Multi-agent-Driven Safety Information

Dissemination in V2I Networks

Ramesh B. Koti and Mahabaleshwar S. Kakkasageri

Abstract The intelligent information exchange techniques in vehicular adhoc net-

works (VANET) have gained huge attention in intelligent transport system (ITS).
ITS connects people on the road, cars, and fixed infrastructure using wireless com-
munication technologies to solve different traffic-related difficulties. Dissemination
of information is the basis of communication that plays a major role in VANETs and
has recently become an active field of research. Vehicle-to-infrastructure (V2I) com-
munication is an essential study topic for enhanced data exchange for implementing
security and integrity in communication. In this, we are proposing an effective data
dissemination mechanism in V2I network scenario using multiple software agents.
These agents are set up and used for the construction of high-quality communication
paths between the nodes. The intermediary connections are picked based on each
of the moving nodes radio signal strength and relative mobility. This protocol cre-
ates an environmentally beneficial system that also improves fuel economy by opti-
mum traffic management techniques. To gather and distribute safety information, we
employed a three-layer vehicle to infrastructure (V2I) network architecture with static
and dynamic intelligent agents in the proposed study. The suggested algorithm per-
forms better in terms of coverage area, lossless transmission, and decreased latency
in a highway networks scenario with medium traffic density. Finally, qualitative com-
parison is made with present V2I system and found the significance improvement in
its performance metrics.

Keywords V2I · Multi-agent · Vehicle manager agent (VMA) · Knowledge base

(KB) · Routing node (RN) · High-quality link

R. B. Koti (B)
Electronics and Communication Engineering Department, Gogte Institute of Technology
(Autonomous), Belagavi 590008, Karnataka, India
M. S. Kakkasageri
Electronics and Communication Engineering Department, Basaveshwar Engineering College
(Autonomous), Bagalkot 587102, Karnataka, India
e-mail: mskec@[Link]

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 785
V. Suma et al. (eds.), Evolutionary Computing and Mobile Sustainable Networks,
Lecture Notes on Data Engineering and Communications Technologies 116,
[Link]
786 R. B. Koti and M. S. Kakkasageri

1 Introduction

Vehicular adhoc networks (VANETs) provide new ways to increase traffic safety and
information distribution efficiency. V2I communication involves the broadcasting of
recognized hazardous occurrences to neighboring vehicles using unicast, point to
multipoint (PMP), or broadcast communication systems, in which a single vehicle’s
message is received by many receivers. When vehicles are traveling in difficult cir-
cumstances such as heavy fog, rain, or scenarios such as difficult to see cyclists or
pedestrians running across the road, when these events are communicated in advance
with all neighbors to overcome the damage together, this is referred to as a cooperative
intelligent transportation system (C-ITS). A signal at a crosswalk, a digital signboard
beside the road, and specialized equipment known as roadside units (RSU) are all
examples of infrastructure-based network. The received signals from the source cars
can be amplified and routed by RSU for onward transmission. The C-ITS used in
connected vehicles focuses on digital technology for the information sharing via
wireless communication channels.

1.1 Related Works

The vehicular adhoc network (VANET) is a viable answer to today’s complicated

transportation problems. Vehicles on the road that identify dangerous circumstances
and are capable of sharing information collaboratively to prevent additional harm are
an obvious example of ITS. The goal of ITS is to improve transportation safety and
mobility by combining a variety of technologies and applications. It will enhance
people’s productivity while minimizing traffic’s adverse effects. Road congestion
and accidents are one of the most prominent causes for such an adverse impact on
traffic. Road accidents have claimed the lives of millions of people and wounded
millions more. The impact of these occurrences can be mitigated if information is
disseminated quickly. We need to handle these challenges as a global concern. By
employing appropriate distribution algorithms, ITS will assist us in improving road
safety. Researchers, academia, and industry are now paying close attention to the
dissemination of safety information in order to enhance vehicle traffic conditions,
but they really make the transport industry more convenient and effective. Post-crash
notification, proactive collision warning, parallel parking assistance, emergency elec-
tric brake light, left/right turn assistance, traffic signal violation warning, and other
information broadcasts are all subject to strict time limitations.
Researchers have proposed a variety of works on information distribution to
reduce packet transmission delay. In [1], it provides a comprehensive examination of
dissemination focusing on the critical problems of privacy and data security, as well as
data forwarding solutions that use fixed infrastructure or centralized administration.
Vehicles in a priority-based traffic flow scheme are assigned a sequence ID depending
on their speed, direction, and opposite side front. By combining a continuous mon-
Multi-agent-Driven Safety Information … 787

