=Paper=
{{Paper
|id=Vol-2326/paper8
|storemode=property
|title=Internet of Things Agents Diagnosis Architecture : Application to Healtcare IoT System
|pdfUrl=https://ceur-ws.org/Vol-2326/paper8.pdf
|volume=Vol-2326
|authors=Sofia Kouah,Ilhem Kitouni
|dblpUrl=https://dblp.org/rec/conf/icaase/KouahK18
}}
==Internet of Things Agents Diagnosis Architecture : Application to Healtcare IoT System==
https://ceur-ws.org/Vol-2326/paper8.pdf
Internet of Things Agents Diagnosis Architecture: Application to
Healthcare IoT System
Sofia KOUAH Ilham KITOUNI
RELA(CS)2 Laboratory, University of Larbi Ben M’Hidi, MISC Laboratory, Constantine 2-Abdelhamid Mehri
Oum El Bouaghi, Algeria University- Algeria
kouahs@yahoo.fr ilham.kitouni@univ-constantine2.dz
Abstract
concise integration of the physical world into
Internet of Things (IoT) is an advanced computer systems through network infrastructure. It
research area which provides deeper sees a huge number of internet enabled devices that
automation, analysis and integration of can network and communicate with each other and
physical world into computer systems with other web-enabled gadgets. IoT refers to a state
through network infrastructure. It allows where Things (e.g. objects, environments, vehicles
interaction and cooperation between a large and clothing) will have more and more information
variety of pervasive objects over wireless associated with them and may have the ability to
and wired connections, in order to achieve sense, communicate, network and produce new
specific goals. Since IoT systems present information” [Gil16].
several properties such as distribution, IoT systems are widely open, dynamic and need large
openness, interoperability and dynamicity, amount of communications, interactions and data
their developing presents a great challenge. exchanges that should be achieved in a coherent
For that reason, a generic, multi-layer and manner. An IoT system is consisting of heterogeneous
agents based architecture for IoT systems physical and virtual connected components, using
has been proposed, called IoTAA (Internet different languages, platforms and intelligent policies
of Things Agents Architecture). IoTAA is [Int15]. In fact, interaction allows limitless
composed of four layers; each layer possibilities for objects to control and access their
maintains and schedules special features, environment and compensates for a lack of self-
such as connectivity and communication sufficiency. Connected to the internet, things acquire
between things, local coordination, global intelligence since they tap into an exceedingly rich
coordination, domain dependent environment [Weg97]. According to these properties,
functionalities. On the other hand, Diagnosis the design domain of IoT systems is becoming
of the IoT systems is an important issue increasingly attractive; as several IoT functionalities
which aims to detect abnormalities from the have to be modeled. While, developing IoT system is
system desired behavior, identify the failure an innovative field for the imminent future of
causes and localize failure components. computing and communication, the development of
Accordingly, the paper aims to show the efficient IoT systems still facing many challenging
efficiency of IoTAA for modeling IoT issues. Several approaches have been proposed in the
diagnosis problem. The four layers are literature, among others [Kat08] [Kaz09] [Che10]
developed and furthers processing are [For12] [Aya12] [Cha18]. These approaches are
proposed. The proposed IoTAA based application domain dependent and do not take into
diagnosis architecture is illustrated by an account almost IoT systems functionalities and
example of healthcare monitoring. properties. In particular, they do not support the
intelligent feature which is important to manage and
Keywords – Internet of Things, IoTAA, Fault diagnosis, control properly the objects’ traffic flow and improve
Multi-agents systems, Healthcare monitoring, Arduino. decision making. Recently, an alternative approach
has been proposed, called IoTAA for Internet of
1. Introduction Things Agent Architecture [Kou18]. It focuses on the
Multi-Agents System (MAS) paradigm and on the
Nowadays, Internet of things (IoT) is becoming a principle of separating responsibilities. MAS is an
promising technology which can modernize and efficient paradigm which considers IoT systems
simplify our daily life style. Moreover, it provides a characteristics (i.e. distribution, heterogeneity,
intelligence …etc.) and models its behaviors in well-
Copyright c by the paper’s authors. Copying permitted for structured manner with respect to its functional
private and academic purposes.
