Automatic Vehicle
`Guidance:
`The Experience of the ARGO
`Autonomous Vehicle
`
`Alberto Broggi
`Massimo Bertozzi
`Alessandra Fascioh
`Gianni Conte
`
`World Scientific
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Automatic Vehicle
`Guidance:
`The Experience of the ARGO
`Autonomous Vehicle
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This page is intentionally left blank
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Automatic Vehicle
`Guidance:
`The Experience of the ARGO
`Autonomous Vehicle
`
`Alberto Broggi
`Massimo Bertozzi
`Alessandra Fascioli
`Gianni Conte
`Universtty of Parma,
`Italy
`
`World Scientific
`Singapore • New Jersey 'London • Hong Kong
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Published by
`World Scientific Publishing Co. Pte. Ltd.
`P O Box 128, Farrer Road, Singapore 912805
`USA office: Suite IB, 1060 Main Street, River Edge, NJ 07661
`UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
`
`British Library Cataloguing-in-Publication Data
`A catalogue record for this book is available from the British Library.
`
`AUTOMATIC VEHICLE GUIDANCE:
`The Experience of the ARGO Autonomous Vehicle
`Copyright © 1999 by World Scientific Publishing Co. Pte. Ltd.
`All rights reserved. This book, or parts thereof may not be reproduced in any form or by any means,
`electronic or mechanical, including photocopying, recording or any information storage and retrieval
`system now known or to be invented, without written permission from the Publisher.
`
`For photocopying of material in this volume, please pay a copying fee through the Copyright
`Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
`photocopy is not required from the publisher.
`
`ISBN 981-02-3720-0
`
`This book is printed on acid-free paper.
`
`Printed in Singapore by Uto-Print
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`"... and at night, by the skill of Argo,
`they reached broad-Bowing Phasis,
`and the utmost bourne of the sea."
`
`Apollonius, THE ARGONAUTICA, Book II
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This page is intentionally left blank
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Preface
`
`During the last decade, the subject of Intelligent Transportation gained
`strategic importance and widespread relevance. Many projects were
`launched worldwide aimed at analyzing the problem of people's mobility
`and goods transportation from a number of different perspectives; and it is
`in the last few years that the first prototypes of both vehicles equipped with
`automatic driving facilities and road infrastructures supporting these func­
`tionalities, are being tested and demonstrated to the public. This book
`surveys the history of intelligent vehicles, discusses some of the different
`approaches developed worldwide by a large number of research institu­
`tions, and presents the solutions adopted by the Universita di Parma in
`the ARGO project, which started about 10 years ago within the Eureka
`PROMETHEUS project. In particular, this book illustrates the problem,
`proposes some of the different solutions, and details the design, the de­
`velopment, and the engineering of a hardware and software platform for
`Automatic Vehicle Guidance, as well as the set-up of two prototype vehi­
`cles.
`Among the main results of this research, the GOLD (Generic Obstacle
`and Lane Detection) system is presented; it is an automatic driving system
`which has been integrated on ARGO -a Lancia Thema 2000 passengers'
`car- allowing to drive the vehicle autonomously in real traffic conditions
`along highways and freeways, with no requirements of additional specific
`road infrastructures.
`The results of this long term research was demonstrated to the interna­
`tional scientific community and to the public in the first week of June 1998
`with a journey through Italy, the MiUeMiglia in Automatico tour, during
`
`vii
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`viii
`
`Preface
`
`which the vehicle drove autonomously for about 2000 km. The experience
`of this demonstration is discussed in the book, along with a description of
`the main advantages and problems encountered.
`This book is divided into three parts. The first part presents the motiva­
`tion of this research and a brief history of the main projects launched world­
`wide aimed at vision-based vehicle driving. The second and third parts are
`related to the ARGO project. Part II describes both the algorithms and
`hardware platforms developed during the whole project -starting from the
`very first implementation, up to the current one- and presents the equip­
`ment installed on the ARGO prototype vehicle. Part III shows a report of
`the extensive test that was performed on ARGO and analyzes the problems
`encountered and the overall system performance.