itored and disciplined style of dissemination protocol, the wireless communication

module (NRF24L01) and Arduino module are utilized for sustainable traffic control.
Using the GDSOM-P2P method, the data searching algorithm collects the critical
data about the road characteristics and then distributes it to neighbors [2, 3]. In [4,
5], real-time traffic management can be achieved by monitoring V2I communication
using an agent-based framework that comprises heavyweight static cognition (based
on belief desire intention: BDI) and lightweight mobile agencies in a road works
scenario. Based on the information relevancy, it does push (gather/store and dissemi-
nate) and pull (gather/store) actions. With a central traffic control database connected
by multihop wireless technologies, the vehicle preemption approach improves the
possibilities of emergency message distribution. The dataflow control is provided
by the different dashboard controls on the logical platform architecture, which are
based on DSRC connection [6].
The authors used Wi-Fi, DSRC, and LTE to demonstrate the potential of V2V
and V2I in a Het-Net context, ensuring the best possible use of current commu-
nication options while reducing backhaul telecommunications network and taking
into consideration connected vehicle application needs [7]. The proposed technique
has important implications in [8], including overcoming the limitations of K-means-
based clustering and increasing clustering accuracy as a crucial tool in data mining
and expert systems.
A systematic reading and analysis of the relevant literature produced extremely
valuable insights such as motivations, challenges, issues, and suggestions in connec-
tion to V2I systems and performance measures using traffic data and methodologies
[9, 10]. In [11], it represents an adaptive load balancing method (RBO-EM) for
effective data distribution with reduced end-to-end latency, whereas in [12], it talks
about the secure cryptography-based cluster mechanism (SCCM) for MANET. Safe
routing, encryption, signature creation, signature verification, and decryption were
all steps in the development of the AOMDV routing protocol.
The global timeout method and the anti-packet broadcast scheme, where con-
trol signals are given in social-based end-to-end and local-based adhoc ways, are
two common systems for lossless data transfer [13]. A trust evaluation approach is
developed using the Dempster–Shafer theory. The method of evaluation is based on
two forms of trust: direct and indirect trust, with the latter being more important
[14]. Better results in terms of connection latency and transmission speed proved the
reliability of the fragmentation technique. For effective traffic management, vibrant
mapping of highway traffic with the collection of a digital map via swarm mapping
with grouping is used to enable path optimization to reduce reaction time delays [15].
Our suggested work’s limitations are limited to increased vehicle density, which is
prevalent in highway scenarios in order to maintain consistent intermediate linkages.
In the future scope, the issue of decreased vehicle density will be addressed.
788 R. B. Koti and M. S. Kakkasageri

