Multi Agent Protocol for Cooperative Intersection Collision

Avoidance System

Noor Cholis Basjaruddin

, Dwi Hendratmo Widyantoro

, Saufik Ramadhan

and Umar Zaenal

Abidin

Department of Electrical Engineering – Bandung State Polytechnic, Bandung, Indonesia

School of Electrical Engineering and Informatics - Bandung Institute of Technology, Bandung, Indonesia

Keywords: cooperative intersection collision avoidance system, wireless access in vehicular environment, agent protocol

Abstract: One of the safety devices embedded in autonomous vehicles is the Intersection Collision Avoidance System

(ICAS). In connected vehicle environment, ICAS can be developed into cooperative ICAS (CICAS), namely

ICAS which has the ability to cooperate with other vehicles. CICAS can be realized using V2V

communication, among others, with Wireless Access in Vehicular Environment (WAVE) technology. Data

exchange between vehicles equipped with CICAS is governed by a special protocol. This paper presents the

multi agent protocol for the purposes of the CICAS that are designed with the WAVE architecture

environment. Simulation results demonstrate that the proposed protocol can improve the safety index of

CICAS, length of intersection passing time, and total time for data exchange.

1 INTRODUCTION

The emergence of WAVE (Wireless Access in

Vehicular Environments) encourages cooperative

vehicle development, namely vehicles that can work

together to improve safety and driving comfort.

Through the WAVE between vehicles can

communicate by exchanging important data such as

the position, speed, and direction of the vehicle.

These important data can then be used by drivers,

Advanced Driver Assistance Systems (ADAS)

(Basjaruddin, et al., 2018), or autonomous vehicles to

choose the right maneuver so that the vehicle can

avoid accidents and / or inconvenience.

Intersection Collision Avoidance System (ICAS)

is a device that can help the driver to avoid accidents

at intersection. ICAS uses sensors that can alert or

braking vehicles suddenly when they will collide with

other vehicles that will cross the intersection.

Weakness of ICAS is a very limited operating area,

when the vehicle is near the intersection. Sudden

braking due to working ICAS certainly reduces

passenger comfort.

The weakness of ICAS is overcome by the

development of ICAS which has cooperative

capability or known as Cooperative Intersection

Collision Avoidance System (CICAS). CICAS

allows multiple vehicles to work together when

crossing the intersection so safety is guaranteed. The

wider scope of the CICAS area is also useful for

maintaining the comfort of passenger vehicles.

Research in the development of ICAS is carried

out in (Basma, et al., 2011) , (Yang, et al., 2016), and

(Elleuch, et al., 2017). Basma,et.al. (2011) develops

the intersection collision avoidance (ICA) using

telematics and wireless sensor networks (WSN) to

monitor approaching traffic, detect possible collision,

and then transfer information to drivers, warning

them of high collision probability. Cooperative

driving model for unsignalized intersections

developed by Yang, et.al. (2016). In this research

model was developed with reduplicate dynamic game

theory. Elleuch, et al. (2017) study cooperative

intersection collision avoidance based on V2V

communication and real-time databases.

This paper will discuss the design of the protocol

which serves to regulate the exchange of data

between vehicles so that passing the intersection can

be carried out safely. The protocol design was then

tested with a simulator program developed with a

cooperative multi agent system approach using BDI

(belief desire intention) agents (Basjaruddin, et al.,

2015).

Basjaruddin, N., Widyantoro, D., Ramadhan, S. and Abidin, U.

Multi Agent Protocol for Cooperative Intersection Collision Avoidance System.

DOI: 10.5220/0009011003350340

In Proceedings of the 7th Engineer ing International Conference on Education, Concept and Application on Green Technology (EIC 2018), pages 335-340

ISBN: 978-989-758-411-4

335

2 RESEARCH METHOD

In a connected vehicle environment between vehicles

can exchange information on speed, acceleration,

heading, position, and other attributes periodically

every 20 ms. This information exchange can be done

using WAVE technology. In addition to periodic data

exchange, each vehicle can also exchange special

data related to special conditions such as notifying

that the vehicle will cross a intersection. Periodic data

exchange is carried out by broadcast, while special

data exchange is carried out by unicast. Unicast mode

in WAVE can be used to support cooperative

technology.

