ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER

AT91SAM7S256 FOR PLATOONING MULTI-AGENT SYSTEM

IMPLEMENTATION

José M. Rodríguez, AbdelBaset M. H. Awawdeh, Felipe Espinosa, Julio Pastor, Fernando Valdés

Department of Electronics, University of Alcalá, 28805 Alcalá de Henares-Madrid, Spain.

Miguel A. Ruiz, Antonio Gil

High Tecnology and Homologation Center, University of Alcalá, 28805 Alcalá de Henares-Madrid, Spain.

Keywords: Platooning using multi-agent systems, electronic implementation of MAS, AT91SAM7S256 guidance

application, nRF2401A communication module.

Abstract: In this work a low cost electronic solution adapted for control and communication of a convoy of electrical

vehicle prototypes based on multi-agent system -MAS- is presented. From the obtained results in previous

works, focused on mobile platforms with PC architecture and Bluetooth communication module, a new

electronic system has been designed based on the 32 bits microcontroller AT91SAM7S256 and the

communication module nRF2401A. With which, it obtains a greater integration of the final mobile

prototype and a greater communication capability between the devices connected in wireless network.

1 INTRODUCTION

Road transport accidents and contamination as well

as traffic congestion have been some of the most

important problems to solve in the last years

(Ioannou, R, 2002; Zuylen, H. J., et al., 2000). Many

of the proposed solutions based on cooperative-

driving, especially vehicles platooning (Yamamura,

Y., et al., 2002), which is defined as a group of

vehicles whose actions in the road are coordinated

by means of communication to reduce platoon

oscillations and eliminate the so called “slinky

effect” or “string-instability” which refers to the

amplification of the peak of the error from vehicle to

vehicle up stream the last one.

This justifies the dedicated effort to realize

studies and projects in the platoon guidance field.

Such as: NAVLAB of Carnegie University (Gowdy,

J., et al., 1997), PATH related to University of

California (PATH, 1998; PATH, 2003),

PROMETHEUS handles a part of the “Eureka”

European program (Clarke, N., 1995), ARGO of

Parma and Pavia Universities (Broggi, A., et al.,

1999), PRAXITELE proposed by INRIA (Massot,

M.H., et al., 1999; Laugier, C., et al., 1999),

CHAUFFEUR (Chauffeur, 2006), CARTALK 2000

which brings up by the cooperation among

automotive electronics companies (Bosch, Siemens,

etc.) and Colonia and Stuttgart universities

(Reichardt, D., et al., 2002; Maihöfer, C., et al.,

2004), MIMICS carried up by university of Murcia

and Valencia Polytechnic School (Martínez-Barberá,

H., et al., 2003).

In 2002 the European Commission lunched an

initiative called CIVITAS (Civitas, 2002) to improve

the urban transport by means of innovators methods,

strategies, technologies and infrastructures. From

these is possible to emphasize:

a) Platoon of vehicles public or private, clean

and profitable as far as costs and energy, and the

necessary infrastructure.

b) Strategies of demand management based on

the access restrictions to the centric zones and other

sensible zones.

c) Innovators strategies for management of

mobility demand.

d) Integration of the transport management

systems and the associated information services.

In the Department of Electronics in University of

Alcala various works, in platoon guidance, have

202

M. Rodríguez J., M. H. Awawdeh A., Espinosa F., Pastor J., Valdés F., A. Ruiz M. and Gil A. (2006).

ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER AT91SAM7S256 FOR PLATOONING MULTI-AGENT SYSTEM IMPLEMENTATION.

In Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics, pages 202-209

DOI: 10.5220/0001216202020209

 SciTePress

been realized in different aspects: on-line / off-line

trajectory generation and representation as a set-of-

points considering linear as well as non-linear

trajectory (Awawdeh, A.M.H., et al., 2004-a), multi-

agents system applied to vehicles platooning

(Awawdeh, A.M.H., et al., 2002; Awawdeh,

A.M.H., et al., 2004-b), lateral and longitudinal

control to ensure string-stability and to reduce lateral

and longitudinal oscillations (Awawdeh, A.M.H., et

al., 2004-c; Awawdeh, A.M.H., et al., 2004-d).