Fig. 1 V2I communication scenario in VANETs

1.2 Proposed Work

The suggested work uses several static and dynamic agents to fulfill two primary func-
tions via the module’s node agency and RSU agency. In a vehicle-to-infrastructure
(V2I) architecture, the link quality network protocol (LQNP) is a multihop wireless
protocol that allows for the fast broadcast of safety information. The nodes in the net-
work under examination might fall into one of three categories, as described below.
The communication method distributes traffic across its enabling devices in a tree-
like topology. The architecture of the LQNP protocol is seen in Fig. 1. It comprises
three tiers of nodes: road side units (RSU), routing nodes (RN), and end nodes.
RSU and RN will handle packet routing, while mesh equipped end nodes will link
the RSU to the forwarding nodes for information dissemination. Hundreds of end
nodes, numerous routing nodes (RN), and few RSUs along the road side are presented
for redundancy and workload sharing to form flexible network environment. Based
on their distance and radio signal strength (RSS) level, any end node can be set to
operate as RN. A LQNP end node works in power sleep simplex mode for battery
operation and does not execute information redistribution until it is awakened by
external stimuli or a clock. Table 1 shows the predicted relationship between RSS
levels and related connection possibilities for a vehicle density of 100 in a road length
of 1000 m, based on current network scenarios.
Multi-agent-Driven Safety Information … 789

Table 1 RSS thresholds with connection possibilities

Case RSS levels (db) Vehicle mobility Predicted connections
(km/h) (%)
1 30 80 70
2 40 80 60
3 50 80 50
4 60 80 40
5 70 80 30

1.3 Our Contributions

There are many methods for disseminating safety information that are currently in
use. We proposing consistent connection-oriented safety information dissemination
protocol in this part, which comprises the following set of actions: (1) development
of static and dynamic cognitive agents for the gathering and dissemination of safety
information. (2) The belief desire intention (BDI) model was used to identify critical
and non-critical road danger occurrences. (3) Using the RSS levels of receiving cars,
establish congestion-free wireless connections for information transfer. (4) Static
agents are triggered to collect information on target cars, and mobile agents are acti-
vated to disseminate that information via error-free bandwidth-oriented networks.
For instance, when a set of connections is available, the static agent always chooses
the higher bandwidth link for dissemination. Rest of the paper is presented as follows.
Safety information dissemination using cognitive agents is described in Sect. 2. Sim-
ulation model and performance parameters are described in Sect. 3, Sect. 4 includes
the result analysis, and finally Sect. 5 concludes the paper.

2 Multi-Agent-Based Safety Data Dissemination Scheme

The roles of static and mobile agents used to improve the data dissemination process
are described in this section. The techniques for node categorization and node search
are described mathematically, and a network scenario is provided.

2.1 Network Scenario

The suggested network scenario includes a large number of densely packed cars on the
road separated by a significant distance. Each vehicle has a GPS navigation system,
event collection sensors, communication (TX and RX) capability, and a processing
unit. All of the cars are expected to respect the road lane directions and be able to
790 R. B. Koti and M. S. Kakkasageri

interact with their neighbors utilizing the 802.11 wireless communication standard.
Clusters are produced by considering the city environment with curving road lanes
and top structures, as well as the relative geographical area in the direction of vehicle
moments. The equipment in the cars’ onboard systems is expected to enable agent-
based platforms. Platform-independent agent programmers launched on each vehicle
resolve compatibility concerns. Each end node has a mobile agent, a static agent, and
a knowledge base expert system, all of which help with information forwarding via
a secure communication channel. The codes used by agents are used to interact with
other agents in order to create a greater level of collaboration in order to achieve the
given goal.

2.2 Software Agents

An intelligent agent is a self-contained entity that uses its knowledge base to achieve
its objectives. Software agents are self-contained programmers that run on a host’s
agent platform. Agents employ their own knowledge base to accomplish certain tasks
without interfering with the host’s operations. Mobile agents are modular, adaptable
entities that may be built, moved, deployed, and destroyed in real time. Mobile code
should be platform agnostic, meaning it can run on any remote host in a mixed
network environment. Node information retrieval agent (NIRA), vehicular manager
agent (VMA), RSU vehicular manager agent (RVMA), and dissemination agent
(DA) are some of the static and dynamic agents used in the proposed dissemination
protocol.