In: Proceedings of the 3rd Edition of the International
Conference on Advanced Aspects of Software Engineering Page 62
(ICAASE18), Constantine, Algeria, 1,2-December-2018, pub-
lished at http://ceur-ws.org
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
requirements and global coherency [Kou18]. IoTAA Then, implementation issues are illustrated through
architecture is structured in four layers: three an example (healthcare monitoring system).
horizontal layers and a transversal layer. The first one This paper is structured as follows: Firstly, section 2
is a smart layer that ensures connection and provides the necessary background allied to our
communication between things and the system. The subject, mainly the IoT system, diagnosis and IoTAA
second one constitutes the intelligent core of the Architecture. Next, section 3 introduces the proposed
system which ensures local coordination and provides approach. Section 4 discusses some implementation
internal functioning. The third layer ensures issues and provides a case study dealing with
coordination between the local system and the diagnosis of a IoT based healthcare monitoring
externals ones. The transversal layer provides domain system. Finally, Section 5 concludes the paper and
dependent behavior and users interface. The gives prospects to be continued in the future.
intelligent aspect is ensured by means of MAS
intelligence feature. Each layer is scheduled and 2. Background
maintained by a set of particular agents which interact
together to achieve their goals. 2.1. Internet of Things
On the other hand, fault diagnosis is an important This section reviews some IoT properties, applications
issue for developing IoT systems especially that are and challenges. IoT system is “a system of
made up of heterogonous and distributed physical and interrelated computing devices, mechanical and digital
virtual components. machines, objects, animals or people that are provided
Since these components are strongly coupled, a failure with unique identifiers and the ability to transfer data
that occurs in a given component can propagate to over a network without requiring human-to-human or
others with respect to the existing relationships. This human-to-computer interaction” [Mar14]. The main
can have disastrous consequences on the IoT system feature of IoT systems is the pervasive presence of a
functioning. Accordingly, developing fault diagnosis variety of connected things or objects, such as:
is an important issue for improving system sensors, actuators and Radio-Frequency IDentification
maintenance activities [Kos03]. In other words, their (RFID) tags, etc. These objects are able to interact
reliable execution is an important design concern. with each other and cooperate with their neighbors to
Our paper focuses on the development of diagnosis reach common goals. IoT systems present numerous
IoT systems, where we intend to provide a generic and properties and functionalities, such as:
efficient solution. openness, distribution, dynamicity and
Fault diagnosis research area has received wide interoperability;
attention over the last decades [Kos03] [Muc05]. A possibility to interconnect small devices and
variety of fault diagnosis approaches have been computers to the Internet in cheaply and easily
developed including: model based approaches, manners by means of uniquely identifiable IP
knowledge based approaches, qualitative simulation (Internet Protocol);
based approaches, neural-network based approaches capabilities of sensing and actuations, embedded
and classical multi-variate statistical approaches intelligence and communication;
[Muc05]. self-configurability and ubiquity.
Considering the distributed aspect of the IoT system Examples of IoT systems can cover practically
and the nature of handled information related to the numerous real world applications, they goes beyond
usual observations and the expected system the following field: Smart home, wearables objects,
descriptions, the proposed methods [Kat05] suffer of connected cars, Smart cities, Smart retail, Smart
an overall view state of the system which influences farming, etc.
greatly diagnosis quality. From diagnosis view point, Although IoT is promising research area, it is lacking
it is quite difficult to obtain concise and correct results of experimentation for modeling and carrying out IoT
from a centralized diagnosis process. On the other systems. Hence, some challenging issues should be
hand, a completely distributed diagnosis solution pinpointed such as: privacy and security, power
converges generally to incoherent and imprecise consumption, analyzing and managing real-time data,
results. Thus, combining both alternatives gives rise to resource constraints and QoS supports, Data storage
an appropriate and efficient answer, that is, diagnosis (Big Data issues), standards and connectivity
process should be semi centralized. (Protocols, Norms, and Platforms).
