Content uploaded by Suresh Chavhan
Author content
All content in this area was uploaded by Suresh Chavhan on Sep 16, 2020
Content may be subject to copyright.
IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020 6023
IoT-Based Context-Aware Intelligent Public
Transport System in a Metropolitan Area
Suresh Chavhan ,Member, IEEE, Deepak Gupta , B. N. Chandana ,Member, IEEE,
Ashish Khanna, and Joel J. P. C. Rodrigues ,Fellow, IEEE
Abstract—The public transportation system (PTS) in a
metropolitan area is a nonlinear, dynamic, and complex system.
Managing and providing suitable public transportation services
are difficult. In this article, we propose an Internet of Things-
based intelligent PTS (IoT-IPTS) in a metropolitan area. An
IoT is used to interconnect transportation entities, such as vehi-
cles, commuters (mobile phones), routes (sensors), roadside units
(RSUs), etc., in a metropolitan area. The IoT provides the
seamless connectivity between different networking technologies
whenever the commuters or vehicles move from one location to
another location. Hence, IoT provides the suitable seamless pub-
lic transportation services in the metropolitan area. In addition,
we have used context information of transportation entities, such
as routes condition, traffic density, number of routes available,
traffic congestion, vehicles’ movement, and their mobility, which
are stored in the cloud. The stored context information in cloud
along with the IoTs are used to find the relevant routes, alterna-
tive modes, departure times, and many more for providing public
transportation services in a metropolitan area. The proposed IoT-
IPTS makes use of static and mobile agents with the emergent
intelligence technique (EIT) for collecting, analyzing, and shar-
ing context information. The analyzed context information is used
to form the policies to provide the best available public trans-
portation services to the commuters in a metropolitan area. The
software-defined network is used to enable the cloud computing
and EI network to manage the public transportation services to
the commuters.
Index Terms—Agent, cloud computing, context information,
intelligent public transport system, Internet of Things (IoT),
metropolitan area, policy, software-defined network (SDN).
Manuscript received August 31, 2019; revised October 11, 2019 and
October 28, 2019; accepted November 6, 2019. Date of publication
November 22, 2019; date of current version July 10, 2020. This work was
supported in part by the Fundação para a CiênciaeaTecnologiaunder
Project UID/EEA/50008/2019, and in part by the Brazilian National Council
for Scientific and Technological Development under Grant 309335/2017-5.
(Corresponding author: Joel J. P. C. Rodrigues.)
S. Chavhan was with the Department of Electrical Communication
Engineering, Indian Institute of Science, Bengaluru 560012, India. He is now
with the Automotive Research Center, Vellore Institute of Technology, Vellore
632014, India (e-mail: suresh.chavhan@vit.ac.in).
D. Gupta and A. Khanna are with the Department of Computer Science
Engineering, Maharaja Agrasen Institute of Technology, New Delhi 110085,
India (e-mail: deepakgupta@mait.ac.in; ashishkhanna@mait.ac.in).
B. N. Chandana is with the Department of Electronics and
Telecommunication Engineering, SIT Tumkur, Tumkur 572103, India
(e-mail: chandanabn65@gmail.com).
J. J. P. C. Rodrigues is with the PPGEE, Federal University of Piauí,
Teresina 64049-550, Brazil, and also with the Instituto de Telecomunicações,
1049-001 Lisbon, Portugal (e-mail: joeljr@ieee.org).
This article has supplementary downloadable material available at
http://ieeexplore.ieee.org, provided by the author.
Digital Object Identifier 10.1109/JIOT.2019.2955102
I. INTRODUCTION
METROPOLITAN areas are more populated, and they
represent source of congestion, pollution, and over-
crowding. This puts more pressure on public transportation,
which results in issues, such as traffic jams, underutiliza-
tion of resources, waiting time, etc. The migration of public
to private transportation leads to an increase in traffic con-
gestion and jams. To avoid or reduce the above-mentioned
issues, there is a need of fast and intelligent public trans-
portation. At the same, time prediction of traffic and their
abnormalities [17], [20]–[22] are also important. In order to
manage demands and assist commuters and transportation
authorities, there is a necessity to design and deploy an intel-
ligent decision making and control system. Hence, there is
a requirement of advanced information and communication
technologies (ICTs), which is fulfilled by introducing an IoT-
based intelligent public transportation system (IoT-IPTS). The
main aim of IoT-IPTS is to enhance comfortability, efficiency,
and effectiveness of the transportation system. In addition, it
improves quality of mobility (seamless connectivity), safety,
traffic management, multimodality, etc. Fig. 1 shows the IPTS
in a metropolitan area.
Despite the growth of ICTs, the commuters and trans-
portation authorities are not able to get accurate information
of transportation entities, which causes many problems. To
avoid/reduce these problems, an IoT is used to interconnect all
the transportation entities and collect their context information.
It analyzes various circumstances and estimates the number of
routes available, traffic densities, vehicles movement and their
mobility, etc. In IPTS, an IoT-based context information uti-
lization could provide rich services, such as relevant routes,
alternative modes, etc. Hence, an IoT-based context-aware
IPTS makes the public transportation system (PTS) more
reliable, provides seamless connectivity to all transportation
entities to make the smart city, and provides accurate and more
relevant seamless services to commuters’ and transportation
authorities hands.
Existing works, such as Google [2], urban transport [3],
and context-based services for IPTSs [1], [4], are used to
provide real-time information to commuters based on the cur-
rent location, destination location, and context information
of mobile devices used by commuters in an urban city. To
provide the realistic scenario information, current and his-
torical context information of entities has to be considered.
This context information is collected, analyzed, and estimated
2327-4662 c
2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6024 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
Fig. 1. Scenario of IPTS for a metropolitan area.
and it predicts the realistic scenario, and then decentralized
decisions are taken over different regions in a metropolitan
area. This decision information is communicated to central
management authority using the emergent intelligence tech-
nique (EIT). The main function of EIT is to build IPTS
by collecting context data from various sources [like com-
muters, roadside units (RSUs), vehicles, routes, etc.]. The
EIT adapts to dynamic changes in transportation entities con-
text and demands of commuters’ requirements using available
and historical transportation data. In this article, the EIT and
context information of entities realize smart metropolitan PTS.
The proposed system makes use of both static and
mobile agents (MAs), and the MA gathers context
information. Depending upon the collected and analyzed con-
text information, the static agent (SA) forms policy to predict
relevant routes, alternative modes, and departure times. The
predicted information is used for managing various trans-
portation services, traffic flow redistribution, traffic and travel
demand control, road capacity control, etc. The proposed
system effectively and efficiently manages various services at
different situations using context information. The proposed
system’s performance improves by adding MAs in the EI
technique, as compared to the traditional system client–server
(CS), because they potentially reduce the communication
network traffic, are highly adaptable, remove centralized bot-
tleneck, provide high scalability, and are more efficiently used
in distributed heterogeneous environment.