These works have been developed as a part of

COVE project, which deals with platoon

implementation (formation and driving) using multi-

agent system in transport contexts characterized by

great number of potential users where also

transportation units circulate with low speeds (less

than 40km/h). Such as: airports, historical-cultural

centres of millennial cities, thematic parks, business

parks, university’s campus, etc. In these cases, the

traffic ordination based on conventional vehicles is

problematic and even prohibitive. Furthermore;

because of the curvature variation in trajectory (soft

and hard variation), the tracking becomes more

important than vehicle’s velocity. For that; one or

more platoons formed by electrical vehicles have

been proposed. In those the first unit (the lead one)

drives the platoon to reach the asked station (stop

points), which has been solicited by user by means

of a cell-phone. As seen in figure 1.

Figure 1: Vehicles platooning attending to the user petition

of transport service.

It was shown that for identical vehicles, it is

impossible to achieve string stability for spacing

distance error using decentralized identical

controllers (Maziar, E., et al., 2004; Chaibet, A., et

al., 2005). For that, new control strategies and

architectures have been designed to communicate

Platoon’s members together, in order to get

information about the intentions of the lead vehicle

and the preceding one. Two of the most important

architectures that have been implemented in the

platoon guidance field are: layers architecture

(Varaiya, et al., 1994; Varaiya and Horowitz, 2000)

and agents’ architecture (Matsui, et al., 2000; Hallé,

et al., 2003). In layer architecture, the hierarchical

structure controls the vehicles, but in second

architecture each vehicle represented by one agent is

able to communicate with the others to achieve the

cooperative driving and the vehicle platooning.

The object of the mentioned works and

researches is vehicles platooning (guidance and

formation) in linear and cuasi-linear trajectory

(highways), where the lateral velocity is

approximately zero compared with the longitudinal

one. Nevertheless, the COVE project focuses on

transport scenarios where the non-linear trajectory

and the route selection are more important than the

speed. For that, new architecture has been carried

out including three agents for the followers and two

for the leader.

The rest of this paper has been structured in form

that: the second section presents the multi-agent

architecture developed for guidance and formation

platooning. The third part details the agent structure

and simulation tool designed for its evaluation.

Furthermore it details the interaction between

agents. The fourth part presents the novel low cost

electronic solution proposed to implement the MAS

based on the 32 bit microcontroller:

AT91SAM7S256. Summing up, part fifth is

dedicated to conclusions.

2 MULTI-AGENT SYSTEM

DESCRIPTION

The platoon prototype under test is formed by full

automated followers and a leader remotely

controlled. This one deals with the trajectory

generation for the rest of platoon members, and the

linking-maneuvers organization such as splitting;

merging and following.

Figure 2 presents the control hierarchy designed

to realize the basic tasks of platoon guidance

(Awawdeh A.M.H et al, 2004-b). The leader control

logic must solve the trajectory generation based on

its actual location and the anticipated curvature, and

broadcasts this information to the rest of platoon

member, the decision making takes into account the

norms and the traffic conditions in addition to the

number of units in the platoon and, finally, the

possible communication with other leaders of other

convoys to guarantee safe displacements and the

traffic flow balance.

The followers control logic must give solutions

to the exchange of information relative to the state of

movement of the leader and other followers (the first

two followers and the preceding unit). Furthermore,

it has to govern the system of traction and direction

ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER AT91SAM7S256 FOR PLATOONING

MULTI-AGENT SYSTEM IMPLEMENTATION

203

in order to reach individual stability as well as the

string one, without forgetting the platoon units

manoeuvres (splitting, merging and following). In

addition, the leader as well as the followers requires

information related to the mechanical characteristics

of each unit, supposing that its size can condition the

route and the convoy following, especially in

nonlinear trajectories. This kind of information has

been called mechanical performance as seen in

figure 2.

3 AGENT STRUCTURE

The approach used here for the agent design,

suggests that its intelligent behavior can be

generated by giving it the minimum symbolic

representation of its environment, in which these

symbolic representations are first order logical

formula or/and mathematical formula in the

trajectory generation case; lateral control execution

level and longitudinal one. In addition, the decision

making is implemented in some form of direct

mapping from situation to action. That is to say, the

decision function action is realized through a set of

behaviors, together with an inhibition relation

holding between these behaviors.

In this way, the agent’s behavior is determined

by the agent’s deduction rules (its “program”) and

the environment mathematical modeling (Awawdeh

A.M.H et al, 2004-b). For that, agent architecture

has been developed using a composition of three

ones: logic based; reactive and layered, as seen in

figure 3.

Cooperation layer (agent, layers) - deals with the

social interactions, i.e. , the rules that control and

“draw” the layer-to-layer, agent-to-agent

interactions, intra-platoon cooperation in the

follower’s circulation agent and inter-platoon

cooperation in the leader’s circulation agent.

