereabouts
years. Investigation concerning can be monitored properly.
Alzheimer's are made over the variety
It has been noted that a large number
of factors, symptoms and impacts for
of Alzheimer’s patients suffer from
the diagnostic and treatment.
agitated behaviour, which may create a
The current paper focuses on the stressful work environment for their
people with mild dementia, for whom caregivers. A number of AI-based
driving is a high-risk activity. The main technologies have been designed to
aim of this project is to devise a manage this agitated behaviour in
cognitive test that can predict the Alzheimer’s patients, but it came
Innovative Computing Review

62
Sultan
inappropriate prediction models and Dementia care frameworks, hence the

misleading data representation became patients’ powerful conduct perceptions
a challenge in them. So, this paper (user’s area, objects used) should be
proposes an ontology-based coding investigated against the semantic
model for managing the agitated information available about their
behaviour of the Alzheimer’s patients. condition (disease history). The current
The study here came up with Dementia‐ paper reveals the capacity of semantic
Related Agitation Non‐Pharmacological innovations to reason based upon
Treatment Ontology (DRANPTO). The complex interrelated occasions that
method used by it is the Neon originate from the conduct checking
methodology [5]. Through this study, it sensors used to gather information [7-
has been deduced that DRANPTO can 10]. There has been a lot of work done
come up with better knowledge in the domain of ontologies using Big
representation and standardization. It Data [11], Blockchain [12], and drug
can successfully communicate with the interaction models [13, 14]. Also,
readers about the disease related ontologies are very important in the
medical domain concepts and their domain of Natural Language Processing
relationships [6]. (NLP) [15-18]
The current paper describes another III. EXAMPLE USAGE
methodology that depends on To exemplify the usage of our
propositional analytics to isolate model, we have used a driving scenario
complex occasions that can be in which the patient is suffering from a
interpreted through semantic thinking mild case of Alzheimer’s disease. The
derived from occasions that require data location for the scenario is a road and a
beyond semantic engine’s own room where equipment related to
understanding. Regression analysis was driving is kept. While driving, the
used to evaluate the response time and patient needs to carry home keys, car
the study concludes that response time keys, a mask, and sanitizer due to the
of the systems to identify and notify the present COVID-19 regulations.
danger level is tolerable.
Dividing the location into zones
This paper showcases the helps in the proper documentation of the
difficulties in creating a semantic-based patient’s activities and deducing danger
Dementia Care Decision Support
System used to check discretely a levels. In this study, there are two zones:
patient’s conduct. It should be noted a room where keys, masks, and sanitizer
that semantic-based methodologies are are kept and a road where the patient
appropriate when there are dynamic will drive.
behaviour observations as seen in

63
The scenario is of the patient going with Alzheimer’s patients to detect

to a nearby park for a walk. However, things. We also assume that the patient
the park is at least 5 miles away, so the is wearing a wearable device such as a
patient needs to drive to get there. watch. Our ontological approach to
activity recognition and danger
Now, for this hypothetical scenario,
assessment uses a scenario in which an
the patient needs to carry home keys, car
Alzheimer’s patient is driving a car and
keys, mask, sanitizer and a water bottle
it can be used to assist them with several
(optional). Our model analyses the
activities before and during driving. In
actions of the patient. It checks if the
this particular scenario, we need to
user has taken along the home keys, car
address challenges such as the speed of
keys, and a mask. If the patient has not
the vehicle, distance from the next car,
taken these items, the danger level rises.
and distance from a vehicle to the left or
As soon as the system gauges that the
right of the car, as well as traffic signals.
patient has taken all these things, the
For example, if the patient reaches a
danger level decreases. If the patient
traffic signal, then the sensor in the car
forgets to take along these items, an
should detect which traffic signal is on
alert is generated. Afterwards, when the
and which are off. An alert would be
patient sits in the car, another alert is
sent to the patient’s wearable device and
generated to let the patient know
also to the car, so the patient can be
whether the lights in the car are on or off
alerted through the car speakers.
based on the situation. Similarly,
another alert is prompted to check the IV. ANALYSIS OF STRATEGY
temperature in the car. Some alerts are AND TOOLS
generated to warn the patient when they The instruments utilized for the
exceed a certain speed limit, while execution of this scenario include
others are generated when the patient is Protégé, OWL, SWRL, RDFS, the Jena
near the traffic signals. Through our API, and Jess API. These make an
safety model, the patient is notified Action Driven Model and consequently,
about the traffic light and told to take an make our safety model an Action
action based on the corresponding Service Ontology.
traffic signal. Similarly, an alert can also
be generated to urge the patient to check A. Protégé
if the hand brake is down or not as soon Protégé is a free open-source stage
as they enter the car. created by the Stanford University for
A. Scenario to Address the development of area models and
information-based applications utilizing
For this analysis, we assume that the ontologies [19].
patient is using the sensors affiliated