The rest of this article is organized as follows. Section II
presents the literature survey. Section III discusses the rele-
vant concepts and definitions required for the proposed work.
Section IV presents the IoT-based context-aware PTS architec-
ture. Section V analyzes the IPTS execution time. Section VI
presents the simulation scenario. Section VII presents the sim-
ulation and performance evaluation of the proposed system.
Section VIII presents the discussion of the proposed work
compared with the existing works, and conclusions are drawn
in Section IX.
II. LITERATURE REVIEW
In this section, we present few related works on the context
information, services, and PTS in the various networks.
Google [2] is the ubiquitous service provider, which learns
user’s past habits, predicts user’s behavior, and provides
real-time relevant information. Transport urban [3] project
developed a basic route planner for public transportation which
takes inputs as current and destination location. In this project,
inputs are manually inserted which does not consider the
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
CHAVHAN et al.: IoT-BASED CONTEXT-AWARE INTELLIGENT PUBLIC TRANSPORT SYSTEM IN METROPOLITAN AREA 6025
TAB LE I I
TERMINOLOGY/TECHNOLOGY,ABBREVIATIONS,AND SYMBOLS
context information of vehicles, routes, weather, staff, and
commuters in a metropolitan area.
Context-based service for intelligent transportation systems
was designed in [1] and [4] for providing services to com-
muters in order to take informed decisions for public trans-
portation. In addition, they have developed mobile applica-
tions, which understand the user’s behavior and react accord-
ingly with it. The context-aware reflective middle-ware system
for mobile applications (CARISMA) [5] is a mobile computing
middleware, and adaptive and context-aware services for appli-
cations. Depending upon the type of applications developed,
the system provides policies for delivering particular services
and configures context.
The context-aware middleware for resource management
in the wireless Internet (CARMEN) [7] determines context
information by taking into account metadata as a basis,
which includes policies, user profiles, terminal capabilities,
and resource characteristics. The context-based approach to
vehicle behavior prediction in [15] analyzes the situations
using required context information and collected vehicle
data used for deriving location context information. The
authors have presented a process to infer high-level vehicle’s
state information using context information. The policy-based
network management [16] used context information to form
policies for taking decisions while enforcing business rules
and service priorities. Mobility as a service (MaaS) [28]–[31]
combines transport services from public and private transport
providers through a unified gateway to create and manage trips
and makes user pay for it with a single account. MaaS offers
travelers mobility solutions based on their travel needs.
Table I shows the comparisons of few existing intelligent
public transport system works against the proposed work.
III. DEFINITIONS,COMMUNICATION TECHNOLOGIES,
ASSUMPTIONS,AND CONCEPTS USED IN
THE PROPOSED SYSTEM
Table II shows the terminology, technology, abbreviations,
and symbols used in this article. Each of these along with
the assumptions considered in this article are defined and
discussed in the supporting supplement material.
IV. PROPOSED IOT-BASED CONTEXT-AWA RE PUBLIC
TRANSPORTATION SYSTEM ARCHITECTURE
The proposed IoT-based context-aware PTS architecture as
shown in Fig. 2(a) is available at each public transport depot,
and its each module is explained in detail as follows.
1) Context Acquisition Module: In the vicinity of pub-
lic transport depot, the context acquisition module
acquires static and dynamic context information of vehi-
cles, staff, commuters, routes, RSUs, etc. This module
uses EIT with MAs to collect all transportation enti-
ties context information and provides to the SA for
refinement.
2) Context Analysis and Sharing Module: In this module,
the collected context information is analyzed probabilis-
tically using historical context information. This module
shares analyzed context information with agents and
among transportation entities by using the EI technique
with MAs in the vicinity of public transport depot.
3) Route Estimation Module: This module uses collected,
analyzed, historical, and refined context information
to estimate routes. It probabilistically estimates set of
routes and relevant routes from source to destination. It
provides relevant routes information to the SA module.
4) SA Module: The SA resides in public transport depot.
Its main objective is to coordinate all modules as shown
in Fig. 2(a). It creates and dispatches MAs in the vicin-
ity of public transport depot. It collects, analyzes, and
refines context information of all transportation entities.
The refined context information used to form policies
to provide relevant public transportation services to
commuters in the metropolitan area.
5) Public Transportation Management Module: This mod-
ule manages transportation on estimated relevant routes
using EIT along with communications among agents,
vehicles, and RSUs. It provides public transportation
services depending upon the policies.
6) Software-Defined Network (SDN) Controller: The con-
text information of transportation entities is collected
from various IoT-based embedded devices and given to
the SDN controller. The SDN controller refers to the
cloud data along with the collected information to pro-
cess and manage the transportation services, such as
transport resource allocation, alternative routes, trans-
portation modes, redistribution of public transporta-
tion, etc.
A. Context Information Collection, Analysis, and Sharing
In the vicinity of RSUs, context information of vehi-
cles, staff, etc., is acquired and provides SA for refinement
and analysis as shown in Fig. 2(b). Fig. 2(b) shows how
EIT forms cluster/group with neighbor transportation enti-
ties and communicates, and during communication these
transportation entities share with each other their context
information.
This collected context information is shared with agents and
also among transportation entities by using an EIT with MAs
in the vicinity of public transport depot. The collected context
information of vehicles, staff, commuters, routes, and envi-
ronment is analyzed probabilistically in the vicinity of public
transport depot in a metropolitan area as follows.
1) Vehicle (V): The state of vehicle is determined by its cur-
rent available context (c) information, historical context
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6026 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
TAB LE I
COMPARISONS OF RELATED WORKS AND PROPOSED WORK
Fig. 2. (a) Architecture of IoT-based context-aware IPTS at a public transport depot. (b) Context information collection and sharing in the vicinity of RSUs.
(h) information, and dynamic relationships with neigh-
bor vehicles. The vehicle context model is given as
follows:
PVi
t|
vVv
t−1,CIv,h
t−1,CIv,c
t(1)
where Vi
tindicates the dynamic state of the ith vehicle
at time t(which contains low-level information, such as
position, orientation, velocity, and acceleration), CI is
the context information (situation the vehicle is engaged
in), and vis total number of context information,
v{Type,Speed,Capacity,Location,Heading,Failure
rates,Availability,...}.