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Acknowledgment
`
`During the last ten years of research a large number of people helped
`and supported this project. Among them we would like to acknowledge:
`Francesco Gregoretti, Leonardo Reyneri, Claudio Sansoe, and Roberto
`Passerone from the Politecnico di Torino, Italy, for the joint development of
`the PAPRICA hardware system; Giorgio Quaglia and Sandra Denasi from
`the Istituto Elettrotecnico Nazionale Galileo Ferraris, Torino, Italy, and
`Giovanni Adorni from our department for the set-up of MOB-LAB; Lucio
`Bianco, Leopoldo Chinaglia, and Agostino Scognamiglio from Progetto Fi-
`nalizzato Trasporti II, CNR, Italy, for supporting the research; Sergio Tem-
`peri and Edmondo Pietranera from Telecom Italia Mobile, which sponsored
`the MilleMiglia in Automatico tour and all our friends who welcomed us as
`guests, hosted our seminars and helped in the organization of this exciting
`journey; Mrs. Lucia Caccioli for her kind help in improving the style of the
`presentation; and finally our administrative and technical staff whose con­
`tribution was essential for the complete development of the whole project.
`We would also like to acknowledge the support we received from Giorgio
`Quaglia, Giulio Vivo, Uwe Franke, Wilfried Enkelmann, Pierre Charbon-
`nier, Dean Pomerleau, Umit Ozguner, Keith Redmill, Ichiro Masaki, and
`Nicole Love, who provided up-to-date material regarding the state of the
`art of their projects.
`
`ix
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This page is intentionally left blank
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Contents
`
`Preface
`
`Acknowledgment
`
`PART I INTELLIGENT VEHICLES
`
`1
`
`2
`
`Introduction
`
`Intelligent Vehicles and Machine Vision
`2.1 Evolution of Intelligent Transportation Systems
`2.2 Requirements of Intelligent Transportation Systems
`2.3 Sensing the Environment
`2.4 Machine Vision
`
`3 State of the Art
`3.1 Road Following
`3.1.1 Lane Detection
`3.1.2 Obstacle Detection
`3.2 Worldwide Projects
`3.2.1 Research carried out on the MOB-LAB Vehicle
`3.2.2 Research carried out at the Centro Ricerche FIAT
`. ..
`3.2.3 Research carried out at the Universitdt der Bundeswehr
`3.2.4 Research carried out at Daimler-Benz
`3.2.5 Research carried out at the Fraunhofer-Institut fiir In­
`formations und Datenverarbeitung
`
`xi
`
`vii
`
`ix
`
`1
`
`5
`
`9
`9
`12
`14
`15
`
`21
`21
`21
`23
`25
`25
`29
`34
`37
`
`38
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`xii
`
`Contents
`
`3.2.6 Research carried out at the Laboratoire Central des Ponts-
`et-Chaussees de Strasbourg
`3.2.7 Research carried out at the Defence Evaluation & Re­
`search Agency
`. .
`3.2.8 Research carried out at Carnegie Mellon University
`3.2.9 Research carried out at The Ohio State University
`. ..
`3.2.10 Research carried out at the University of Michigan . ..
`3.2.11 Research carried out at the Massachusetts Institute of
`Technology
`3.2.12 Research carried out at the Phoang University of Science
`and Technology
`
`PART II THE ARGO PROJECT
`
`4 Algorithms for Image Processing
`4.1 Lane Detection: a Model-Based Approach
`4.1.1 The multi-resolution approach
`4.1.2 The algorithm structure
`4.1.3 Performance analysis
`4.1.4 Critical analysis and evolution
`4.2 Obstacle Detection: a Model-Based Approach
`4.2.1 The vehicle detection algorithm
`4.2.2 Performance analysis
`4.3 The GOLD System
`4.3.1
`Inverse Perspective Mapping
`4.3.2
`Inverse Perspective Mapping and stereo vision
`4.3.3 Functionalities
`4.3.4 An extension of Inverse Perspective Mapping to handle
`non-flat roads
`4.3.5 Discussion
`
`5 Hardware Support for Real-Time Image Processing
`5.1 The PAPRICA Architecture
`5.1.1 Architectural issues
`5.1.2 Hardware system description
`5.2 Critical Analysis of the PAPRICA Architecture
`5.2.1 Memory organization and processor virtualization
`5.2.2
`I/O problems
`
`. ..