Our contributions twofold:
First, the IoTAA- based diagnosis approach is 2.2. Diagnosis
proposed. The choice of this IoTAA fits IoT Fault diagnosis is usually integrated into the largest
diagnosis problem requirements. The four layers framework of monitoring, supervision and
are refined and fault diagnosis is introduced. maintenance. Generally, fault diagnosis can be defined
International Conference on Advanced Aspects of Software Engineering Page 63
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
as a process of three complementary tasks: fault achieve the further suitable functioning for horizontal
detection, fault isolation and fault identification layers with respect to the application domain, system
[Muc05]. These tasks are defined hereafter. requirements (i.e. desired behavior) and layer nature.
Fault detection: it consists in deciding whether the Furthermore, this layer interfaces the application with
system works in normal conditions or whether a the potential users and participants. For more details,
fault has occurred. It is a logical operation whose reader can refer to [Kou18].
answer must be true or false.
Fault isolation: it is triggered whenever fault has 3. Proposed Approach
occurred. It aims at localizing the potential
components causing the fault. 3.1. IoTAA Diagnosis Architecture
Our IoTAA based diagnosis architecture is structured
Fault identification: it intends to identify the nature
as follows:
of the fault. In other words, it analyses specific
Physical Component Management Diagnosis
faults parameters among others its size, criticality
Layer (PCM-D): is the lower horizontal layer. It
and significance.
monitors, manages and controls the IoT
The fault detection operates on desired behavior infrastructure. It is directly related to the IoT
model whereas both fault isolation and identification systems sensors, actuators, tags, smart devices
involve a faulty system behavior model under the and other terminals. PCM-D ensures the
considered faults. following functionalities: perception,
Topics concerning fault diagnosis field point toward communication and connection between things,
several aspects and issues, among others: devices and agents. Also it enables achieving
The representation model of the behavior and management and control of IoT resources by
diagnosis process. providing local fault diagnosis, taking the
The reasoning model (i.e. algorithm) and the adequate decision and acting accordingly in real-
associated decision making process which should time. Agents of this layer are named Agents of
comply with the representation model. Things (AoT). PCM-D layer interacts directly
The global diagnosis strategy and the with LMC-D layer.
corresponding system architecture.
The consideration of some particular aspects and
characteristics of the system being to be
diagnosed.
The ontological structure which describes the
Global Management and Coordination Layer S
different faults concepts and their relationship. O
System diagnosability.
M
Ontology
Local Management and Coordination Layer
L
An important remark should be highlighted, which A
Y
concerns fault diagnosis: in this paper, we are Physical Component Management Layer
E
R
concerned by the third point (i.e. global diagnosis
strategy). We assume that the system is diagnosable. Interlayers and external Communication GCA
An ongoing paper addresses the representation model Intra-Layer communication
Ontology Share
Control
LCA
GC
AoT
and its allied reasoning algorithm that are Fuzzy Logic Sensor, Actuator, Tag, smart, etc.
device
IA
based. Figure 1: Internet of Things Architecture [Kou18].
2.3. Internet of Things Agent Architecture Local Management and Coordination
IoTAA [10] is multi-layer architecture (See Figure 1); Diagnosis Layer (LMC-D): it represents the
it aims to provide a general framework for developing middle layer. LMC-D ensures management of
IoT systems. It is made up of four layers, structured collective intelligent behavior that requires a
into three horizontal layers; named respectively: particular decision making and enables
Physical Component Management Layer, Local coordination between the local agents in order to
Management and Coordination Layer and Global achieve a global fault diagnosis of the local
Management and Coordination Layer; alongside a system. Such responsibility can be achieved by
transversal layer, called Specialized Operative using sophisticated protocols and exploiting
Management Layer. information delivered from the PCM-D layer.