2) Staff (S): The behavior of staff depends upon current
and historical context information which includes skill,
fatigue, experience, and many others. The probabilistic
context model of staff is given as follows:
PSi
t|
sSs
t−1,CIs,h
t−1,CIs,c
t(2)
where Si
tindicates the dynamic state of staff iand sis
the total number of context information, s{Health con-
dition,Skill,Fatigue,Punctuality,Experience,...}
3) Routes (Rt): The context of route depends upon
the present and past context information, and its
probabilistic context model is given as follows:
PRi
t|
rRr
t−1,CIr,h
t−1,CIr,c
t(3)
where Rtis the set of routes (discussed in the subse-
quent sections), Ri
tis the dynamic state of route i, and
ris the set of context information, r{Type,Density,
Movements,Mobility,Accidents,Traffic jam,...}.
4) Environment (E): The environmental conditions will
alter dynamic properties of staff, vehicles, and routes.
The environmental condition depends upon the past and
current context information and its model is given as
follows:
P(E|CIc,CIh)=
3
i=1
P(E|CIci,CIhi)(4)
where i=1israiny;i=2 is dusty/sunny; and i=3is
foggy.
5) Commuters (C): Depending upon the type of commuters
and their conditions, transportation modes are going to
select, and commuters’ conditions depend upon the past
and current context information and is represented as
follows:
PCi
t|
nSn
t−1,CIn,h
t−1,CIn,c
t(5)
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
CHAVHAN et al.: IoT-BASED CONTEXT-AWARE INTELLIGENT PUBLIC TRANSPORT SYSTEM IN METROPOLITAN AREA 6027
Algorithm 1 Context Information Acquisition, Analysis, and
Sharing in the Vicinity of Public Transport Depot
1: Begin
2: Define depot vicinity, number of SAs and MAs.
3: Let Ris total number of transport depot regions and zis total number of bus station
zones.
4: for l=1to R do
5: Static agent (SA) is deployed in lth region, i.e., SAl
6: SAlcreates and dispatches mobile agents (MAs) in its zones
7: for i=1to z do
8: MAiis created and dispatched in zone i
9: for j=1to b do
10: MAi,jinteracts and collects context information of following entities in
jth bus station of ith zone:
11: RSUs, vehicles, staff, commuters, routes and environment.
12: end for
13: MAianalyzes and shares collected context information among neighbor MAs,
vehicles and RSUs
14: MAiprovides analyzed, collected and shared context information to SAl.
15: end for
16: end for
17: End
where Ci
tindicates dynamic state of commuters (like
patient, kid, and normal commuters) and nindicates the
number of context information of commuters, including
location, type, severity, profile, preferences, etc.
For example [referring to Fig. 2(b)], the SA creates MAs
and dispatches in the vicinity of transport depot. The dis-
patched MA migrates to nearby transportation entities (such
as vehicles, RSUs, and mobile phones of people), all these
entities are interconnected with the IoT and execute their pro-
cess and collect context information and again migrate to the
next entity. The collected and analyzed context information
are shared with SA periodically, which is presented in
Algorithm 1.
B. Context-Based Route Estimation
The collected, analyzed, shared, and historical context
information is used to estimate relevant routes to reach the
desired destination. Let rbe the route from starting to the
desire destination point, which consists of various edges along
the route (that is e1,e2,...,el, where lindicates the total num-
ber of edges available along the route r). Let C(r)denote the
cost of route r, which is defined as follows:
C(r)=Cmw(r)+Cs(r)+Cv(r)+Cc(r)+
l
i=1
Ci(6)
where Ciindicates the ith edge cost function which depends
on the edge eiattributes ai,j. The edge cost function is defined
as follows:
Ci=
m
j=1
wjcjai,j(7)
where mis the total number of attributes of edge eiand the
value of cjis the weighted linear combination of cost functions
of mnumber of attributes and is given as
cj:R→ R≥.(8)
Functions Cmw,Cs(r),Cv(r), and Cc(r)are nonadditive
route costs related to the current weather, staff, vehicles, and
commuters conditions.
Let vector w=[wj],j=1,2,...,mindicate the trans-
port entities profile which contains their current and historical
context information. Equation (6) can be rewritten as follows:
C(r)=Cmw(r)+Cs(r)+Cv(r)+Cc(r)+
l
i=1
m
j=1
wjcjai,j
=Cmw(r)+Cs(r)+Cv(r)+Cc(r)+w.uj(r)(9)
where vector u(r)is defined as
uj(r)=
l
i=1
cjai,j.(10)
The probability of preferring route rtthan qtat time twith
given past context information wis given as
P(rt>qt|w)=1
1+eC(qt)−C(rt).(11)
The probability of preferring the set of routes Rtfrom the set
of available routes Rat time tgiven past context information
wis given as
P(Rt|w)=
R
t=1
P(rt>qt|w)=
R
t=1
1
1+eC(qt)−C(rt).(12)
The SA uses this refined context information and set the of
routes Rtto manage PTS in the vicinity of public transport
depot in a metropolitan area.
C. Travel Time of Public Transport Vehicles
on Estimated Routes
Estimation of travel time of public transport vehicles
depends upon variables, such as traffic flow, traffic volume,
capacity, and free flow speed on estimated routes. Estimation
of travel time of public transport vehicles is the function of
combination of Soltman’s and Smock’s functions [43], [44].
The estimation of travel time of public transport vehicles on
route Ris given as
TR=Tf,R+1+α.QR
CRβ+γ.QR(13)
where Tf,Rdenotes the free flow travel time on route R;QR
represents the traffic volume on route R;CRdenotes the route
capacity; and α, β, and γrepresent the parameters.
The complexity of estimating the public transport vehicles’
speed and travel time depends upon the relationships between
the distance between intersections, traffic volume, the number
of bus stops, and average bus stop service time. The public
vehicles’ average travel time is defined using the following
relationships between intersections on route R:
Tavg
R=64.7L0.5
Re0.19NBS
R+0.68 QR
CR+1.21NBS
R.DTavg,BS
R(14)
where Tavg is the average travel time between intersections
(in s), Lis the distance between intersections (km) on route
R,NBS is the number of bus stations, DTavg,BS
Ris the average
dwell time at station (s), QRis the traffic volume (vehicles per
hour), and CRis the estimate route (R) capacity (vehicles per
hour).
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6028 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
Fig. 3. (a) Different communications involved in the vicinity of public transport depots in a metropolitan area. (b) Migration of MAs in the zones in the
vicinity of public transport depot.
D. IoT-Based Intelligent Public Transportation System
Using EIT
In the jurisdiction of public transport depot, there exist
multiple communications, including agents-to-agents (A2A),
agents-to-RSU (A2R), RSU-to-RSU (R2R), RSU-to-vehicle
(R2V), and vehicle-to-vehicle (V2V) as shown in Fig. 3(a).