`
`40
`
`42
`44
`47
`49
`
`51
`
`52
`
`55
`
`59
`59
`60
`61
`69
`70
`74
`75
`80
`82
`83
`88
`94
`
`107
`114
`
`123
`123
`124
`125
`128
`128
`129
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Contents
`
`xiii
`
`130
`Instruction set
`5.2.3
`131
`5.2.4 Architectural evolution
`132
`5.3 The PAPRICA-3 Architecture
`133
`5.3.1 Hardware system description
`137
`5.3.2 Obstacle Detection on PAPRICA-3
`142
`5.4 The MMX Technology
`143
`5.4.1 MMX optimization issues
`146
`. . ..
`5.4.2 Obstacle Detection on an MMX-based processor
`5.5 Comparison between PAPRICA-3 and MMX Processors
`. . 152
`5.5.1 Algorithms implementation
`152
`5.5.2 Performance evaluation
`155
`5.5.3 Discussion
`156
`
`6 The ARGO Vehicle
`6.1 Description
`6.1.1 The name
`6.2 The Data Acquisition System
`6.2.1 The vision system
`6.2.2 The speed sensor
`6.2.3 The user interface
`6.2.4 The keyboard
`6.3 The Processing System
`6.4 The Output System
`6.4.1 The acoustical devices
`6.4.2 The optical devices
`6.4.3 The mechanical devices
`6.5 The Control System
`6.5.1 Lane change maneuvers
`6.6 Functionalities
`6.7 Other Vehicle Equipments and Emergency Features
`
`PART III THE PROJECT'S RESULTS
`
`in Automatico Tour
`
`7 The MilleMiglia
`7.1 Description
`7.1.1 Dates and schedule
`7.1.2 Data logging
`7.1.3 Live broadcasting of the event via Internet
`
`161
`161
`162
`163
`163
`165
`165
`166
`166
`167
`168
`168
`169
`171
`171
`173
`174
`
`177
`
`181
`181
`182
`183
`183
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`xiv
`
`Contents
`
`8 Performance Analysis
`8.1 System Performance
`8.1.1 The vision system
`8.1.2 The processing system
`8.1.3 The visual processing
`8.1.4 The control system
`8.1.5 The man-machine interface
`8.1.6 Environmental conditions
`8.1.7 The data transmission system
`8.2 Statistical Analysis of the Tour
`8.2.1 Detailed analysis of one hour of automatic driving
`8.3 Discussion
`
`187
`187
`187
`188
`188
`189
`190
`190
`192
`193
`. . . 194
`197
`
`9 Closing Remarks
`
`APPENDICES
`
`A Morphological Implementation of the DBS Filter
`
`B PAPRICA-3 Programming Environment
`B.l Low Level Programming Language
`B.2 High Level Programming Language
`B.3 Assembly Code Optimization
`B.3.1 Deterministic optimization
`B.3.2 Stochastic optimization
`B.3.3 Parallel implementation of the code optimizer
`
`201
`
`203
`
`205
`
`211
`211
`213
`214
`214
`214
`218
`
`C Global Communications on PAPRICA-3
`C.l Concurrent Communications on the ICN
`C.2 Determining the Sets of Compatible Communications . . ..
`References
`Biographic Notes
`
`219
`221
`225
`227
`241
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`PART I
`INTELLIGENT VEHICLES
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This page is intentionally left blank
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This first part of the book presents the motivations that triggered off
`the high interest towards Intelligent Transportation Systems and describes
`the social and economical impact that automatically driven vehicles will
`cause in the future. Then, after the analysis of the main requirements and
`consequential benefits, a brief history of Intelligent Transportation Systems
`is presented. Finally, it describes the results obtained in this field by various
`research institutions worldwide.
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`This page is intentionally left blank
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Chapter 1
`Introduction
`
`Automatic Vehicle Driving is a generic term used to address a technique
`aimed at automating -entirely or in part- one or more driving tasks. The
`automation of these tasks carries a large number of benefits, such as: a
`higher exploitation of the road network, lower fuel and energy consump­
`tion, and -of course- improved safety conditions compared to the current
`scenario.
`The tasks that automatically driven vehicles are able to perform range
`from the following: the possibility to follow the road and keep within the
`right lane, maintaining a safe distance between vehicles, regulating the vehi­
`cle's speed according to traffic conditions and road characteristics, moving
`across lanes in order to overtake vehicles and avoid obstacles, helping to
`find the correct and shortest route to a destination, and the movement and
`parking within urban environments.