The horizontal layers ensure management of physical Agents of this layer are named Local
components and the coordination between the Coordination Agent (LCA). LMC-D layer
different parts of the system. However, the transversal interacts directly with both PCM-D and GMC-D
one takes in charge the responsibility to assign and layers. PCM-D and LMC-D layers constitute
International Conference on Advanced Aspects of Software Engineering Page 64
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
jointly the local system that needs to interact and process involved at different layers. Along this
coordinate with other systems, or acquires needed layer, Interface Agents (IA), are set in and
services from application servers (e.g. Cloud interact with users.
servers). It can also need to be coordinated by a Relationships between these kinds of agent together
remote part. and the connected components are depicted by Figure
Global Management and Coordination 2.
Diagnosis Layer (GMC-D): It is the higher 3.2. Agent description and Interaction
horizontal layer which can be seen as a social one IoTAA Diagnosis agents are described as follows:
due to its relationship with other systems. It 3.2.1. Agent of Things (AoT)
ensures interaction and communication between AoT is a hybrid agent which intends to monitor and
agents of the current IoT MAS and remote agents diagnoses locally a set of components (i.e. sensors,
or external parts in order to provide a common things, RFID). AoT internal architecture is depicted
global fault diagnosis for several IoT systems. In by Figure 3. It is structured into five components:
other words, it empowers the openness property Perception and Monitoring; Local Diagnosis; Decision
which is quite fundamental for both IoT systems Making and Data Storage; Internal Interaction
and MASs, thus, possible exogenous reasons Management and Action.
could be found. Agents of this layer are similar to Perception and Monitoring: detects
LCA-D, named Global Coordination Agents environment changes and gathers relevant real-
(GCA). They interface the current system with time data issued from different connected
other MASs or related popular platforms (i.e. terminals, such as sensors measurements, new
social network, Cloud …etc.) in order to device identification. AoT communicates
exchange relevant data and services, achieve a periodically the sequence of percepts to Decision
complex task or resolve possible conflicts. Agents Making and Data Storage component.
of this layer, which are generally remote, are also Local Diagnosis: it consists of diagnosing faults
liable for coordinating several local systems (i.e. locally (i.e. connected things).
systems of systems) and can be located in the Decision Making and Data Storage: based on
Cloud whenever the privacy property can be the sequence of perceived Data and diagnosis
preserved or is trivial. results, Decision-Making is a simple mapping
from situations to actions (i.e. stimulus /
responses). Each AoT has its own library which
assists and resumes its behavior. Such decision
doesn’t require any reasoning capability. It
consists of determining the suitable course of
actions and the required interaction to perform the
desired behavior. This sub module decides the
relevant data to be stored (i.e. locally or on the
cloud).
Internal Interaction Management: ensures
internal interaction management between the AoT
agents together and with LCA agents.
Figure 2: Class Diagram of IoTAA-D Diagnosis Action: After deciding about the adequate actions
System. to be carried out, AoT acts upon its environment
Specialized Operative Management Diagnosis through its actuators.
Layer (SOM-D): is deliberative layer. It models AoT’s Knowledge Base includes behaviors for
the additional functionalities that should be taken analyzing, controlling and storing IoT data. It is also
into account by the system. It depends on the enhanced by further knowledge which is related to the
functioning requirements, desired behavior and fault diagnosis (i.e. KBsop).
system’s application domain. SOM-D transverses Things, sensors and actuators should be assigned to
the horizontal layers by assigning and modeling, the different AoTs. Device assignment could be done
at each level (i.e. Layer), the suitable with respect to particular parameters related to them,
functionalities; with respect to horizontal layer such as their nature, geographic location and
nature and its requirements. Such assignment and functioning capabilities. For instances, things of the
modeling are done during MAS design process; same kind could be controlled by the same AoT. Also,
they are specialized operational functions. The things, actuators and sensors located in the same
main SOM-D’s functionalities are fault diagnosis region could also be managed by the same AoT.
International Conference on Advanced Aspects of Software Engineering Page 65
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
Figure 4: Local Coordination Agent architecture.