The EIT makes use of these communications for monitor-
ing, gathering, and sharing context information in the vicinity
of public transport depot (i.e., region) in a metropolitan area.
The IEEE 802.11p wireless access in wireless environments
(WAVEs), 4G, and LTE communication technologies are used
for vehicle-to-anything (V2X) and agent-to-everything (A2X)
during information collection and sharing in the PTS in a
metropolitan area. The SA is deployed at each public transport
depot which creates and dispatches MAs to zones and sub-
zones of the cluster in the vicinity of public transport depot
as shown in Fig. 3(b). Whenever MAs migrate to bus sta-
tions in the vicinity of the transport depot, they use the EI
technique to form groups (which consists of neighbor MAs,
commuters, vehicles, and RSUs) and interact among them
for collecting and sharing context information. The MA pro-
vides collected context information to the SA and stores in the
cloud periodically during interaction. The SA and SDN con-
troller analyzes these context information, provides the type
of services required at that moment, and manages transporta-
tion services in the vicinity of the transport depot. Steps of the
EI technique for providing required services in a metropolitan
area are as follows.
1) It autonomously responds to each event in the vicin-
ity of the public transport depot to reduce conditions
worsening further.
2) It forms a group which consists of neighborhood agents,
vehicles, commuters, and RSUs. Also, it periodically and
spontaneously monitors these entities by making MAs
to gather their current and historical context information
and share among them.
3) It makes SA analyze and refine the collected context
information.
4) It makes SA use this analyzed, refined, and historical
context information to form policies for predicting rel-
evant routes, modes, departure times, traffic densities,
etc.
5) SA provides the required services to the commuters
and transit operators according to the policies in a
metropolitan area.
6) The SDN controller manages the transportation services
by referring to the cloud data and collected context
information of transportation entities.
In each public transport depot’s vicinity, the SA forms
policy using the collected, shared, analyzed, and refined con-
text information of various transportation entities, in order
to provide a particular type of services required at that
movement. We have considered the following situations to
provide services timely in a metropolitan area.
1) During the nonpeak period, there is a sudden occurrence
of commuters’ arrival to the bus stations in the vicinity
of transport depot D1. The SA of D1dispatches MAs
for acquiring the context information of vehicles, routes,
and staff. The SA develops the following policy Pall:
Pall =IF<Environment :=Sunny>∧<Staff :=Not
sufficient>∧<Vehicle :=Not available>∧<Time
:=Non-peak period>∧<Routek:Freeway>∧
<Routek:High mobility pattern>∧<Commuters :
High density>.
Execution of this policy provides status and next bus
timings to commuters, collects resources (such as vehi-
cles and staff) from neighborhood depots, and the SA
of D1allocates the resources.
2) Whenever there is a heavy traffic density at nonpeak
hour on a particular route in the vicinity of public trans-
port depot (say D1), the SA of D1executes the following
policy (Pred):
Pred =IF<Environment :=Sunny>∧<Staff :=
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
CHAVHAN et al.: IoT-BASED CONTEXT-AWARE INTELLIGENT PUBLIC TRANSPORT SYSTEM IN METROPOLITAN AREA 6029
Experienced>∧<Vehicle :=Bus>∧<Time :=
Non-peak period>∧<Routek:Heavey congestion>
∧<Routek:High density>∧<Routek:Freeway>
∧<Routek:Low mobility pattern>∧<Commuters :
Normal passenger>∧<Commuters :Heading to D>.
The executed policy provides conditions on route kand
relevant routes available information to the commuters.
The SA uses relevant routes and context information for
redistributing the public transportation, i.e., bus.
3) At the bus station of D1, during heavy rain, the com-
muters’ density is increased and public transport buses
are scarce. The SA of transport depot 1 (D1) causes Mas
to migrate in the vicinity; using the EI technique, they
acquire context information of routes, vehicles, com-
muters’, etc. The acquired context information is used
to form the following policy (Palt):
Palt =IF<Environment :=Rainy>∧<Staff :=
Fatigue>∧<Vehicle :=Bus>∧<Time :=Non-peak
period>∧<Routek:Heavey congestion>∧<Routek:
High density>∧<Routek:Freeway>∧<Routek:
Low mobility pattern>∧<Commuters :Normal
passenger>∧<Commuters :Heading to D>.
The executed policy provides suitable transportation
modes and alternative public transportation modes, such
as trams, metro, etc., in the metropolitan area.
4) During occurrence of a disaster in the vicinity of public
transport depot D1, like accident and sudden breakdown
of vehicles on route k, this results in the growing number
of vehicles on that route. The EI technique with MAs
acquires context information of disaster, routes, victims,
etc. This acquired context information is given to the
SA. The SA develops the following policy (Pmanag):
Pmanag =IF<Enviroment :=Dusty ∧Sunny>∧
<Staff :=Punctual>∧<Routek:Accident ∧Vehicle
crashed>∧<Routek:=Slow mobility vehicle>∧
<Time :Peak period>∧<Routek:=High density>
∧<Routek:=Freeway>∧<Commuters :Critical
condition>∧<Vehicle :Bus>∧<Commuters :
Heading to D>.
The executed policy provides alternative route (Rt)
information for the vehicles which are stuck and head-
ing to D1through route kand for victims it provides an
alternative public transport vehicle.
The above-mentioned services are obtained by using the
context-based policies, and these services are managed by the
SDN controller in a metropolitan area.
V. ANALYSIS OF IPTS EXECUTION TIME
In this section, we analyze the performance of the CS-based
computing and EIT with MAs using the execution time of
IPTS in a metropolitan area.
Execution time is the amount of time used to complete pro-
cessing a public transportation service. In the CS-based model,
execution time is the time spent for processing the send out
data and results generated at the server. Whereas in the EI tech-
nique with the MA-based model, execution time is the time
spent for forming groups, generating and migrating MAs, pro-
cessing the data, and resulting in a generation at the SA. There
are three components required for estimating the execution
time: 1) transferring time; 2) overhead time; and 3) process-
ing time. In addition, the execution time is also affected by
few more components, including the number of MAs (m); size
of SA (Sp); the number of nodes each MA migrates (n); the
overhead of context-data file (reading and writing context data
into the file) (Ocf ); overhead of MAs (Om); data processing
rate (rd); network transfer rate (rn); size of the collected con-
text information file (the raw context data size at each MA
collects) (Scf ); MA size (Sma); size of policy (Sp); the over-
head time of policy formation (Opf ); and the overhead time of
reconfiguration of policy (Oprc).