2.3 Definitions

Belief Generation: Based on the installed sensors, beliefs are formed and compared
to the node parameters. For the data collecting agent, the information tuple consists
of source ID, destination ID, TTL, RSS level, mobility, and distance. Because the
node with distance and mobility is a critical parameter for the RN selection for the
node agency, beliefs are updated with these characteristics.
Desire Updation: The distance and mobility properties of nodes are critical for
RN node selection; the desires are to identify nodes with distance and mobility
parameters.
Intention Based on Desires: Based on the parametric values for mobility and dis-
tance for all the database collected in the interval t, the lowest value of k in KNN
model generates the intention of lowest value of distance and mobility.
Cluster Mobility Pattern: It is defined as the collective pattern of vehicle movement
in a cluster.
Lane Intersection Pattern: It is the collective segment of lane intersection points
with one another.
Multi-agent-Driven Safety Information … 791

Vehicle Manager Agent (VMA)

Belief Generator Variable Desire Extraction Agent

Declaration Migration /
Node Arrival
Information
Retrival
Routing Event Detection Agent
knowledge Sensors Revisions Intensions (NIRA)
Base
(RKB) Agent
Migration /
Arrival
Node Discovery BDI Model Based Vehicle
Node Search Configuration
Agent
(VCA)

Fig. 2 Node agency functional model

2.4 Node Agency

Each vehicle will have a node agency and an RSU that will send invitation commands
to the lower tier nodes; if the request is accepted, the end nodes will join the cluster.
Figure 2 depicts the functional components of node agency.

• Node Information Retrieval Agent (NIRA): It is a mobile agent that will be

activated by VMA at regular intervals to update the knowledge base’s node state. By
saving a bit in the header for acknowledgment status, it hosts the HELLO packets
in the cluster. The acknowledgment status of end nodes and RN in the coverage
range is collected by this agent. As a result, it establishes communication across
three layers of nodes, forming a network of high-quality links. NIRA updates the
latest network topology dynamics by refreshing the HELLO packet broadcasting
every 5ms. The low refresh rate aids the updated topology nodes in catching all
essential events.
• Routing Knowledge Base (RKB): This is a centralized data storage system that
keeps track of the state of network nodes. NIRA keeps this up to date and uses it
to help the VMA in detecting end nodes. RN gateways will use this information
to conduct optimal routing. Every 10 ms, the knowledge base will be updated.
For fast uplink information distribution, RKB implements data gathering in a tree
hierarchical structure. For data integrity, it employs the BDI software model.
• Vehicle Manager Agent (VMA): VMA is installed on each end node and activates
the mobile agent NIRA, which updates the knowledge base with vehicle status
parameters. It manages and coordinates the BDI model’s functions. It is aware
of all node parameters, including node position, movement, RSS level, and road
conditions. VMA, when used with a BDI model, generates beliefs based on the
data obtained.
792 R. B. Koti and M. S. Kakkasageri

2.5 K-Nearest (KNN) Neighbor Algorithm

It is a nonparametric instance-based classification technique that looks for nodes

utilizing computation approximation by majority parametric values and K value. The
KNN algorithm determines the criticality of data. Select the number K of neighbors
in a cluster and compute the Euclidean distance between them as follows:

d( p, q) = (q1 − p1 )2 + (q2 − p2 )2 + · · · (qn − pn )2 (1)

In Eq. (1), take the K-nearest neighbors as per the calculated Euclidean distance and
among these k neighbors find the RSS level for each node to satisfy the condition
RSS>=threshold value. 
 n

d( p, q) =  (q1 − p1 )2 (2)
i=1

Extract the safety data for each category by counting the number of nodes in each
cluster. Assign the system ID and device ID to the node that has the highest RSS. We
found the nearest neighbors as three nearest neighbors in category A and two nearest
neighbors in category B by computing the Euclidean distance indicated in equation
(2). This formula is used by KNN to calculate the distance between each RSS level
of the node in question and the threshold RSS level. It then calculates the likelihood
of a node having high-quality connections to RN among all available nodes in the
cluster, as shown in Eq. (3)

1
p(Y = j|X = x) = I (Y i = j) (3)
k i=A

where X is output class, k is the neighbor distance, and I(Y ) is the decision boundary.
The node discovery by the mobile agent with respect to available information tuple
is given by 
 n