`Besides the obvious advantages linked to increasing safety and reducing
`road accidents -thus saving human lives- the possibility of having vehicles
`riding in a much closer proximity than they do today, produces an increment
`of road capacity, and, together with the intelligent modulation of vehicles'
`speed, also causes an appreciable reduction of fuel consumption. In other
`words, Automatic Vehicle Driving tends to achieve optimal use of current
`infrastructures, improve mobility, and minimize risks, travel times, and
`energy consumption. Moreover, commercial and industrial vehicles which
`repeatedly move along a given path, benefit from a stronger control of their
`routes and require less personnel to manage their moves. Co-operative
`driving and dynamic fleet management can therefore play a fundamental
`role for the reduction of industrial costs.
`
`5
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`ti
`
`Introduction
`
`Two main cooperative solutions are possible to achieve automatic driv­
`ing functionalities: they require to act on
`• infrastructures or
`• vehicles.
`Both scenarios have their own pros and cons, depending on the specific
`application. As a general rule, public transportation and industrial robotics
`applications -based on repetitive and prescheduled routes- can generally
`benefit from a specific and fixed road infrastructure. Instead, for private
`vehicles, moving along extremely large and extended road networks, the
`organization, set-up, and maintenance of nation-wide road infrastructures
`becomes cumbersome and extremely expensive.
`Conversely, the establishment of a structured environment specifically
`designed for this purpose can be possibly considered for a reduced subset
`of the road network, for example for building a fully automated highway
`on which only automatic vehicles -public or private- can drive.
`Moreover, infrastructure-based systems generally require a more exten­
`sive preparation time. For this reason, systems that are expected to be
`achieved on a short-term basis can only be vehicle-autonomous or at the
`most, simple versions of infrastructure-based systems. Such systems do not
`require a dedicated complex infrastructure or fit into already existing traffic
`infrastructure (RDS, GSM, beacons for automatic toll systems and traffic
`data banks).
`Furthermore, Automatic Vehicle Driving hides another extremely im­
`portant benefit, that first triggered the interest in this new field of appli­
`cation. The on-board equipment and technology that allows a vehicle to
`drive autonomously can also be exploited as a safety enhancement in a more
`general meaning. The autopilot can be switched to the more static task
`of monitoring the driver's activities. In case of dangerous situations -such
`as driver's illness or drowsiness-, the equipment understands the risk and
`warns the driver or even takes control of the vehicle, starting an emergency
`maneuver and stopping the vehicle on the emergency lane, or keeping the
`vehicle on the driving lane at a constant speed until the driver resumes
`control. The autopilot, considered in this second function, behaves like
`an active security system able to prevent accidents, like the Anti Breaking
`System, ABS.
`Recently a large emphasis has been given to security systems -both ac­
`tive and passive ones- and almost every automotive company is integrating
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`7
`
`such systems into its latest car models. For this reason, the techniques and
`the technologies supporting Automatic Vehicle Driving are currently being
`studied by various parties worldwide, focusing on different perspectives.
`Automobile manufacturers view automatic driving systems as an interesting
`product; public and government institutions, such as public transportation
`companies, mainly focus on the possibility of increasing safety, reducing
`energy consumption, and optimizing comfort and traffic control. Similarly
`these kind of activities have also been tackled by the academic world which,
`together with military institutions, sees the problem of Automatic Vehicle
`Guidance as a process of technological transfer from avionics, underwater,
`military, or robotics applications in general, and as an engineering problem
`to produce low-cost systems for the private market.
`In the last decade, a large number of research institutes worldwide have
`been involved in national and international projects related to the analysis
`of the Intelligent Transportation Systems problem, and a number of proto­
`types of intelligent vehicles have been designed, implemented, and tested
`on the road.
`The design of these prototypes was preceded by the analysis of solu­
`tions deriving from similar and close fields of research, and exploded with
`a great flourishing of new ideas, innovative approaches, and novel ad-hoc
`solutions, robotics, artificial intelligence, computer science, computer ar­
`chitecture, telecommunications, control and automation, signal processing
`are just some of the principal research areas from which the main ideas and
`solutions were first derived. Initially, underlying technological devices -such
`as head-up displays, infrared cameras, radars, sonars- derived from expen­
`sive military applications, but, thanks to the increased interest in these
`applications and progress of industrial production, today's technology of­
`fers sensors, processing systems, and output devices at very competitive
`prices. In order to test a wide spectrum of diverse approaches, these proto­
`types of automatic vehicles are equipped with a large number of different
`sensors.