For CS-based computing, we have time required for trans-
ferring data (Ttra =[mnScf /rn]); time spent for reading and
writing the context data in file, which is called overhead time
(2mnOcf ), and the overhead time of policy formation and
reconfiguration of policy ([2mp(Opf +Oprc)Scf ]/rnd.
Therefore, using the CS-based model, the total execution
time is
TCS =mnScf
rn+2mnOcf +2mSpOpf +OprcScf
rnd .(15)
For the EI technique with MA-based computing, we have
time spent for migration of MAs from SA to neighbor nodes
[like RSUs, vehicles, handheld devices (like smartphone), and
so on] (mSma/rn) and from neighbor nodes to SA (nSa/rn),
the overhead time used for SA to dispatch and receive mMAs
(2mOma), where all neighbor nodes receive and send each MA
(2nOma); and time spent for executing code and processing
task locally at neighbor nodes ([(m+n)Sma]/rd).
Therefore, the EI technique with MAs-based model execu-
tion time is given as
TEMA =(m+n)Sma
rn+2(m+n)Oma +(m+n)Sma
rd
+2mSpOpf +OprcScf
rnd .(16)
VI. SIMULATION SCENARIO
A realistic metropolitan area scenario considered for the
simulation is shown Fig. 4(a). The realistic simulation scenario
(creation, traffic flow density generation, context information
analysis, prediction, policy formation, services, and also
wireless communication aspects) is scripted and coded in
MATLAB. In addition, Mobile-C agent platform [9] is used
for context information collection, analysis, and sharing. Both
MATLAB and Mobile-C are interfaced to analyze and provide
relevant context-based services as shown in Fig. 4(b). In the
proposed system performance evaluation, we used different
mode real-time traffic information, which has been extracted
from the Caltrans performance measurement system (PeMS)
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6030 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
(a) (b)
Fig. 4. (a) Simulation scenario and (b) MATLAB and mobile-C interface.
TABLE III
SIMULATION PARAMETERS
database [8] on weekdays. The Mobile-C agent platform is
installed in every public transport depots, which deploys SA.
The SA creates and dispatches MAs to zones and regions.
The EIT with MAs forms a synchronized group (consisting of
neighbor RSUs, MAs, and vehicles) and communicates among
them. During communication, they collect and share context
information. In addition, SA analyzes, refines, and predicts
relevant routes, modes, departure time, alternative routes, etc.
In the simulation, we have set 2 GHz as beacon shipping
frequency to avoid the overloading of the network with wide
density of vehicles, and at every 1 s the program updates the
frequency of propagation efficiency. The results of the EI tech-
nique obtained are the average value of 20 repetitions with the
95% confidence interval of each point in the simulation.
The performance evaluation of the proposed system was car-
ried out using the simulation parameters as shown in Table III
on a dual-CPU Intel Core i5-2400 at 3.10-GHz desktop
computer with 12-GB RAM running Fedora version 25.
VII. SIMULATION AND ANALYSIS RESULTS
In this section, we discuss the proposed system performance
using the execution time and also traffic flow redistribution,
choosing public transportation alternative modes, and change
of departure time in a metropolitan area.
Fig. 5(a) shows the traffic flow redistribution to differ-
ent routes available at a location in a metropolitan area.
The EI technique with MAs gathers context information of
vehicles, routes, staffs, commuters, environmental, and RSUs.
The context information of these entities, such as density,
movements pattern, mobility patterns, travel time, congestion,
etc., is used to redistribute the traffic flow among all the
available routes in a metropolitan area. In the current system,
traffic flows are distributed in different routes as shown in
Fig. 5(a), which shows routes 3 and 1 distributed in the high-
est traffic flows and lowest traffic flows in routes 2 and 4.
Hence, the current system faces traffic congestion in routes 3
and 1. The proposed EIT context-based system redistributes
traffic flows among all the four routes depending upon their
context information as shown in Fig. 5(a).
The proposed system encourages commuters to avoid using
their own vehicles and use public transportation in a metropoli-
tan area during high traffic, and travel demand at peak is shown
in Fig. 5(b).
The proposed system motivates commuters to avoid peak
hour commuting with their own vehicles and encourages to use
public transportation at peak hour. In addition, it also motivates
commuters to avoid departure time at peak hour to save time
and fuel in a metropolitan area, as shown in Fig. 5(c).
The developed analysis model of IPTS services the esti-
mate execution time. In the simulation, we have considered
the random waypoint model as the mobility model which is
used for nodes to choose random destinations and move at a
speed of 10 m/s. Table IV consists of parameters to determine
the execution time of CS and MA-based model.
We assume that each MA migrates to an equal number
of nodes. Fig. 6(a) shows the constant profile (execution
time =890 s) irrespective of a varying number of MAs in
the CS-based model. In the case of the MA-based model, exe-
cution time is always less than the CS model because of the
large number of nodes (n). The execution time of the MA-
based model decreases as there is an increase in MA density,
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
CHAVHAN et al.: IoT-BASED CONTEXT-AWARE INTELLIGENT PUBLIC TRANSPORT SYSTEM IN METROPOLITAN AREA 6031
(a) (b) (c)
Fig. 5. Simulation results: (a) traffic flow redistribution, (b) usage of public transportation, and (c) commuters commuting using public transportation and
own vehicles at peak hour and nonpeak hours in a metropolitan area.
(a) (b) (c) (d)
Fig. 6. Analysis results: Execution time with varying MA size and per agent comparison and multiagent systems (MAS) cumulative times. (a) Execution
time with a varying number of mobile agents. (b) Execution time with varying mobile agent size. (c) Per agent comparison. (d) MAS cumulative times.
TAB LE I V
EXECUTION TIME PARAMETERS
and it achieves the lowest execution time when there are five
MAs. From this point, the execution time starts increasing
gradually. This is due to more MAs, which will reduce the
number of nodes that each MA is migrating, hence it will
reduce the time of execution. However, the increase of the
number of MAs causes more connections overhead.
Fig. 6(b) shows the effect of MA size on the execution time
of the CS and MA-based models. Here, we change the MAs
size and keep other parameters as is and consider maximum
MA size to be 10 KB. When Sma <8 KB, the MA-based
model spends less execution time as compared to the CS-based
model. Similarly, when Sma >8 KB, the CS-based model
spends less execution time as compared to the MA-based
model.
Fig. 6(c) and (d) shows the precise difference between con-
text information with and without EIT while taking decisions,
both on a per-agent and a cumulative time basis.
Fig. 7(a) shows the relationship between travel time and
each vehicle’s time headway. Vehicles’ arrival rates are
Poisson distribution at the origin in the real-time traffic system.