E(d) = (X, X i ) =  (X j − X i j )2 (4)
i=1

Since the usage of software static and dynamic agents works concurrently with the
inherent routing functions in the V2I model’s three-layer architecture, there is less
end-to-end latency and congestion. The mobile agents may move between nodes and
establish a database, which is accessed by the RSU for intelligent routing and forward-
ing information that is updated in real time. The current system, which employs an
agent-based forwarding technique, can provide routing flexibility while also enabling
heterogeneous communication for quick message delivery. The system is organized
in a layered design that allows for distributed processing. For greater coverage, traffic
signals and digital sign boards can participate as additional RSUs for location-based
services and information dissemination in the neighborhood region.
Multi-agent-Driven Safety Information … 793

2.6 RSU Agency

It uses the static and dynamic agency components, namely RVMA, KBES, IFA, and
DA for its optimum functionality. Here the system finds the target identification and
route discovery and establishes the error-free links approaching the target vehicle.
The functions of RSU agency components are as follows.
• Knowledge Base Expert System (KBES): It is the data center storage area where
the routing node information and optimum hops to the target nodes are fetched and
updated by the IFA. VMA uses this data during safety information dissemination
after route discovery. The information includes RN ID, Node ID, TTL time stamp,
and available bandwidth.
• Information Fetch Agent (IFA): It is a mobile agent which will be triggered by
RVMA at regular intervals to collect the information about RN nodes, RSU, and
available bandwidth for error-free dissemination. IFA agent is a mobile agent that
configures the end nodes on the air with RSS level threshold setting, data com-
munication modes, and required packet TTL intervals. This provides the flexible
operating modes of system function. It uses four byte addressing scheme for the
header excluding the payload. IFA agent in connection with the KBES synchro-
nizes all the events triggered by three-level network nodes that are associated with
regional cluster.
• RSU Vehicle Manager Agent (RVMA): It is an static agent deployed in RSU
having the responsibilities of coordinating the activities of information dissemi-
nation. It triggers the IFA to collect the information about target vehicles based
on the pre-stored route maps and finding the error-free route based on the prob-
abilistic node searching algorithm. It uses the knowledge base data to take the
appropriate decisions on route calculation and triggers the safety information dis-
semination. This unit is responsible for discovering an appropriate neighbor based
on node geographic position, vector information, and predicted future direction.
It uses HELLO messages to discover neighboring nodes. The next hop selection
algorithm is used to find an appropriate neighbor for data forwarding.

2.7 Features of Proposed Model

The well-defined procedural steps in algorithms aid in the construction of program-

ming and make the relationship between input and output parameters easier to com-
prehend. The following are some of the key aspects of the suggested model:
• Node Hierarchy: During node classification process, broadcasting higher level
of RSS level for accepting the connections creates fever communication links in
network which leads to error-free information dissemination.
• Dynamic Connections: Self-healing nature of end nodes in mesh networks con-
tinuously checks the alternate connections by comparing the hop level and signal
794 R. B. Koti and M. S. Kakkasageri

strength of received HELLO packets. In cases where the primary link becomes
failed, the device will automatically change to the alternate routing if such routing
is possibly available. If the alternate routing is also unresponsive, the device will
enter a state where it searches for new routing possibilities.
• Cognitive Agents: Cognitive agents perspective in the dissemination process ini-
tiates the faster end to delay during packet transfer. Information is preserved in the
knowledge base until the valid connection is established, thus this scheme ensures
the guarantee of service.

2.8 Algorithm Properties

The characteristics of proposed algorithm are well defined in terms of sequence of

actions performed, during information dissemination. Some of the useful properties
of algorithm are as follows.
• Property 1: The dissemination process is secure and fast because of high per-
formed link participation in the communication. The system has lower control
overhead during connection setup due to the simple registration setup with finite
size HELLO packets.
• Property 2: The knowledge base provides the nearby digital data centers along
the road and traffic signal centers for enhanced target coverage which significantly
increases the packet delivery.
• Property 3: This algorithm tends to minimize the delay by reducing the frequent
handover using RSU wired communication which provides the stable connectivity
for dissemination.
• Property 4: Automatic registration process of cluster members, RN and on air
configuration of end nodes for RSS thresholds gives the better control over the
speed and congestion in the network.