`These vehicles are currently undergoing extensive tests both on struc­
`tured environments, such as closed roads or private tracks, and on pub­
`lic roads in real traffic conditions. The first preliminary results of these
`tests demonstrated that full automation of traffic (at least on motorways
`or sufficiently structured roads) is technically feasible. Nevertheless, not all
`problems related to automatic vehicle driving that still remain to be solved
`are of technical nature: there are some aspects that must be considered
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`8
`
`Introduction
`
`and carefully evaluated in the design of such systems. First of all, prior
`to having an automatic driving system sold and installed on a commercial
`vehicle, all the legal aspects related to the responsibility in case of faults
`and incorrect behavior of the system must be solved. Secondly, in case
`no specific roads are built and dedicated to automatic vehicles only, the
`possibility of driving on a motorway along with automatic vehicles must be
`considered and its impact on human drivers evaluated. Although technical
`aspects seem to have a higher importance, these problems must be dealt
`with and solved as well, since they represent the basics and prerequisites
`on which future automatic highways will rely.
`The research carried out on intelligent vehicles -after its first explo­
`rative stage- has now reached its second stage: many different approaches
`have been evaluated during these years and the most promising ones are
`being engineered, installed on prototype vehicles, and extensively tested
`on the road. However, a long period of exhaustive tests and refinement
`must precede the availability of these systems on the general market, and
`a fully automated highway system with intelligent vehicles driving and ex­
`changing information is not expected for a couple of decades. For the time
`being, complete automation will be restricted to special infrastructures such
`as industrial applications or public transportation. Then, automatic vehi­
`cle technology will be gradually extended to other key transportation areas
`such as the transportation of goods, for example on expensive trucks, where
`the cost of an autopilot is negligible with respect to the cost of the vehi­
`cle itself and the service it provides. Finally, once technology has been
`stabilized and after freezing the most promising solution and the best al­
`gorithms, a massive integration and a widespread use of such systems will
`also take place with private vehicles, but this will not happen for another
`two or more decades.
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Chapter 2
`Intelligent Vehicles
`and Machine Vision
`
`2.1 Evolution of Intelligent Transportation Systems
`
`After completion of the first explorative phase, that started around the
`beginning of the eighties, the field of Intelligent Transportation Systems
`(ITS) is now entering its second phase characterized by a maturity in its
`approaches and by new technological possibilities which allow the devel­
`opment of the first test products. Nevertheless, the evolution towards the
`third stage, i.e. the massive production and integration of ITS technology
`on commercial vehicles, is expected to begin not before a couple of decades.
`The interest in ITS technologies and their related issues was born about
`20 years ago, when the problem of people and goods mobility began to arise:
`the saturation of the most common way to increase mobility, namely the
`extension of the road network, focused the interest of governments and
`research institutions towards new alternative solutions.
`The main target of this first phase was to create a technical substrate to
`be used in the following prototyping stages, namely a formal basis covering
`different subjects. For this reason, automatic vehicle driving, intelligent
`route planning, and other extremely high-level functionalities were selected
`as main goals.
`Government institutions activated the initial explorative phase by means
`of different projects worldwide. The first results were a deep analysis of the
`problem and the development of a feasibility study to understand the re­
`quirements and possible effects of the application of ITS technology. One
`of the main advantages of these activities was the possibility to have a
`large number of different and complementary research units -with differ-
`
`9
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`10
`
`Intelligent Vehicles and Machine Vision
`
`ent backgrounds- working in a cooperative way. This research stage, in
`fact, mainly characterized by non-competitive research, produced a large
`number of different prototypes and possible solutions, all based on rather
`different approaches.
`In this book, only in-vehicle applications will be considered and dis­
`cussed, while road infrastructure, inter-vehicle communication, satellite
`communication, and route planning issues will not be covered.