The simulation result shows that public transportation vehi-
cles’ travel time drops at average time headway to 3 s. The
average travel time becomes constant after 3 s of average time
headway of vehicles.
At each bus station, the regression method is used as a
mathematical function to adjust the empirical data with the
correlation coefficient. In the simulation, we have assumed the
constant value for the ratio of QRand CRasshowninFig.7(b),
and assumed a constant number of bus stations in Fig. 7(c).
The simulation results obtained are shown in Fig. 7(b) and (c),
for the public transport vehicles average travel time which is
estimated depending upon the parameters, such as distance,
the number of bus stations, average dwell time at station, dis-
tance between intersections, average traffic volume, and travel
time between intersections at each hour. The estimated average
travel time of public transport vehicles is heavily dependent
upon the collected and analyzed context information of roads,
vehicles, trip duration, length, etc. However, the simulation
result confirms that the estimation of public transport vehicles’
travel time depends upon important factors, such as traffic vol-
ume and route capacity, similarly, we can estimate the travel
time of private transport vehicles.
VIII. DISCUSSION
Overall, in this article, the context information of vehicles,
staff, routes, etc., is acquired from the IoT network with nec-
essary sensors by the MAs and provided to the SA. The SA
analyzed acquired context information using historical data,
and shares the analyzed context information. The acquired,
analyzed, and shared context information is used to form the
policies to provide the relevant services of intelligent public
transport system in a metropolitan area.
The existing CS model without EIT provides various
services and shows quite better results. But these existing mod-
els/methods have not used historical database, contextual data,
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6032 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
(a) (b) (c)
Fig. 7. Simulation results: Travel time and vehicles time headway from the origin when the vehicles’ arrival rate is assumed as Poisson distribution.
(a) Relationship between the travel time of the system and each vehicle’s time headway from the origin. (b) Estimating average travel time (Tavg
R) from LR,
NBS
R,DTavg,BS
R,andQR/CR=0.5. (c) Estimating average travel time (Tavg
R) from LR,NBS
R,DTavg,BS
R,QRand CR
TAB LE V
PERFORMANCE COMPARISON
TAB LE V I
EXECUTION DELAY COMPARISON
and the EI technique with agents. The proposed IoT-based
context-aware IPTS is the best compared to the existing mod-
els, as it uses the EI technique with multiple agents (both MA
and static-based model), compared with the traditional CS-
based model. The performance comparison of the proposed
system with the existing model is shown in Table V.
The amount of time required to complete the processing
of public transportation services is the execution time. The
obtained results for the EIT-based MA system show the least
execution time, i.e., 250 s as compared to 890 s (CS model).
This is due to the EI technique with MAs, where they pro-
vide the parallel execution of tasks in a synchronized agents
group. In addition, we compared the proposed system’s exe-
cution time delay (average delay = 33.45 s and average delay
simulation density error = 32.10 s) involved during processing,
analyzing, and sharing task information to the existing meth-
ods, such as epidemic (average delay = 52.97 s and average
delay simulation density error = 51.78 s), E-GPSR (average
delay = 106.50 s and average delay simulation density error
= 106.67 s) as shown in Table VI.
In the simulation analysis, we have used the Poisson arrival
rate λ(1±r∗10%), where ris the random number which takes
value between 0 and 1. The simulation results of the proposed
system show the efficient utilization of public transportation
especially during the heavy traffic and travel demand at peak
hours, and motivate the commuters’ to avoid peak hour depar-
ture for saving time and fuel. In addition, the results confirm
that estimation of public transport vehicles travel time depends
upon the factors, such as traffic volume, travel time, vehicle
condition, and route capacity.
The proposed system provides timely, necessary, and
efficient transportation services by considering the context
information, such as weather conditions, route conditions,
vehicles conditions, demands, and resources availability in a
metropolitan area. According to our knowledge, this is the first
paper to use an emergent intelligence along with context aware
IPTS in a metropolitan area.
IX. CONCLUSION
In this article, we have presented IoT-based IPTS in a
metropolitan area using context information and the EI tech-
nique. The proposed system reduced: 1) the execution time
(i.e., 250 s) of IPTS services compared to the traditional
system (i.e., 890 s); 2) reduced the average delay simulation
density error (32.10 s) as compared to the epidemic (51.78 s)
and E-GPSR (106.67 s); 3) improved the intermodal time
(150 s) compared to the without EIT and context information
(300 s); 4) increased the usage of public transportation than
the private; and 5) avoided the departure time at peak hour or
provided an alternative departure time. The above-mentioned
performance measures justify the efficiency of the proposed
system for management of IPTS services in a metropolitan
area. In the future, incorporating the blockchain-based security
framework will provide a secured IPTS data and service
management in a metropolitan area.
REFERENCES
[1] V. D. Stanciu, C. Dobre, and V. Cristea, “Context-based service for intel-
ligent public transportation systems,” Int. J. Adapt. Innov. Syst.,vol.2,
no. 2, pp. 180–196, 2015.
[2] (2017). Google Now. Accessed: Apr. 28, 2017. [Online]. Available:
http://www.google.com/landing/now/
[3] (2017). Transport Urban. Accessed: Apr. 28, 2017. [Online]. Available:
http://www.transporturban.ro/
[4] P. Bellavista, A. Corradi, R. Montanari, and C. Stefanelli, “Context-
aware middleware for resource management in the wireless Internet,”
IEEE Trans. Softw. Eng., vol. 29, no. 12, pp. 1086–1099, Dec. 2003.
[5] L. Capra, W. Emmerich, and C. Mascolo, “CARISMA: Context-aware
reflective middleware system for mobile applications,” IEEE Trans.
Softw. Eng., vol. 29, no. 10, pp. 929–945, Oct. 2003.
[6] S. Garg, K. Kaur, G. Kaddoum, S. H. Ahmed, and D. N. K. Jayakody,
“SDN based secure and privacy-preserving scheme for vehicular
networks: A 5G perspective,” IEEE Trans. Veh. Technol., vol. 68, no. 9,
pp. 8421–8434, Sep. 2019.
[7] V.-D. Stanciu, C. Dobre, and V. Cristea, “Context-based service for
intelligent public transportation systems,” in Proc. IEEE Int. Workshop
Auton. Distrib. Syst. (ADiS) 8th Int. Conf. Complex Intell. Softw.
Intensive Syst. (CISIS), 2014, pp. 53–57.
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
CHAVHAN et al.: IoT-BASED CONTEXT-AWARE INTELLIGENT PUBLIC TRANSPORT SYSTEM IN METROPOLITAN AREA 6033
[8] (Jan. 2016). Real Time Traffic Flow Data. [Online]. Available:
http://pems.dot.ca.gov/
[9] A Multi-Agent Platform for Mobile C/C++ Agents. Accessed: Jun. 2015.