64
Sultan
B. OWL F. Pellet
OWL or the Web Ontology Language Pellet is an open-source JAVA-
is a semantic web computational based semantic reasoner for OWL2. A
rationale-based language intended to semantic reasoner is a program that can
state information about concepts, derive sensible results from a bunch of
groups of concepts, and the declared realities and maxims.
relationships between concepts. It may With regard to this paper, Activity
be utilized to confirm the consistency of Driven Model is characterized as a
information patterns or to deduce semantic model that takes into
verifiable information (W3C, 2018) consideration user actions.
[20].
Activity Service Ontology is
C. RDFS characterized as formal and computable
RDFS or the Resource Description representation of real-life processes for
Framework Schema is an augmentation client activities that can be given as a
of the fundamental RDF vocabulary that Web Service.
takes into consideration information Utilizing the above devices and
demonstration. RDF itself is a standard ideas, a Significance Level Model is
model for information trade on the proposed that can be executed to
internet (W3C, 2018). It is used to support and help with action
communicate the rules for semantic web acknowledgment, give acknowledgment
given by SWRL or the Semantic Web of activities, and respond to them as
Rule Language [21]. needed. Utilizing the instrument’s
D. Jess accessibility, we can fabricate an OWL
model that can be used to portray the
Jess Rule Engine is a standard
environment around the occupant,
engine for the JAVA platform that can
express the connections between the
generate new information from pre-set
objects in the environment in RDFS, use
principles.
SWRL to characterize rules, and Pellet
E. Apache Jena to reason the outcomes. Apache Jena
Apache Jena is an open-source and the Jess Rule engine can be utilized
semantic web structure for JAVA that to make new principles and supplement
gives an API to separate information new axioms dependent on user conduct.
from and compose information to RDF For Pellet to comprehend our ontology,
diagrams. data about the associated knowledge to
patient entity is provided as Jess facts
functions and SWRL rules are
interpreted as Jess rules. The service,

65
further, deduces from provided health G., & Rinaldi) to utilize the existing
guidelines and add the inductions to the COA model; however, in our model we
Course of Action (COA) information also add different classes, object
base lastly control this through a user properties, data properties, and
interface. instantiation.
V. ONTOLOGY MODEL AND SYSTEM System architecture is designed by
ARCHITECT making the full use of Protégé’s features
of classification, consistency checking,
System architecture is planned by
and ontology testing. It acts as an
utilizing Protégé’s highlights of
extension of the underlying APIs adding
characterization, consistency checking,
classes to represent COA and danger
and ontology testing. It goes about as an
levels using RDFS and OWL DL.
augmentation of the basic APIs, adding
classes to speak to COA and predicting Consequently, we can support
danger levels utilizing RDFS and OWL inferencing with SWRL rules using the
and RDFS and SWRL, respectively. Drool Engine. We have created and
edited classes and their characteristics,
A. Evaluation of the Implemented
accessed reasoning engines, executed
Approach
queries and rules, added individuals and
We have created an ontology-based visualized relationships between
reasoning model to assist the patient in concepts through a customizable
the given scenario. We have used graphical user interface.
ontology reuse approach (Caldarola, E.
Fig. 1. COA model

66
Sultan
OWL is chosen because it allows us application independent core model that

to set property restrictions such as can be reused to support other danger
functional, inverse-functional, transitive, identification scenarios.
symmetric, asymmetric, reflexive, and I
Our ontology is organized as a
reflexive. These allow us to better hierarchy of classes, that is, objects
capture relationships between objects. which classify individuals (instances)
RDF defines the relation between with relational properties represented as
two concepts as a semantic triple based object properties and states classified as
presentation, which allows structured data properties.
and semi-structured data to be exposed, The relational properties and states
mixed, and shared across different of the objects capture the relevant
applications, regardless of the characteristics and describe the
difference between the underlying environment in terms of logical
schemas. concepts.
SWRL is also used to write rules in The recommended Course of Action
the form of an implication between an (COA) is sub-divided into three main
antecedent and consequent. In SWRL, if categories: phase, activity, and
the conditions in the antecedent hold, condition. COA depicts a plan to
then the conditions in the consequent undertake the activities of daily life
also holds. successfully and it is further divided
The COA model is built from into phases. There are many sequences
classes that represent the objects and of activities that must be performed for
users in the environment as more the successful completion of a task.
abstract concepts, such as danger level. Individual tasks that are not part of a
To clarity, the instance of the sequence are designated as activities. A
constructed COA with attached condition is a state of being that is a
guidelines, will be from now on referred prerequisite for a certain activity. For
to as objects in the domain ontological example, making a cup of tea has hot
context. While objects can change the water as a condition.
main features of the COA model, that According to our scenario, we add some
are designed to be separate from the classes and relationships as follow:
daily activity data to provide an