2.1 Wireless Access in Vehicular

Environments

Wireless Access in Vehicular Environments

(WAVE) is the operating mode used by IEEE 802.11

devices to work on Dedicated Short Range

Communications (DSRC) bands. DSRC is the name

of the 5.9 GHz band allocated for communication in

the Intelligent Transportation System (ITS) (Weigle,

2008).

WAVE architecture can be seen in Figure 1.

WAVE consists of 5 layers and refers to the OSI

(Open Systems Interconnection) model that layer

occupies layers 1-4 and a combination of layers 5, 6,

and 7. At the top layer, the application layer, IPv6

supports access internet wirelessly via TCP

(Transmission Control Protocol). IP-based

communication allows vehicle users to access the

internet from vehicles. Communication services that

are not IP based are supported by the WAVE Short

Message Protocol (WSMP). This communication

service is generally used to develop active safety or

ITS system (Baccelli, et al., 2010).

The WAVE architecture supports multi-channel

operations that support control channels (CCH) and

service channels (SCH) to transmit data in general

and data in the form of WAVE Short Messages

(WSMs). WSMs are transmitted to CCH, especially

if they contain information related to safety, both are

event-driven messages and beacon-driven messages.

Even-driven messages will warn drivers of hazardous

situations, such as accidents and generally have high

priority. While beacon-driven messages will send

traffic information in general and periodically, such

as the location of other vehicles (Ghandour, et al.,

2014). Even-driven messages are distributed

unicastically, while beacon-driven messages are

distributed broadly. Illustration the broadcast mode

can be seen in Figure 2.

Data contained in broadcast messages can be seen

in Table 1. This message is sent every 20 ms in 50 Hz

data rate (Liu, 2012). The unicast mode is used for

special purposes, namely for the delivery of messages

related to safety. One example of the use of this mode

for active safety device is sending safety messages in

the Cooperative Overtaking System (CACC) (Ploeg,

et al., 2011).

Figure 1: WAVE Architecture.

Figure 2: Sending messages in broadcast mode on WAVE.

Table 1: Format and Data Source.

Field

Format

Data source

Time

hhmmss.sss

GPS (UTC)

VID

00:00:00:00:00:00

MAC OBU

Position

xxx.xxxxxx:yyy.yyyyyy

GPS (DD format)

Speed

ss.ss

GPS (m/s)

Heading

hhh.hh

GPS (degree)

Attribute

MAC OBU

WAVE specifications are as follows (Armstrong,

n.d.):

• special frequency range in the 5.9 GHz band (5.860

GHz - 5.920 GHz)

• data rates of 6-27Mbits / s.

EIC 2018 - The 7th Engineering International Conference (EIC), Engineering International Conference on Education, Concept and

Application on Green Technology

336

• 7 licensed channels

• Dual stack protocol namely Transmission Control

Protocol (TCP) and WAVE Short Message

Protocol (WSMP)

• support operations in range up to 1000 meters

• Supports high speed of the vehicle (~ 500 km / h

relative speed)

• support communication on the On Board Unit

(OBU) and Road Side Unit (RSU) and between

OBU.

• Command-response & peer to peer

2.2 Intersection in Connected Vehicle

Environment

Illustrations of intersection in connected vehicle

environments can be seen in Figure 3 and Figure 4.

Figure 3 shows how connected vehilce works at

intersections with traffic lights (signalized

intersection). Each vehicle has an On Board Unit

(OBU) and a Road Side Unit (RSU) is installed in a

traffic light system. In this situation there is an

exchange of data between RSU and OBU also

between OBU and other OBUs. At the intersection

without traffic lights (unsignalized intersection) there

is a data exchange between the OBU and OBUs in

other vehicles. See Figure 4.

2.3 Cooperative Intersection Collision

Avoidance System

CICAS is the development of ICAS. ICAS utilizes a

proximity sensor so that it has weaknesses in the

working range. Communication technology between

vehicles enables the development of CICAS.

CICAS devices function to prevent collisions at

intersection, especially at intersections without traffic

lights. Between vehicles that have been equipped

with OBU communication devices will exchange

information such as speed, position, and direction

periodically every 20 ms. This data exchange allows

each vehicle to decide whether it is safe to cross the

intersection or not. If one vehicle gets priority

crossing then another vehicle will cooperatively

maintain speed or reduce speed so that the vehicle that

gets priority will be safe crossing the intersection.