[Online]. Available: http://www.mobilec.org/overview.php
[10] I. Satoh, “Mobile agents,” in Handbook of Ambient Intelligence and
Smart Environments. New York, NY, USA: Springer, 2010, pp. 771–791.
[11] S. Chavhan and P. Venkataram, “Emergent intelligence: A novel com-
putational intelligence technique to solve problems,” in Proc. 11th Int.
Conf. Agents Artif. Intell. (ICAART), vol. 1, 2019, pp. 93–102.
[12] C. B˘
adic˘
a, Z. Budimac, H.-D. Burkhard, and M. Ivanovic, “Software
agents: Languages, tools, platforms,” Comput. Sci. Inf. Syst., vol. 8, no. 2,
pp. 255–298, 2011.
[13] S. Garg, K. Kaur, N. Kumar, G. Kaddoum, A. Y. Zomaya, and R. Ranjan,
“A hybrid deep learning-based model for anomaly detection in cloud
datacenter networks,” IEEE Trans. Netw. Service Manag., vol. 16, no. 3,
pp. 924–935, Sep. 2019.
[14] S. Chavhan and P. Venkataram, “Emergent intelligence based QoS
routing in MANET,” Procedia Comput. Sci., vol. 52, pp. 659–664,
Jun. 2015.
[15] S. Worrall, G. Agamennoni, J. Nieto, and E. Nebot, “A context-based
approach to vehicle behavior prediction,” IEEE Intell. Transp. Syst.
Mag., vol. 4, no. 3, pp. 32–44, Aug. 2012.
[16] A. R. Choudhary, “Policy-based network management,” Bell Labs Techn.
J., vol. 9, no. 1, pp. 19–29, 2004.
[17] Y. Lai, Y. Chen, Z. Liu, Z. Yang, and X. Li, “On monitoring and
predicting mobile network traffic abnormality,” Simulat. Model. Pract.
Theory, vol. 50, pp. 176–188, Jan. 2015.
[18] S. Garg, K. Kaur, S. H. Ahmed, A. Bradai, G. Kaddoum, and
M. Atiquzzaman, “MobQoS: Mobility-aware and QoS-driven SDN
framework for autonomous vehicles,” IEEE Wireless Commun., vol. 26,
no. 4, pp. 12–20, Aug. 2019.
[19] S. Chavhan and P. Venkataram, “Emergent intelligence technique-based
transport depot resource management in a metropolitan area,” J. Veh.
Routing Algorithms, vol. 2, nos. 1–4, pp. 23–40, 2019.
[20] Y. Wang, J. H. van Schuppen, and J. Vrancken, “On-line distributed
prediction of traffic flow in a large-scale road network,” Simulat. Model.
Pract. Theory, vol. 47, pp. 276–303, Sep. 2014.
[21] Y. Wang, J. H. van Schuppen, and J. L. M. Vrancken, “On-line dis-
tributed prediction of traffic flow in a large-scale traffic network,” in
Proc. ITS World, Oct. 2012, pp. 276–303.
[22] W. Suh, M. P. Hunter, and R. Fujimoto, “Ad hoc distributed simulation
for transportation system monitoring and near-term prediction,” Simulat.
Model. Pract. Theory, vol. 41, pp. 1–4, Feb. 2014.
[23] G. Rzevski and P. Skobelev, “Emergent intelligence in large scale multi-
agent systems,” Int. J. Educ. Inf. Technol., vol. 1, no. 2, pp. 64–71,
2007.
[24] W. D. Hillis, “Intelligence as an emergent behavior; or, the songs
of Eden,” in Daedalus. Cambridge, MA, USA: MIT Press, 1988,
pp. 175–189.
[25] Z. Li, C. H. Sim, and M. Y. H. Low, “A survey of emergent behavior
and its impacts in agent-based systems,” in Proc. IEEE Int. Conf. Ind.
Informat., 2006, pp. 1295–1300.
[26] P. J. Angeline, “Genetic programming and emergent intelligence,” in
Proc. Adv. Genet. Program., vol. 1, 1994, pp. 75–98.
[27] S. Garg et al., “En-ABC: An ensemble artificial bee colony based
anomaly detection scheme for cloud environment,” J. Parallel Distrib.
Comput., vol. 135, pp. 219–233, Jan. 2020.
[28] G. Smith, J. Sochor, and I. C. M. Karlsson, “Mobility as a service:
Development scenarios and implications for public transport,” Res.
Transp. Econ., vol. 69, pp. 592–599, Sep. 2018.
[29] S. Hietanen, Mobility as a Service. The New Transport Model,ITS
Canada, 2014, pp. 2–4.
[30] Transport Systems Catapult. (Oct. 2015). Traveller Needs and U.K.
Capability Study. Transport Systems Catapult. [Online]. Available:
https://ts.catapult.org.uk/wp-content/uploads/2016/04/Traveller-Needs-
Study.pdf
[31] M. Kamargianni, M. Matyas, W. Li, and A. Schafer, Feasibility Study
for “Mobility as a Service” Concept in London FS-MaaS Project,UCL
Energy Inst., London, U.K., May 2015.
[32] S. Chavhan and P. Venkataram, “Prediction based traffic manage-
ment in a metropolitan area,” J. Traffic Transport. Eng., Jul. 2019,
doi: 10.1016/j.jtte.2018.05.003.
[33] S. Garg, K. Kaur, S. Batra, G. Kaddoum, N. Kumar, and A. Boukerche,
“A multi-stage anomaly detection scheme for augmenting the security
in IoT-enabled applications,” Future Gener. Comput. Syst., vol. 104,
pp. 105–118, Mar. 2020.
[34] J. A. Guerrero-Ibanez, S. Zeadally, and J. Contreras-Castillo,
“Integration challenges of intelligent transportation systems with con-
nected vehicle, cloud computing, and Internet of Things technologies,”
IEEE Wireless Commun., vol. 22, no. 6, pp. 12–28, Dec. 2015.
[35] K. Dar, M. Bakhouya, J. Gaber, M. Wack, and P. Lorenz, “Wireless
communication technologies for ITS applications [topics in automo-
tive networking],” IEEE Commun. Mag., vol. 48, no. 5, pp. 15–62,
May 2010.
[36] S. Ghosh, T. Lee, and T. S. Lee, Intelligent Transportation Systems: New
Principles and Architectures. Hoboken, NJ, USA: CRC Press, Jan. 2002,
p. 192.
[37] B. U. Rai, Ed., Handbook of Research on Emerging Innovations in Rail
Transportation Engineering. Hershey, PA, USA: IGI Glob., May 2016.