Figure 3: Agent of Things Architecture.
3.2.3. Global Coordination Agent (GCA)
3.2.2. Local Coordination Agent (LCA)
GCA is a deliberative agent which aims to manage
LCA is a deliberative agent which intends to manage
and control social behaviors of the system. GCA
and control internal behaviors of the system. LCA
internal architecture is depicted by Figure 5. It is
internal architecture is depicted by Figure 4. It is
structured into five components: Perception,
structured into five components: Perception,
Reasoning and Decision Making, Social Interaction
Reasoning and Decision Making, Interlayer
Management, Meta Diagnosis and Action.
Interaction Management, Partial Global Diagnosis and
Perception: it aims at collecting relevant real-
Action.
time data and detecting environment changes
Perception: consists of gathering and extracting
issued from its own sensors. The particularity of
relevant data from the basic layer (i.e. PCM-D
such data is that it concerns both local data and
layer) and its own sensors. Recall that terminals
social relationship with external parts. Perception
concerned by IoT system are directly connected
component communicates periodically the
to AoTs.
sequence of perceived Data to Decision Making
Reasoning and Decision Making: it is the core component.
intelligent component of LCA. The designer
Reasoning and Decision Making: it intends to
should specify and model an adequate reasoning
deliberate the appropriate course of actions that
mode such as logic based reasoning, cases based
can achieve agent’s goal. The designer should
reasoning, theorem prover, fuzzy reasoning, etc.
implement an adequate reasoning mode by taking
Such choice depends on designer viewpoint and
into account system’s external interactions. In the
its adequacy with the application nature. The
same way as LCA, the reasoning process of GCA
reasoning process is followed up by a decision-
is followed up by a decision making process. The
making process which is based on the reasoning
knowledge base KBs is made up of social
process results.
knowledge that can be distinguished into two
Global Diagnosis: it represents the internal categories: knowledge related to fault diagnosis
intelligent desired behavior that should be and those associated to the norm, regulation and
specified by designer as stated above. The governance conformity [Lez17].
specialized operative internal function should
Meta Diagnosis: it represents the social
implement a reliable diagnosis mechanism which
intelligent desired behavior that should be
is based on the local diagnosis results of the
specified by designer to achieve an overall
involved IoT agents.
diagnosis of several systems.
Interlayer Interaction Management: Particular
Social Interaction Management: it aims to
protocols should be specified in this component
ensure interactions with both GCA and LCA
which concern interaction between AoTs, LCAs
agents, as well, with external systems or
and GCAs involved in the same IoT MAS.
platforms. Thus, designer should specify a well-
Standard protocols such as contract Net [Smi80],
defined interaction protocols by reusing existing
Message Queue Telemetry Transport (MQTT)
standards or proposing new ones.
[Tra09] can be reused or new ones could be
Action: GCA acts upon its environment through
proposed.
its own actuators by executing selected sequence
Action: After deciding about the adequate actions
of actions which impact among other the LCA
to be carried out, LCA acts upon its environment
behavior.
through its own actuators.
3.2.4. Interface Agent (IA)
KBint is the internal knowledge base that includes IA ensures system initialization, environment
local knowledge manipulated over LMC-D layer. discovering, device configuration and all interactions
International Conference on Advanced Aspects of Software Engineering Page 66
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
between IoT MAS and the potential users; For which are chemicals appear in the body when the
instance responding to user request, showing the body fat is used for energy instead of glucose.
result, making suggestions and giving advices. It Acetone which is part of ketone bodies; is an acidic
supports and provides an active assistance to users. In substance, naturally occurring in very small amounts
other words, it does not represent a simple interface in blood and urine. These substances are normally
between the application and users, but it contributes made by the body from fat and eliminated in the urine.
with them to a collaborative process for performing However, they sometimes accumulate in the body,
desired behavior. which can no longer eliminate them completely.