In Figure 5 two vehicles with the same relative

speed will cross the intersection in different

directions. Vehicle A drove from West to East

(W2E), while vehicle B was from South to North

(S2N). The distance of vehicles A and B to the

intersection is relatively the same, d

= d

. Similarly,

the speed of the two vehicles is also relatively the

same, namely v

= v

Figure 3: Signalized intersection in connected vehicle

environment.

Figure 4: Unsignalized intersection in connected vehicle

environment.

Figure 5: An illustration of the studied intersection.

Multi Agent Protocol for Cooperative Intersection Collision Avoidance System

337

2.4 Protocol Design

The function of the cooperative algorithm on a

cooperative vehicle system is to regulate information

exchange and cooperation mechanisms. Algorithms

that function to regulate information exchange are

known as application protocols. The developed

protocol refers to protocol which is used in the

cooperative multi agent system (CMAS)

environment. The language commonly used in multi

agent systems is agent communication language

(ACL). Almost all ACLs are developed from the

speech act theory, such as The Knowledge Query and

Manipulation Language (KQML) and The

Foundation for Intelligent Physical Agents (FIPA)

ACL (Vasudevan, 1998). We have developed special

ACL used in vehicle agents such as: maneuver =

crossing, sequence = behind, and cooperative =

slower. The design of the protocol for cooperative

intersection crossing can be seen in Figure 6.

Figure 6: Protocol for cooperative overtaking.

An example of the use of ACL in the protocol that

we have designed can be seen in Table 2. In proposed

protocol the information exchange process devide

into three phases, namely the Intersection Passing

Request (IPR), Intersection Passing Action (IPA),

and Intersection Passing Confirmation (IPC).

When two vehicles A and B approach the

intersection the two vehicles periodically exchange

information on speed, acceleration, position and

direction. The distance to the intersection and the

speed of the two vehicles are relatively the same.

According to traffic rules in Indonesia vehicle A will

get priority to cross the intersection compared to

vehicle B. The first step A will send an IP request to

B (IPR). Then B replies the request from A by

sending ‘agree’ message. If ‘agree’ message has been

received by A, then vehicle A will cross the

intersection (IPA). After finishing crossing the

EIC 2018 - The 7th Engineering International Conference (EIC), Engineering International Conference on Education, Concept and

Application on Green Technology

338

intersection vehicle A will send a message to B

containing information that A has finished crossing

the intersection (IPC).

Table 2: An Example of The Use ACL in Protocol.

Sender

00:00:00:00:00:01

Receiver

00:00:00:00:00:02

Performative

request

Content

cooperation(intersectionPass)

ACL

(request

:sender (00:00:00:00:00:01)

:receiver (00:00:00:00:00:02)

:content

"cooperation (intersectionPass)"

:language fipa

)

3 RESULTS AND ANALYSIS

The developed protocol was tested using software

developed using a multi agent system approach.

Protocol performance is determined by calculating

the Protocol Total Time (PTT) which means the

entire time is used for data exchange for intersection

crossing and is calculated by Eq. (1).

PTT = Nbc.Cbc + Nuc.Cuc (1)

where,

Nbc : number of broadcast messages

Cbc : average message delivery time in broadcast

communication

Nuc : number of unicast messages

Cuc : average message delivery time in unicast

communication

Safety Index (SI) represents the percentage of

simulation episodes that result in the success of the

vehicle reaching the target without collision with

other vehicles. The safety index is expressed in Eq.

(2) (Yen & Pfluger, 1995), (Cang, 1999), and (Ngai

& Yung, 2011).

kmc /

(2)

with m is the total simulation episode without

collision and k is the number of all simulation

episodes.

We define the Length of Intersection Passing

Time (LIPT) is the time needed by the vehicle to cross

the intersection from a distance of 100 meters to the

intersection. If the vehicle speed is considered to be

fixed at v m / s, LIPT is obtained as shown in Eq. (3)

LIPT = 100 / v (3)

In the simulation program, LIPT is calculated

from the time taken by the vehicle to cross the

intersection from a distance of 100 m. This is to

approach a more realistic situation because there is no

guarantee that the speed of the vehicle will remain.

Protocol testing is done by observing PTT, SI, and

LIPT for cooperative and non-cooperative situations.

In non-cooperative situations it is assumed that the

vehicle does not have ICAS devices and the algorithm

works on autonomous vehicles.