Figure 3: Hierarchical architecture for trajectory;

navigation; and circulation agents designed for COVE

project.

Figure 2: Hierarchical control of semi-autonomous platoon of vehicles using multi-agent system technology.

ICINCO 2006 - ROBOTICS AND AUTOMATION

204

Knowledge layer (agent, layers) - deals with the

representation of the agent and its environment at

different levels of abstraction. In this way the

knowledge base is divided to three levels: highest

level represents the plans and actions of other agents

in the environment; the middle-level represents the

plans and actions of the agent itself; and the lowest

level represents “raw” information about the vehicle

and its environment.

Reactive layer - provides a possible immediate

response to the changes that occur in the

environment. It is implemented as a set of situation-

action rules.

World model- represents the relation between

perceptions and actions. In trajectory agent case,

lateral and longitudinal control which are presented

as a mathematical model; in navigation agent case it

presents the mathematical model of trajectory

generation adapted to the curvature. Finally, in the

circulation agent it presents the logical model of the

traffic rules, inter-platoon, intra-platoon and linking

rule.

Planning layer (agent, layers) – in the leader’s

and follower’s agents, it is responsible for

determining the agent goals and for the layers tasks

allocation. In addition, in a leader’s circulation

agent, it is responsible of the platoon formation plan

with other vehicles using the linking rules and of the

platoon performance information. Then, according

to this information, it plans split or merge tasks.

See-to-perception- represents the output of the

see function which is implemented in hardware and

software to obtain information about the vehicle and

agents environment. This perception is the

transformation of low-level raw sensor data into

higher-level (processed data) in order to make them

suitable for further processing by the navigation and

trajectory functions.

Commands-to-action – uses hardware and

software actuators (traction, steering, notification

signals, etc) to transform the agent commands to

percept action in its environment.

Finally, this architecture has been used to design

the leader agents and the follower agents. The

differences here among agents are the tasks of the

agent's layers, as indicated before.

To study and evaluate the behavior of the

proposed multi-agent system architecture, a

simulator with graphical-user-interface has been

designed to demonstrate a platoon of vehicles

guidance using MAS (Awawdeh, A.M.H., et al.,

2004-b). In addition, each agent’s behaviors and

actions have been designed as an individual program

that can be enabled/disable or active/inactive. These

agents are active depending on the operation mode

(platoon or free vehicle) and in the platoon case on

the members’ driving mode (full-automatic for

followers; and manual for platoon’s leader).

Figure 4 presents the block-diagram of

communication among agents, data base and

designed GUI. The communication is carrying up

through the global data base. Imagine that an agent

(i.e. trajectory agent) needs information from

another one (i.e. Navigation agent), the first send a

request message to navigation agent, the last one

claims the access to the database, once it is accepted

the navigation agent emits the data to the database

and sends an acknowledgment to the trajectory

agent. The access control program allows the

trajectory agent to copy this data once this agent has

the priority.

Figure 4: In-vehicle agents interactions and communications.

ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER AT91SAM7S256 FOR PLATOONING

MULTI-AGENT SYSTEM IMPLEMENTATION

205

The algorithms and strategies designed by COVE

project group have been validated by means of

simulation tools (QT, Matlab, Simulink, and

Toolboxs), emulation, and laboratory tests. Using

the mentioned MAS architecture the computational

time was reduced to 30% of the time required

without this method.

As seen in figure 4, each agent has two parts:

hardware which represents the associated sensors

and actuators, and software part deals with the agent

actions and behaviours.

4 NEW PROPOSAL OF

HARDWARE

IMPLEMENTATION

Communication and control represent the mainly

important tasks to address in the hardware

implementation of vehicles prototypes for

cooperative driving, especially for platoon guidance.

The first experimental test was carried out using ad-

hoc mobile platforms (i.e. ad-hoc electrical vehicles

prototype designed in the Electronics Department of

Alcala University, as shown in figure 5) (Awawdeh,

A.M.H., et al., 2005). The low and high level control

tasks are executed in a digital system based on PC.

For the communication among transport units a

commercial Bluetooth module (Module class 1,

Conceptonic Bluetooth USB Adapte, CBT 100U) is

used.

Figure 6 shows the first hardware solution

implemented using personal computer architecture

(Pentium II), with external CPLD cards as interface

with power electronic cards. On the other hand, the

Bluetooth communication technology presents

various limitations, such as: reach distance 100m,

the maximum number of network’s nodes is 8 (1

master and 7 slaves). On the other hand, in the

‘piconet’ communication mode the master is the

necessary link among the followers communication,

etc.