Concerning Interface Agent structure, we admit Acidification of the blood occurs and is referred to as
that it could be implemented differently with respect ketoacidosis. This acidification is toxic and can lead to
to users’ preferences and designer’s requirements. disorders that can go as far as coma [Vee04]. It’s also
Thus, we make a full abstraction about its internal a serious complication of type 1 diabetic mellitus
architecture. This agent interacts with other agents in called a diabetic ketoacidosis (DKA). Acetone is
order to transmit requests to the concerned part in the known as a biomarker of diabetes [Wan10], and
system or to acquire the needed information that assist breath acetone concentration is reported to be elevated
users. in type 1 diabetes mellitus, and it can be used to
diagnose the onset of diabetes [Den04]. Therefore,
measuring of Acetone level can help controlling and
monitoring the diabetic patients’ condition as the large
number of ketones means diabetes is out of control.
b) Description of the developed prototype device
“Diab-check”
The prototype device called “Diab-check” consists of
a hardware and software as an Internet of Things (IoT)
system for breath test to monitor the condition of
diabetic patients. The monitoring device for ketone
Figure 5: Global Coordination Agent architecture. level by using breath measurement (See Figure 6) is
constructed and the required software is developed by
Remark: All diagnosis modules are structured into the team in computer science laboratories and medical
three complementary modules, which represent faults research laboratory LR2M, university of Constantine,
detection, faults isolation and fault identification. In
Algeria. It is presented by the logo “ ". The patient
addition, the proposed architecture is generic in the
breath into the prototype, thus the level of the acetone
sense that these modules could be implemented in
is measured, the value is adjusted by using ambient
different manners with respect to application domain
humidity and temperature degrees, the result are
and designer’s choice. The same statement concerns
displayed in an LCD, sent to a mobile application and
the interactions exchanged between agents and the
stored in the control system data base. Thus, the whole
corresponding protocols.
system controls the Acetone level in the breath. Such
4. Case Study: Diagnosis of a IoT Based system can be helpful for different users such as
Healthcare Monitoring Sensors For Diabetic patients, patient assistant and health
Diabetic Patients practitioners for patients monitoring. Accordingly, it
IoTAA-D can be implemented in several manners can be used or enriched by additional features for
with respect to: application domain, diagnosis health automation (Smart healthcare).
algorithms and models, MAS platforms and tools, IoT
platforms, protocols and technologies [Kou18]. As an
example of IoTAA-D application, let us consider an LCD
IoTAA-D diagnosis based application which
diagnoses IoT system sensors of a Health Care
Monitoring for Diabetic Patients. In this section, we
outline the main design stepwise. The main goal of Figure 6: “ Diab-check Device.
this application is to achieve an easy handheld c) Physical realization of “Diab-check” prototype
healthcare on monitoring diabetic via breath. The stepwise to design the “Diab-check” is described
a) Description of medic standing and the as follows:
underlying scientific phenomenon Firs, the following sensors are required: Gas
It’s well known that when the body has too little detection sensor, Temperature sensor and
insulin, it means that the cells of the body cannot take Humidity sensor.
enough sugar (glucose) from the blood. So, ketones
International Conference on Advanced Aspects of Software Engineering Page 67
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
Effectively for detecting breath acetone we use
FIGARO TGS822 gas sensor. The acetone level is in
unit of mmol/l, so the value of the sensor which reads
in the unit of PPM must be converted to mmol/l by
taking the molar weight of acetone which is 58.08.
The stability of the gas sensor is depending on the
value of ambient humidity and temperature. Based Figure 9: Mobile Application.
on the data sheet, gas sensor is stable at around 60%
humidity and 28-29ºC. Thus, the two parameters
should be included, therefore we use DHT11 sensor,
so that the concentration of the acetone gas can be
determined. Concerning microcontroller and
communication technology hardware, we have used,
respectively, Arduino Uno and GSM module. In
addition, other materials are used such as: USB cables,
LEDs, Breadboard, Connection wires, Resistances,
etc...These materials are depicted by Figure 7.
Figure 10. Scenario of health facility monitoring
based on IoTAA-D.