The test results of the developed protocol can be

seen in Table 2. It can be seen that in cooperative

mode PTT is required which is greater than in non

cooperative mode. This is because during cooperative

mode there is unicast exchange of information

between vehicle A and B. In cooperative mode there

is a significant increase in the safety index. On the

contrary LIPT dropped. More shorter LIPT amounts

show smoother traffic and greater fuel savings can be

obtained.

In Figure 7 can be seen a simulator view is created

specifically to test the CICAS protocol.

Table 3: Simulation results.

Mode

PTT (s)

SI (%)

LIPT (s)

Non Coop

82,2

5,2

Cooperative

99,4

4,5

Figure 7: Display of the CICAS protocol simulator.

4 CONCLUSIONS

The cooperative properties used in ICAS can increase

the safety index and shorten the passage time of

Multi Agent Protocol for Cooperative Intersection Collision Avoidance System

339

crossing. Both parameters indicate that the use of

communication between vehicles will improve the

level of safety and smooth traffic.

REFERENCES

Armstrong, L., n.d. Dedicated Short Range

Communications at 5.9 GHz, s.l.: s.n.

Baccelli, E., Clausen, T. & Wakikawa, R., 2010. IPv6

Operation for WAVE - Wireless Access in

Vehicular Environments. Jersey City, NJ, s.n.,

pp. 160-165.

Basjaruddin, N. C., Kuspriyanto, Priyana, Y. &

Husni, E., 2015. Modeling and simulation of

overtaking maneuver using BDI agent.

International Review on Modelling and

Simulations, Vol 8(4), pp. 458-465.

Basjaruddin, N. C. et al., 2018. Hardware Simulation

of Advanced Driver Assistance Systems Based

on Fuzzy Logic. International Review on

Modelling and Simulations (I.RE.MO.S.), Vol

11(1), pp. 24-31.

Basma, F., Tachwali, Y. & Refai, H. H., 2011.

Intersection Collision Avoidance System Using

Infrastructure Communication. Washington,

DC, USA, 14th International IEEE Conference

on Intelligent Transportation Systems.

Cang, Y., 1999. Behavior-based fuzzy navigation of

mobile vehicle in unknown and dynamically

changing environment. Pokfulam: The

University of Hong Kong (Pokfulam, Hong

Kong).

Elleuch, I., Makni, A. & Bouaziz, R., 2017.

Cooperative Intersection Collision Avoidance

Persistent System Based on V2V

Communication and Real-Time Databases.

Hammamet , 14th International Conference on

Computer Systems and Applications.

Ghandour, A., Di Felice, M., Artaila, H. & Bononib,

L., 2014. Dissemination of Safety Messages in

IEEE 802.11p/WAVE Vehicular Network:

Analytical Study and Protocol Enhancements.

Pervasive and Mobile Computing, Volume 11,

pp. 3-18.

Liu, C., 2012. Design of Efficient In-Network Data

Processing and Dissemination for VANETs,

Houghton, Michigan: s.n.

Ngai, D. & Yung, N., 2011. A Multiple-Goal

Reinforcement Learning Method for Complex

Vehicle Overtaking Maneuvers. IEEE

Transactions on Intelligent Transportation

Systems,Vol 12(2).

Ploeg, J., Serrarens, A. F. & J.Heijenk, G., 2011.

Connect & Drive: design and evaluation of

cooperative adaptive cruise control for

congestion reduction. Journal of Modern

Transportation, September.19(3).

Vasudevan, V., 1998. Comparing Agent

Communication Languages. Object Services

and Consulting, Inc.

Weigle, M., 2008. Standards: WAVE / DSRC /

802.11p, s.l.: s.n.

Yang, Z., Huang, H., Yao, D. & Zhang, Y., 2016.

Cooperative Driving Model for Non-Signalized

Intersections Based on Reduplicate Dynamic

Game. Rio de Janeiro, Brazil, 19th International

Conference on Intelligent Transportation

Systems (ITSC).

Yen, J. & Pfluger, N., 1995. A Fuzzy Logic Based

Extension to Payton and Rosenblatt's Command

Fusion Method for Mobile Robot Navigation.

IEEE Transactions on Systems, Man, and

Cybernetics, Vol 25, pp. 971-978.

EIC 2018 - The 7th Engineering International Conference (EIC), Engineering International Conference on Education, Concept and

Application on Green Technology

340