[38] L. S. Passos and R. J. Rossetti, “Intelligent transportation systems:
A ubiquitous perspective. InNes trends in artificial intelligence,” in
Proc. 14th Portuguese Conf. Artif. Intell. (EPIA), Oct. 2009.
[39] R. Nasim, “Architectural evolution of intelligent transport systems (ITS)
using cloud computing,” Ph.D. dissertation, Dept. Math. Comput. Sci.,
Karlstads Universitet, Karlstad, Sweden, 2015.
[40] S. Chavhan and P. Venkataram, “Commuters’ traffic pattern and
prediction analysis in a metropolitan area,” J. Veh. Routing Algorithms,
vol. 1, no. 1, pp. 33–46, 2018.
[41] O. Urra, S. Ilarri, and R. Trillo-Lado, “An approach driven by mobile
agents for data management in vehicular networks,” Inf. Sci., vol. 381,
pp. 55–77, Mar. 2017.
[42] S. Chavhan and P. Venkataram, “Transport management for evacuation
of victims,” IEEE Trans. Emerg. Topics Comput. Intell., to be published,
doi: 10.1109/TETCI.2019.2940832.
[43] T. J. Soltman, “Effects of alternate loading sequences on results from
Chicago trip distribution and assignment model,” Highway Res. Rec.,
vol. 114, no. 114, pp. 122–140, 1965.
[44] R. J. Smock, “An iterative assignment approach to capacity restraint on
arterial networks,” in Proc. Highway Res. Board Bull., vol. 347, 1962,
pp. 60–66.
[45] S. Garg et al., “Edge computing-based security framework for big
data analytics in VANETs,” IEEE Netw., vol. 33, no. 2, pp. 72–81,
Mar./Apr. 2019.
[46] O. Urra and S. Ilarri, “Spatial crowdsourcing with mobile agents in
vehicular networks,” Veh. Commun., vol. 17, pp. 10–34, Jun. 2019.
Suresh Chavhan (M’18) received the B.E. degree
from SDMCET, Dharwad, India, in 2011, the
M.Tech. degree from NIT Karnataka, Surathkal,
India, in 2013, and the Ph.D. degree from the Indian
Institute of Science, Bengaluru, India, in 2019.
He is currently working as a Senior Assistant
Professor with the Automotive Research Center,
Vellore Institute of Technology, Vellore, India. He
has published over 13 in international/national jour-
nals/conferences and a book chapter. His research
areas includes ad hoc networks, autonomous and
electric vehicles, smart grid computing, and intelligent transportation system.
Deepak Gupta received the Ph.D. degree from
Dr. A.P.J. Abdul Kalam Technical University,
Lucknow, India.
He is an Eminent Academician and plays ver-
satile roles and responsibilities juggling between
lectures, research, publications, consultancy, and
community service, with 12 years of rich exper-
tise in teaching and two years in the industry. He
was a Postdoctoral Fellow with Inatel, Santa Rita
do Sapucaí, Brazil. He has authored/edited 37 books
with national/international level publishers. He has
published 92 research publications in reputed international journals and con-
ferences including 45 SCI indexed journals. He focuses on rational and
practical learning.
Dr. Gupta has served as an Editor-in-Chief, a Guest Editor, and an Associate
Editor for SCI and various other reputed journals. He has actively been an
organizing end of various reputed international conferences. He was invited as
a Faculty Resource Person/Session Chair/Reviewer/TPC member in different
FDP, conferences, and journals.
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.
6034 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 7, JULY 2020
B. N. Chandana (M’17) received the B.E. degree
from SIT Tumkur, Tumkur, India, in 2017, and the
M.Tech. degree from IIT Kharghpur, Kharghpur,
India, in 2019.
She is currently working as an Assistant Professor
with SIT Tumkur. Her research areas includes com-
munication network, unmanned aerial vehicles, and
resource allocation in heterogeneous networks.
Ashish Khanna received the B.Tech. and M.Tech.
degrees from GGSIPU, Dwarka, India, and the Ph.D.
degree from the National Institute of Technology,
Kurukshetra, India, in March 2017.
He has expertise in teaching, entrepreneurship,
and research and development. He has completed his
PDF with the Internet of Things Laboratory, Inatel,
Santa Rita do Sapucaí, Brazil. He has around 90
accepted and published research works in reputed
SCI, Scopus journals, conferences, and reputed book
series, including around 40 papers in SCI indexed
journals with cumulative impact factor of above 100. He has authored and
edited 20 books. He is a Series Editor in De Gruyter, Berlin, Germany, of
Intelligent Biomedical Data Analysis” series. His research interest includes
image processing, distributed systems and its variants, such as MANET,
FANET, VANET, and Internet of Things, machine learning, and evolutionary
computing.
Joel J. P. C. Rodrigues (S’01–M’06–SM’06–
F’20) received the five-year B.Sc. degree (licentiate)
in informatics engineering from the University of
Coimbra, Coimbra, Portugal, in 1995, the M.Sc. and
Ph.D. degrees in informatics engineering from UBI,
Covilha, Portugal, in 2006 and 2002, respectively,
and the Habilitation degree in computer science and
engineering from the University of Haute-Alsace,
Mulhouse, France, in 2014.
He received the Academic Title of Aggregated
Professor of informatics engineering with UBI in
2015. He is a Professor with the Federal University of Piauí, Teresina, Brazil,
and a Senior Researcher with the Instituto de Telecomunicações, Lisbon,
Portugal. He is a member of many international TPCs and participated in
several international conferences organization. He has authored or coau-
thored over 800 papers in refereed international journals and conferences,
three books, two patents, and one ITU-T Recommendation.
Prof. Rodrigues is an Editor-in-Chief of the International Journal on
E-Health and Medical Communications and the Editorial Board member
of several high-reputed journals. He has been the General Chair and the
TPC Chair of many international conferences, including IEEE ICC, IEEE
GLOBECOM, IEEE HEALTHCOM, and IEEE LATINCOM. He is the Leader
of the Next Generation Networks and Applications Research Group, the
Director for Conference Development—IEEE ComSoc Board of Governors,
the IEEE Distinguished Lecturer, the Technical Activities Committee Chair of
the IEEE ComSoc Latin America Region Board, the Past-Chair of the IEEE
ComSoc Technical Committee on eHealth and on Communications Software,
and the Steering Committee Member of the IEEE Life Sciences Technical
Community and Publications Co-Chair.
Authorized licensed use limited to: J.R.D. Tata Memorial Library Indian Institute of Science Bengaluru. Downloaded on August 07,2020 at 05:36:37 UTC from IEEE Xplore. Restrictions apply.