3 Simulation

We used C++ developer to evaluate our suggested method, the LQNP protocol, which
takes use of the V2I naming scheme. To generate mobility situations, we randomly
placed a set of vehicles (mainly 100) on a 5-km route. Different road incidents (e.g.,
accidents, poor roads, etc.) can have a significant impact on traffic conditions. The
distribution of event notification to cars traveling to the event location may aid in
making timely decisions, such as rerouting or slowing down. The LQNP uses the
BDI machine learning algorithm to give a smart manner of broadcasting. Vehicle
mobility patterns, as a result of variable traffic dynamics and road behavior, compli-
cate vehicular information distribution. Broadcast storm and inconsistent connection
are the two primary problems in a busy highway traffic scenario.
Multi-agent-Driven Safety Information … 795

Table 2 Simulation parameters

Parameter Considered values
Highway range 5 km
Number of nodes 100
File size 5 kb
Number of segments 4
Transmission range 75–100 m
Antenna model Unidirectional
Vehicle density per segment 10
TTL of packet 500 ms
Traffic pype Moderate
Connecting establishment time 2000 ms
Vehicle mobilty 10 m/s, 20 m/s, 30 m/s

3.1 Simulation Procedure

The simulation input parameters are summarized in Table 2. Simulation procedure for
the proposed scheme is as follows: (1) Generate VANET network scenario in given
road length of 10KM by deploying the vehicles based on geographical clusters. (2)
Each vehicle maintains a data structure to store information as specified by scheme
(RSS level, mobility, and distance from RN node.). (3) Generate the mobile and
static agency to deliver the safety information and vehicle parameters to the RN node
(agents are implemented as objects). (4) Apply mobility to vehicles. (5) Randomly
generate the vehicle parameters at each vehicle and select RN members using the
agency. (6) Use agency to identify RN and announce the intersection mobility pattern.
(7) Compute the performance of system.

3.2 Performance Metrics

Some of the performance metrics evaluated are packet delivery ratio, energy con-
sumption, dissemination delay, target coverage, handoff delay, and success rate.
• Energy Consumption: It is defined as the amount of energy necessary to estab-
lish a connection and transport packets. It is measured in millivolts (mV) and
represented in terms of individual nodes.
• Packet Delivery Ratio: It is defined as the ratio of total packets received to total
packets sent over a certain time span. The maximum time interval for recording
the PDR was one TTL duration. It is expressed in percentage.
• Dissemination Delay: It is referred to as overhead delay because it contains the
time component associated with queuing delay, which occurs when data switching
phases in intermediary nodes retain data for a random length of time. It is measured
in ms.
796 R. B. Koti and M. S. Kakkasageri

90

80
µ = 30
20
µ = 10
Energy Consumption (in mw)

60 µ = 10
µ = 20
µ = 30

40

20

Proposed Scheme
RBO-EM Algorithm

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Number of Nodes

Fig. 3 Energy consumption in mw versus number of nodes

• Handoff Delay: It is the time it takes for packets to get from the end mobile agent
to the RN nodes, the RN nodes to the RSU, and the RSU to the RSU. It is the sum
total of all three delays. It is counted in milliseconds (ms).