These handicaps have been overcome with a

new proposal of hardware implementation based on

a 32 bit microcontroller - μC - (AT91SAM7S256)

and a transceiver (nRF2401A). This low cost

electronic solution achieves a high integration level

and improves the facilities of the mentioned

prototype version.

The Atmel AT91SAM7S256 (Atmel) is an

ARM7TDMI based High-performance 32-bit RISC

μC that includes: USB device interface, real-time

clock and clock generator, 4-channel PWM, 2 wire

interface, 8-channel 10-bit ADC, 256K bytes Flash,

64K bytes SRAM, high-density 16-bit instruction

set, IEEE 1149.1 JTAG boundary scan on all digital

pins. This μC supports a flexible Real-Time

Operating System (ARTX-ARM) including:

multitasking (which allows to manage several jobs

or tasks on a single CPU), Real-Time Control

(which allows to configure tasks so that operations

execute within a defined period of time), and Inter-

Task Communication (which allows to configure

communication between various tasks in the

system).

Figure 6: Block-diagram of the first hardware architecture

of COVE

platform.

Figure 5: First COVE Platform, based on PC hardware

architecture, used as prototype for platoon formation.

ICINCO 2006 - ROBOTICS AND AUTOMATION

206

Another important device in this new electronic

solution for each unit platform of the convoy is the

nRF2401A, a RF transceiver highly integrated for

wireless communication. The entire transceiver

including all inductors and filters is integrated in a

single chip; the only external components needed to

make a complete RF system are a crystal and a

resistor. All configuration of the nRF2401A (Spark)

transceiver is done via a standard serial interface. It

can receive on 2 channels simultaneously. This

module allows to increase the communication rang

to 2km at least.

Figure 7 shows the most important tasks

developed by the μC concerning to the mobile

control.

Thanks to the size and weight reduction of the

hardware architecture a new follower version is used

in COVE project (see Figure 8 and compare with

Figure 5).

The used vehicle prototype corresponds to

Ackerman structure type. For that, in the controllers

design the simplified model (bicycle model) has

been used (Matsui, et al., 2000; Maziar, E., et al.,

2004).

5 CONCLUSIONS

As mentioned before, control and communication

are the most important problems to solve related to

cooperative driving necessary for platooning

applications.

In this work, an improve version of electronics

architecture developed for the COVE project is

presented. The heart of this proposal is a 32 bit

microcontroller (T91SAM7S256), running a

multitasking RTOS. In this way the multi-agent

system designed by the research group is

implemented avoiding the use of PC as in the

previous versions. Besides, a transceiver

(nRF2401A) makes easy the wireless

communication among followers and leader, and

increase the network node-spacing distance.

ACKNOWLEDGEMENTS

The work described in this paper has been possible

by funding from the Spanish Ministry of Education

and Science: Project COVE (reference TRA2005-

05409/AUT).

REFERENCES

Atmel, http://www.atmel.com/

Awawdeh, A.M.H. Espinosa, F. and Núñez, F., 2004-b.

Hybrid multi-agent architecture for vehicles

platooning and cooperative driving. 3ª Conferencia

Iberoamericana en Sistemas, Cibernética e

Informática (CISCI’04, Orlando, Florida, USA.

Awawdeh, A.M.H., and Espinosa, F., 2004-a, A new

strategy of non-linear trajectory generation for a

platoon to reduce inter-vehicle error distance

estimation. First IFAC Symposium on Advances in

Automotive Control (IFAC’04), Salerno, Italy.

Awawdeh, A.M.H., Espinosa, F., and Lázaro, J.L., 2002,

Approach to the application of multi-agent technology

to intelligent transportation system. TELEC’04

Telecommunications, Electronics and Control

Systems. Santiago de Cuba, Cuba.

Awawdeh, A.M.H., Espinosa, F., and Mazo, M., 2004-c,

Non-linear trajectory generation and lateral control

new algorithms to minimize platoon’s oscillations.

American Control Conference (ACC’04). IEEE

Control Systems Society. Boston, Massachusetts, USA.

Awawdeh, A.M.H., Espinosa, F., Tahboub, K.A., and

Ramírez, S., 2004-d, Linear approach for longitudinal

control algorithm applied to a platoon of vehicles in

highways. Telecommunications, Electronics and

Control systems (TELEC’04). 2004. Santiago de

Cuba, Cuba.