67
Fig. 2. Purposed concepts for the particular scenario

68
Sultan
Fig. 3. Class hierarchy

Firstly, we create classes then define the person will have a sensor on the
Object properties according to the shower tap to alert them if they have
scenario. Afterwards, we write SWRL used the shower or not. It will also alert
rules to get the desired result. For them if the shower has not been used for
example, the number of times a person more than 4 days to urge the user to take
has taken a shower can be taken as a a shower.
hypothetical scenario. In this scenario,

69
VI. SWRL RULES (PELLET REASONER TO INFERENCE KNOWLEDGE)
Fig. 4. SWRL
Following is the implementation of rules and we use pellet as a reasoner for
inferencing.
Fig. 5. Asserted value by the sensor to the wearable device of the

user that the signal is green
Fig. 6. Inferred value by the sensor to the wearable device to alert the user
Fig. 7. Asserted value by the sensor to the wearable device of

the user that the signal is red

70
Sultan
Fig. 9. Asserted value by the sensor to the wearable device

of the user that the signal is orange
We write SWRL in ROWL and use The following figure displays our whole
Pellet Reasoner to get the desired result. system.
Fig. 11. Graph of relationships between purposed class and instances

71
Fig. 12. Validation of Ontology by reasoner

The fig 12 displays the validation of recognition. Future work should also be
the system by running the reasoner to based on patients with mild Alzheimer.
check the consistency of our model. The extended research area is
optimization of computerized services
VII. CONCLUSION AND FUTURE
provided in a smart
WORK
This paper discusses a unique and a very REFERENCES
high-end risk scenario in which an [1] M. E. Mlinac and M. C. Feng,
Alzheimer’s patient is driving. We have "Assessment of activities of daily
taken this scenario because driving is an living, self-care, and independence,"
important part of an independent Archives of Clinical Neuropsychology,
lifestyle; however, Alzheimer’s patients vol. 31, no. 6, pp. 506-516, 2016.
struggle with this activity on a daily https://doi.org/10.1093/arclin/acw049
basis. Our aim is to assist Alzheimer’s
patients complete this activity in an [2] J. D. Kasper, "The disproportionate
uncomplicated/easier manner. The impact of dementia on family and
research proposed an ontology-based unpaid caregiving to older adults.,"
semantic model, where all activities of Health Affairs, vol. 34, no. 10, pp.
the patient are monitored, dangers are 1642-1649, 2015.
deduced, and a proper and safe course
of action is suggested for the patient. In [3] K. Jekel et al., "Mild cognitive
this model, we have used free source impairment and deficits in
platforms and sensors that are more instrumental activities of daily
powerful and affordable for the patients. living: a systematic review,"
Though we have strived to determine Alzheimer's Research & Therapy,
major events that may occur in this vol. 7, no. 1, pp. 1-17, 2015.
scenario, more work should be done in https://doi.org/10.1186/s13195-
this regard. Future studies should focus 015-0099-0
on the same thorough activity

72
Sultan
[4] Alzheimer's Research UK, [9] W3C Semantic Web. (11

"Dementia Statistics Hub," 2018. December 2013). Web Ontology
[Online]. Available: Language. [Online]. Available:
https://www.dementiastatistics.org https://www.w3.org/OWL/
/statistics/numbers-of-people-in- [10] W3C Semantic Web. (15 March
the-uk/ 2014). Resource Description
[5] Alzheimer's Society, "Understanding Framework. [Online]. Available:
and supporting a person with https://www.w3.org/RDF/
dementia," 2018. [Online]. [11] Horrocks, P. F. Patel-Schneider, H.
Available: Boley, S. Tabet, and B. Grosof,
https://www.alzheimers.org.uk/get "SWRL: A semantic web rule
-support/help-dementia-care/ language combining OWL and
understanding-supporting-person- RuleML," W3C Member submission,
dementia vol. 21, pp. 79, 2004.
[6] C. K. Andersen, K. U. Wittrup- [12] E. J. Friedman-Hill, "Jessrules,"
Jensen, A. Lolk, K. Andersen, and 2018. [Online]. Available:
P. J. H. Kragh-Sørensen, "Ability https://www.jessrules.com/jess/docs
to perform activities of daily living /71/
is the main factor affecting quality
of life in patients with dementia," [13] Apache Jena. (2019). Available:
Health and Quality of Life https://jena.apache.org/
Outcomes, vol. 2, no. 1, pp. 2-52, [14] A.-R. Tawil, "A Semantic-Based
2004. https://doi.org/10.1186/1477- Safety Model for Alzheimer
7525-2-52 Patients," Unpublished, 2018.
[7] X. Huang, J. Yi, X. Zhu, and S. J. [15] H. Ishii, K. Kimino, M. Aljehani,
S. Chen, "A semantic approach N. Ohe, and M. Inoue, "An early
with decision support for safety detection system for dementia
service in smart home using the M2 M/IoT platform,"
management," Sensors, vol. 16, no. Procedia Computer Science, no.
8, pp. 1224, 2016. https://doi.org/ 96, pp. 1332-1340, 2016.
10.3390/s16081224 https://doi.org/10.1016/j.procs.
[8] Protage Stanford University. (2018). 2016.08.178
Available: [16] F. Zekri, R. Bouaziz, and E. Turki,
https://protege.stanford.edu/products. php "A fuzzy-based ontology for