4 Result Analysis

To design the simulation, we utilized developer C++ and Python scripting. Both are
used to depict traffic flow on roadways that are connected to roadside infrastructure
through a transmission control protocol (TCP) connection, with vehicle movement
represented by changing vehicle mobility. Table 2 lists the additional parameters
utilized in the simulation. The interactive user interface was utilized to deploy nodes
randomly in the simulated scenario. Vehicle movements were planned from source
to destination via several intermediary nodes, and vehicle routes were meant to last
at least 10 s. The vehicles per kilometer were restricted to a minimum of 20 and a
maximum of 25. The vehicle speed ranged from 10 to 30 m/s, and each vehicle’s
transmission range was set at 100 m. The simulation was run ten times, with the
averages presented.
The suggested technique minimizes energy consumption in network infrastruc-
ture by limiting handover between nodes and selection of relaying nodes based on
signal strength. Figure 3 shows the energy consumption simulation results for various
node counts and mobility levels using the proposed method. Due to the availability of
Multi-agent-Driven Safety Information … 797

95

Packet Deliver Ratio (in %) 80

60

40

20

Proposed Scheme
RBO-EM Algorithm

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
RSS Levels in db

Fig. 4 PDR versus RSS Levels

95
90
µ = 10m/s
80
µ = 20m/s
µ = 30m/s
Dissemination Delay (in ms)

60
µ = 30m/s
µ = 20m/s

µ = 10m/s
40

20

Proposed Scheme
RBO-EM Algorithm

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Number of Targets

Fig. 5 Dissemination delay in ms versus number of targets

798 R. B. Koti and M. S. Kakkasageri

50

40 µ = 20m/s
µ = 10m/s
µ = 30m/s
Hand-Off Delay (in ms)

30 µ = 30m/s
µ = 10m/s
µ = 20m/s

20

10
Proposed Scheme
RBO-EM Algorithm

0
0 5 10 15 20 25 30 35 40
Number of Target Vehicle Covered

Fig. 6 Handoff delay in ms versus number of target vehicle covered

reliable connections between the nodes, the proposed approach improves outcomes
with a reduced energy consumption of 45 mv when the number of nodes is approxi-
mately 50. When compared to the RBO-EM method, the proposed algorithm utilizes
10% less energy.
The packet delivery ratio in a communication is the percentage of received pack-
ets to transmitted packets. The effect of a greater RSS level on PDR is seen in
Fig. 4. When the RSS level is low, certain messages fail to deliver under the sug-
gested approach. However, when the RSS level rises, more nodes are likely to join
in communication, lowering the overhead latency and therefore improving PDRs.
The outcome depicts PDR with varying vehicle mobility levels. In comparison with
the RBO-EM method, the suggested technique improves PDRs by 5%, 6%, and 7%,
respectively. The nature of curves for other models has lower bent due to the absence
of quality links in the communication path and delays involved in hand on process.
Figure 5 depicts the dissemination delay with various numbers of information
receiving vehicles (targets). To obtain the best latency and reachability for a large
number of targets, the cognitive multi-agents and probabilistic dispersion approach
are used to establish a trustworthy path to the intended receivers. Because each
target uses bandwidth for information exchange, the dissemination delay rises as the
number of target nodes grows.
The handoff delay is made up of packet transmission notification (handshaking),
path update information transmission (queuing), and mobile node processing delay.
For a effectual analysis, the various mobility levels are taken into account while
calculating the handoff latency. Figure 6 shows the simulated results, which show that
Multi-agent-Driven Safety Information … 799

the suggested system has a 2RBO-EM scheme. The better outcomes are attributable
to the fever intermediary connections, which have larger bandwidth and latency
capabilities for information exchange.

5 Conclusion

We proposed a V2I communication protocol to establish stable communication path

with higher values of RSS levels to achieve lossless and secure information transport
in this research project. The use of agents ensures that vehicle data are updated and
disseminated on a regular basis. Furthermore, having a worldwide navigational sys-
tem in OBU aids in time synchronization and dispersed distribution management.
The three-tier design, along with the delegated communication technique, quickly
generates a broad coverage area for rapid safety information distribution. The produc-
tivity of the V2I system is increased by better network performance characteristics
such as end-to-end latency, energy consumption, and resource use. As a result, a
95 percent success rate is visible. The current innovative multilayer communication
simulation was performed in developer C++ and produces excellent results with a
95 percent success rate in terms of coverage and broadcast efficiency. Working with
reduced node density for single-lane road settings is part of the future scope. For
timely synchronization among the V2I interface, the static and mobile agent main-
tains routing information and refreshes the knowledge base regularly for fast data
delivery.