Broggi, A., Bertozzi, M., Fascioli, A., and Conte, G.,

1999, Automatic vehicle guidance: the experience of

the ARGO autonomous vehicle. World Scientific,

Singapore.

Chaibet, A., Nouveliere, L., Mammar, S., Netto, M., and

Labayrade, R., 2005, Backstepping control synthesis

for both longitudinal and lateral automated vehicle,

IEEE.

Chauffeur, 2006, Main web page of the Chauffeur II

project: http://www.chauffeur2.net

Civitas, 2002, www.civitas-initiativa.org.

Clarke, N., 1995, Anti-collision autonomous support and

safety intervention system / research into optimally

assisted driving systems and technology developed for

European roads.

http://www.ethoseurope.org/ethos/tap.nsf

Gowdy, J., Hebert, H., Thorpe, C., and Stentz, A., 1997,

Intelligent unmanned ground vehicles. Autonomous

navigation research at Carnegie Mellon. Kluwer

Academic Publishers.

Hallé, S., Chaib-draa, B., and Laumonier, J., 2003. Car

platoons simulated as a multi-agent system,

Montpellier, France, pp 57-63, Publications of the

DAMAS Lab.

Ioannou, P., 2002. Evaluation of the effects of intelligent

cruise control vehicles in mixed traffic. University of

Southern California, Final report for task order 4217

(former MOU392). U.S.A.

Laugier, C., Paromtchik, I., and Parent, M., 1999,

Developing autonomous maneuvering capabilities for

future cars. Researchers from INRIA (France).

ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER AT91SAM7S256 FOR PLATOONING

MULTI-AGENT SYSTEM IMPLEMENTATION

207

Maihöfer, C., Coletti, L., 2004, CarTALK 2000: Results

of communication technology and applications for

automotive. Invited presentation at IEEE VTC 2004-

Spring, Milan, Italy.

Martínez-Barberá, H., Zamora-Izquierdo, M.A., Toledo-

Moreo, R., Ubeda., B., and Gómez-Skarmeta, 2003,

A., The mimics project: an application for intelligent

transport systems. 0-7803-7848-2/03 IEEE.

Massot, M.H., 1999, Allouche, J.F., and BÈnÈjam, E.

PRAXITELE: Preliminary results from the Saint-

Quentin experiment.

Matsui, T., Tsugawa, S., Kato, S., and Naganawa, H.,

2000. An architecture for cooperative driving of

automated vehicles. In Procs. IEEE Intelligent

Vehicles Symposium, pages 422-427, Dearbsorn, MI,

USA.

Maziar E. Khatir, and Edward J. Davison, 2004, Bounded

stability and eventual string stability of a large platoon

of vehicles using non-identical controllers, 43rd IEEE

Conference on Decision and Control, , 2004, Atlantis,

Paradise Island, Bahamas.

PATH, 1998, Annual Report. (www.path.berkeley.edu).

PATH, 2003, home page, htttp://www.path.Berkeley.edu.

Reichardt, D., Miglietta, M., Moretti, L., Morsink, P., and

Schulz, W., 2002, CarTALK 2000 - safe and

comfortable driving based upon inter-vehicle-

communication. IEEE intelligent vehicle symposium,

Versailles.

Spark,http://www.sparkfun.com/commerce/product_info.p

hp?products_id=151

Varaiya, P., and Horowitz, R., 2000. Control design of an

automated highway system. In Proceedings of the

IEEE, volume 88, pages 913-925.

Varaiya, P., Hedrick, J.K., and Tomizuka, M., 1994.

Control issues in automated highway systems. IEEE

Control System Magazine, 14(2): 21-32, 1994.

Yamamura, Y., Seto, Y., Hikaru, and Kawabe, T., 2002,

An ACC design method for achieving both string

stability and ride comfort, AVEC’02, Japan.

Zuylen, H. J., Weber, M., Shires, J., and Eriksson, A.,

2002, Options to support the introduction of new

technologies and their implications on transport

policy. FATASIE deliverable 23, report to the

European Commission.

Figure 7: Scheme with the main signals involved in car-like mobile prototype control based on the AT91SAM7S256

micro-controller.

ICINCO 2006 - ROBOTICS AND AUTOMATION

208

Figure 8: Two sides of the new mobile prototype fo

platooning guidance used in COVE project.

ELECTRONIC SOLUTION BASED ON MICRO-CONTROLLER AT91SAM7S256 FOR PLATOONING

MULTI-AGENT SYSTEM IMPLEMENTATION

209