`
`In Europe the PROMETHEUS project (PROgraM for a European Traf­
`fic with highest Efficiency and Unprecedented Safety) started this explo­
`rative stage in 1986. The project involved more than thirteen vehicle man­
`ufacturers and a large number of research units from governments and
`universities of 19 European countries. Within this framework, a number
`of different approaches regarding ITS were conceived, implemented, and
`demonstrated. The vehicle prototypes developed during this period were
`demonstrated at the end of the project in Paris, in October 1994. The
`most interesting approaches that were proposed included the VaMoRs test
`vehicle developed by the group of Universitat der Bundeswehr [67]. It was
`able to follow the lane thanks to a pair of forward looking cameras and a
`custom parallel image processing engine. Daimler-Benz proposed a system
`called VITA [152] equipped with more than a dozen cameras and a large
`number of Transputer nodes and custom DSPs; it was demonstrated to
`move autonomously on highways at speeds up to 110 km/h.
`In the United States a great deal of initiatives were launched to deal with
`the problems of mobility, involving many universities and research centers.
`Amongst them the Carnegie Mellon University research group, active for
`years in this field, developed the first NavLab vehicle [148] in 1985. After
`this, a large number of different approaches were tested on the series of
`NavLab vehicles, all developed by CMU; these included: ALVINN [80],
`a 30x32 neural net able to learn from a driver and to drive the vehicle
`thanks to the analysis of low-resolution visual patterns, and SCARF [40], a
`system based on color segmentation of pairs of images able to detect even
`unstructured roads in slow varying conditions of light. General Motors also
`presented a system for car following, based on custom hardware.
`At the conclusion of this explorative phase -in 1995- the US estab­
`lished the National Automated Highway System Consortium, including
`many different institutions amongst which is the University of California
`and Carnegie Mellon University, who took part in an impressive demon-
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`Evolution of Intelligent Transportation
`
`Systems
`
`11
`
`stration in San Diego in August 1997 [13].
`In the same way, also in Japan, where the mobility problem is much
`more intense and evident, some vehicle prototypes were developed within
`the framework of different projects. Similarly to what happened in the US,
`in 1996 the Advanced Cruise-Assist Highway System Research Association
`(AHSRA) was established amongst a large number of automobile indus­
`tries and research centers, which demonstrated different approaches to the
`problem of Automatic Vehicle Guidance [150].
`
`One of the aspects shared by all these first prototypes developed within
`the first research phase is the massive use of custom hardware. This re­
`current choice was motivated by the fact that during those years (at the
`end of the eighties or beginning of the nineties) the hardware available on
`the market at a reasonably low cost was not powerful enough to provide
`real-time processing. This problem became even more intense, when the
`processing of images was involved. Therefore, besides selecting the proper
`sensors and developing specific algorithms, a large percentage of this first
`research stage was dedicated to the design, implementation, and test of
`new ad-hoc hardware platforms. The processing engines used to speed-up
`image processing tasks were mainly based on either SIMD computational
`paradigms (namely a substantial number of simple processors performing
`the same operation on a large set of data, usually image pixels) exploiting
`the spatial parallelism of images, or MIMD ones where the processors per­
`form different functions on data subsets, therefore exploiting the parallelism
`of the algorithms. It is essential to emphasize that when a new computer
`architecture is built, not only the hardware aspects -such as instruction set,
`I/O interconnections, or computational paradigm- need to be considered,
`but software issues as well. Low-level basic libraries must be developed
`and tested along with specific tools for code generation, optimization and
`debugging.
`Thanks to the improvement of technology, the situation has changed
`in the last few years, and almost all research groups are shifting towards
`the use of off-the-shelf components for their systems, since nowadays com­
`mercial hardware has reached a low price/performance ratio. Nevertheless,
`although the new generation systems are all based on commercial hard­
`ware, the development of custom hardware has not lost significance, but is
`gaining a renewed interest for the production of embedded systems. Once a
`hardware and software prototype has been built and extensively tested, its
`
` Downloaded from www.worldscientific.com
`
`by 2601:282:700:6b90:5054:63cf:a8f2:87b6 on 12/22/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
`
`

`

`12
`
`Intelligent Vehicles and Machine Vision
`
`functionalities must be frozen and integrated in a fully optimized and engi­
`neered embedded system before marketing. It is in this stage of the project
`that the development of ad-hoc custom hardware still plays a fundamental
`role and its costs are justified through a large scale market.
`
`2.2 Requirements of Intelligent Transportation Systems
`
`Any on-board system for ITS applications that is to be sold on the market
`needs to meet some important requirements which are summarized in the
`following.
`
`Robustness
`The final system that will be installed on a commercial vehicle
`needs to be robust with respect to a number of different