References

1. Ghebleh, R.: A comparative classification of information dissemination approaches in vehicular

ad hoc networks from distinctive viewpoints: a survey. Comput. Netw. 131, 15–37 (2018)
2. Azimi, R., Sajedi, H.: A decentralized gossip-based approach for data clustering in peer-to-peer
networks. J. Parallel Dist. Comput. 119, 64–80 (2018)
3. Markus, P., Matt, G., Sielen, N.: “Real-time traffic management analyzing V2I communication
at the edge of the network. In: IEEE International smart cities conference (ISC2). Kansas City,
MO 2018, pp. 1–2 (2018). [Link]
4. Kakkasageri, M., Manvi, Sunil: Safety information gathering and dissemination in vehicular ad
hoc networks: cognitive agent based approach. Wireless Personal Commun. 69 (2011). https://
[Link]/10.1007/978-3-642-17878-8-26
5. Manvi, S., Kakkasageri, M., Pitt, J.: Multiagent based information dissemination in vehicular
Ad Hoc networks. Mobile Inf. Syst. 5, 363–389 (2009). [Link]
6. Chen, J., Mao, G., Li, C., Zhang, D.: A topological approach to secure message dissemination
in vehicular networks. IEEE Trans. Intell. Transp. Syst. 21(1), 135–148 (2020). [Link]
10.1109/TITS.2018.2889746
7. Dey, K.C., et al.: Vehicle-to-vehicle (V2V) and vehicle-to infrastructure (V2I) communication
in a heterogeneous wireless network performance evaluation. Transp. Res. Part C Emerg.
Technol. 68, 68–184 (2016)
800 R. B. Koti and M. S. Kakkasageri

8. Azimi, R., Ghayekhloo, M., Ghofrani, M., Sajedi, H.: A novel clustering algorithm based on
data transformation approaches. Expert Syst. Appl. 76, 59–70 (2017)
9. Malik, R., Alsattar, H., Ramli, K.N.B., Bahaa, B., Zaidan, A., Hashim, Z., Ameen, H., Garfan,
S., Mohammed, A., Zaidan, R.: Mapping and deep analysis of vehicle-to-infrastructure com-
munication systems: coherent taxonomy, datasets, evaluation and performance measurements,
motivations, open challenges, recommendations, and methodological aspects. IEEE Access 1
(2019). 10.1109/ACCESS.2019.2927611
10. Li, T.-H., Khandaker, M.R.A., Tariq, F., Wong, K.-K., Khan, R.T.: Learning the wireless V2I
channels using deep neural networks. arXiv e-prints (2019)
11. Ullah, S., Abbas, G., Abbas, Z.H., Waqas, M., Ahmed, M.: RBO-EM: reduced broadcast
overhead scheme for emergency message dissemination in VANETs. IEEE Access 8, 175205–
175219 (2020). [Link]
12. Mohindra, A.R., Gandhi, C.: A secure cryptography based clustering mechanism for improving
the data transmission in MANET. Walailak J. Sci. Tech. 18(6):8987 (2021). [Link]
48048/wjst.2021.8987
13. Chen, P.-Y., Cheng, S.-M., Sung, M.-H.: Analysis of data dissemination and control in social
internet of vehicles. IEEE Internet of Tings J. 5(4):2466–2477 (2018)
14. Mudengudi, S., Kakkasageri, M.: Agent based trust establishment between vehicle in vehicular
cloud networks. Int. J. Comput. Netw. Inf. Secur. 11, 29–36 (2019). [Link]
ijcnis.2019.07.05
15. Haoxiang, W., Smys, S.: Enhanced Vanet routing protocols for dynamic mapping in real time
traffic. IRO J. Sustain. Wireless Syst. 01, pp. 139–147 (2019). [Link]
2019.3.001