In this application, agent, are implemented by means
Figure 7: “ Diab-check” main Hardware of JADE platform. Such platform choice is motivated
Components. by the fact that is the most popular FIPA-compliant
Connection of these components is made up on agent platform in the academic and industrial
Breadboard and is illustrated by Figure 8. The community. Moreover, it is free, stable software and
program that controls sensors and ensures data an open source framework which is distributed by
transmission between sensors, mobile and server Telecom Italia.
applications is implemented by means of Arduino At this stage, we describe some details that concern
Software [Ard18] which is fully rewritten in C some agents’ behavior:
language. First, the code should be uploaded into the Agent of Things (AoT): the assignment of
Microcontroller. After that, the sensors communicate sensors to agents of things is based on Diab-
the corresponding data to the mobile application via Check devices deployed in each room; thus, we
GSM sensor. In turn, these data are transmitted to the distinguish an AoT associated to each device.
server application and displayed on LCD. Mobile AoT has a repository of stimulus response that
application on Smartphone is implemented under describes the monitoring routine. Diagnosis
Android platform (Figure 9). modules are based on inference engine of Prolog
language, where several inference rules are
defined. Some necessary interactions can take
place when there is interference between the
measured data that are controlled in the same
room.
Figure 8. Connected Things of the “Diab-check” Local Coordination Agent: one agent can be
d) Scenario: Application of Diab-check in Health distinguished by room. Each local coordination
facility agent has to coordinate, locally, agents of things
We consider two treatment rooms R1 and R2 which together. Each Local coordination agent has a
can receive respectively n (n>0) and m (m>0) diabetic local diagnosis module that ensures diagnosis of
patients. Each patient can be monitored by a “Dib- the different devices involved by room. Such
check device”. Each room is supervised by a health module detects anomalies and malfunctioning
manager. The overall health facility is supervised by a related to the devices. In addition, based on an
health supervisor. This sample could be modeled by inference rules, it localizes and fixes possible
means of IoTAA-D as depicted in Figure 10. failures.
International Conference on Advanced Aspects of Software Engineering Page 68
ICAASE, December, 01-02, 2018
Internet of Things Agents Diagnosis Architecture: Application to Healthcare IoT System ICAASE'2018
Global Coordination Agent: only one global of foundation models for Internet of
Things. Frontiers of Computer Science in
coordination agent has to be used to ensure local China, 4(3), 376-385, 2010.
systems coherency and potential interactions. [Den04] Deng C., Zhang J., Yu X., Zhang W., and Zhang
Similarly to local coordination agent, this agent X., Determination of acetone in human breath by
has a diagnosis module which exploits the global gas chromatography-mass spectrometry and solid-
phase microextraction with on-fiber derivatization.
diagnosis results of R1 and R2, to make the Meta J.Chromatogr. B Anal. Technol. Biomed. Life Sci.,
diagnosis decision. vol. 810, no. 2, pp.269–275, 2004.
[For12] Fortino, G., Guerrieri, A., and Russo, W. Agent-
oriented smart objects development. In Computer
Supported Cooperative Work in Design (CSCWD),
IEEE 16th International Conference on (pp. 907-
912), May 2012.
[Gil16] Gilchrist, A, Industry 4.0: the industrial internet of
things. Apress2016.
[Int15] Internet of Things (IoT) Ecosystem Study. 2015.
[Kat05] Katipamula, Srinivas et Brambley, Michael R.
Methods for fault detection, diagnostics, and
Figure 11: Launching agents on JADE platform. prognostics for building systems—a review, part II.