73
Alzheimer's disease decision approach to activity recognition in

support," in 2015 IEEE International smart homes," IEEE Transactions
Conference on Fuzzy Systems on Knowledge and Data
(FUZZ-IEEE), 2015, pp. 1-6: Engineering, no. 24, pp. 961-974,
IEEE. https://doi.org/10.1109/ 2012.
FUZZ-IEEE.2015.7337922 [20] S. Zhang, P. McCullagh, and H.
[17] K.-Y. Lam, N. W.-H. Tsang, S. Zheng, "Situation Awareness
Han, J. K.-Y. Ng, S.-W. Tam, and Inferred from Posture Transition
A. Nath, "SmartMind: activity and Location; derived from smart
tracking and monitoring for phone and smart home sensors,"
patients with Alzheimer's disease," IEEE Transactions on Human-
in 2015 IEEE 29th International Machine Systems, vol. 47, no. 6,
Conference on Advanced pp. 814-821, 2017. https://doi.org/
Information Networking and 10.1109/THMS.2017.2693238
Applications, 2015, pp. 453-460: [21] K. Gayathri and K. S.
IEEE. Easwarakumar, "Intelligent
[18] S. M. Zhang, "An ontology-based decision support system for
context-aware approach for dementia care through smart
behaviour analysis," Activity home," Procedia Computer
Recognition in Pervasive Intelligent Science, vol. 93, pp. 947-955,
Environments, pp. 127-148), 2011. 2016.
https://doi.org/10.1016/j.procs.201
[19] L. Chen, C. D. Nugent, and K.
6.07.281
Wang, "A knowledge-driven

74

A Semantic Agent-Based Safety Model For Alzheimer'sPatients For Driving

Uploaded by

Copyright:

Available Formats

You might also like

A Semantic Agent-Based Safety Model For Alzheimer'sPatients For Driving

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Semantic Agent-Based Safety Model For Alzheimer'sPatients For Driving

Uploaded by

Copyright:

Available Formats

Innovative Computing Review (ICR)

Volume 1 Issue 1, Summer 2021

Article: A Semantic Agent-Based Safety Model for Alzheimer’s Journal QR

Author(s): Yasir Sultan

Received: April 13, 2021

S. Yasir, “A Semantic Agent-based safety model for Alzheimer’s

School of Systems and Technology

University of Management and Technology, Lahore, Pakistan

ABSTRACT: Alzheimer’s disease I. INTRODUCTION

School of Systems and Technology

previous research, 80% of factors driving ability of these patients. Data

Innovative Computing Review

inappropriate prediction models and Dementia care frameworks, hence the

School of Systems and Technology

The scenario is of the patient going with Alzheimer’s patients to detect

Innovative Computing Review

School of Systems and Technology

Fig. 1. COA model

Innovative Computing Review

OWL is chosen because it allows us application independent core model that

School of Systems and Technology

Fig. 2. Purposed concepts for the particular scenario

Innovative Computing Review

Fig. 3. Class hierarchy

School of Systems and Technology

VI. SWRL RULES (PELLET REASONER TO INFERENCE KNOWLEDGE)

Fig. 5. Asserted value by the sensor to the wearable device of the

Fig. 7. Asserted value by the sensor to the wearable device of

Innovative Computing Review

Fig. 9. Asserted value by the sensor to the wearable device

Fig. 11. Graph of relationships between purposed class and instances

School of Systems and Technology

Fig. 12. Validation of Ontology by reasoner

Innovative Computing Review

[4] Alzheimer's Research UK, [9] W3C Semantic Web. (11

School of Systems and Technology

Alzheimer's disease decision approach to activity recognition in

Innovative Computing Review

You might also like