For instance, Figure 11 shows the deployment of Hvac&R Research, 2005, vol. 11, no 2, p. 169-187.
agents, by room, on JADE platform. [Kat08] Katasonov, A., Kaykova, O., Khriyenko, O.,
Nikitin, S., adn Terziyan, V. Y. Smart Semantic
5. Conclusion Middleware for the Internet of Things. ICINCO-
IoT is a promising research area. In this paper we have ICSO, 8, 169-178, 2008
[Kaz09] Kazanavicius, E., Kazanavicius, V., and
proposed an agent based architecture for the diagnosis Ostaseviciute, L. Agent-based framework for
of IoT systems. The proposed architecture takes into embedded systems development in smart
accounts the diagnosis IoT systems requirements. The environments. In Proceedings of International
motivation behind modeling such systems by means Conference on Information Technologies (IT 2009),
Las Vegas, NV, USA (Vol. 2729), April 2009.
of MAS paradigm consists mainly of intelligence, [Kos03] Kościelny, J. M. and Syfert, M. Fuzzy logic
distribution and interactions features. The architecture application to diagnostics of industrial
controls the connected devices and materials, provides processes. IFAC Proceedings Volumes, vol. 36, no
local diagnosis, global diagnosis and Meta diagnosis. 5, p. 711-716, 2003.
[Kou18] Kouah S. and Kitouni I. Multi-Layer Agent Based
Also, it ensures communication and interactions Architecture for Internet of Things Systems. Journal
between different things, agents and external systems. of Information Technology Research (JITR), 11(4),
Likewise, it models further functionalities which are 32-52, 2018.
[Lez17] Lezzar, F., Benmerzoug, D., Kitouni, I. & Osmani,
application domain dependent. The feasibility of A. A GRC-Centric Approach for Enhancing
IoTAA-D implementation is exemplified via a Management Process of IoT-Based Health
healthcare case study. Institution. In International Conference on Mobile,
This work may be continued in several ways: Firstly, Secure, and Programmable Networking(pp. 207-
221). Springer, Cham. June 2017.
we intend to use IoTAA-D architecture to propose a [Mar14] Market, S. P., Self-Powered, E. P., & Module, P.
concise design methodology that enable and assist B. T. (2014). Internet of Things.
Diagnosis IoT systems development. On the other [Muc05] Mucino M. and Li Y. G., A diagnostic system for
hand, we expect to investigate the diagnosis models gas turbines using GPA-Index, Comademc007-
(2005).
and its allied reasoning algorithm in order to more
[Smi80] Smith, R. G. The contract net protocol: High-level
tolerate the system. Moreover, the architecture can communication and control in a distributed
make subject of many case study implementations problem solver. IEEE Transactions on computers,
related to various application domains, for instance (12), 1104-1113. 1980
smart street lighting and smart planning. [Tra09] Transport, M. T. (2009). MQTT. org.
[Vee04] Veech R. L., The therapeutic implications of
6. REFERENCES ketone bodies: The effects of ketone bodies in
[Ard18] Arduino : https://www.arduino.cc/. pathological conditions: Ketosis, ketogenic diet,
[Aya12] Ayala, I., Amor, M., and Fuentes, L. An agent redox states, insulin resistance, and mitochondrial
platform for self-configuring agents in the internet metabolism. Prostaglandins Leukot. Essent. Fat.
of things. In Third International Workshop on Acids, vol. 70, no. 3, pp. 309–319,2004.
Infrastructures and Tools for Multiagent Systems, [Wan10] Wang C., Mbi a. and Shepherd M., A study on
ITMAS (pp. 65-78), June 2012. breath acetone in diabetic patients using a cavity
[Cha18] Chatterjee S. Byun J. Dutta, Pedersen R. U. ringdown breath analyzer: Exploring correlations
Pottathil A. and Xie H. Designing an Internet-of- of breath acetone with blood glucose and
Things (IoT) and sensor-based in-home glycohemoglobin A1C. IEEE Sens. J., vol. 10, no.
monitoring system for assisting diabetes patients: 1, pp. 54–63, 2010.
iterative learning from two case studies. [Weg97] Wegner, P. Why interaction is more powerful
European Journal of Information Systems, 2018. than algorithms. Communications of the
[Che10] Chen, L., Tseng, M., and Lian, X. Development ACM, 40(5), 80-91, 1997.
International Conference on Advanced Aspects of Software Engineering Page 69
ICAASE, December, 01-02, 2018