1. Introduction

   The Intelligent Vehicle Symposium (IV'2002), sponsored by IEEE and the Intelligent Transportation System Council (ITSC), took place from June 17 (Monday) to 21 (Friday) in Versailles, France. Having had the opportunity to attend the symposium, we present this summary report.

   Directed to achieving greater sophistication of intelligent systems for dedicated use on motor vehicles, IV'2002 was administered by the French National Institute for Research in Computer Science and Control (INRIA) and the Interaction of Vehicle/Infrastructure/Driver Laboratory (LIVIC). Approximately 300 participants attended from every part of the world, and one hundred papers were presented. Following the symposium, the French National Institute for Transport and Safety Research (INRETS), INRIA, and the Laboratoire Central des Ponts et Chaussees (LCPC) presented DEMOS 2002, which featured demonstrations of some twenty experimental vehicles on a nearby test course.

2. Symposium Overview

   The symposium opened with an address by the chairman, Dr. Michel Parent (INRIA), who spoke of how the present symposium was at a turning point where adaptive cruise control (ACC), lane keeping assistance, and other technologies that he had dreamed of ten years earlier had now reached the point of practical application. The symposium then moved into technical sessions in the following nine topical areas at which one hundred papers were presented:

Session 1: Vulnerable Road Users
Session 2: Navigation
Session 3: Human-Machine Assistance and Driver Monitoring
Session 4: Vision
Session 5: Hardware/Software
Session 6: Future Transportation Systems
Session 7: Communication and Control Session 8: Lane Recognition
Session 9: Obstacle Detection and Collision Avoidanc

   There were 51 papers presented (four from Japan) at the symposium, and 49 poster sessions (13 from Japan). It was notable that there were presentations not just from universities and research institutions but also from industry, such as DaimlerChrysler (DC) of Germany. The event was better attended than the previous time, with approximately 300 participants, the majority from Europe and America (35 were from Japan, 5 from South Korea).

A report of our impressions from attendance at the main sessions follows:

(1) Session 1

   Five reports were given on the subject of pedestrian detection. This indicates that the quantum advances in microcomputer performance made in recent years have been making it possible to perform pedestrian detection processing including jumping out, which used to present formidable problems.

   This symbolizes how the rapid development of optical sensors, particularly those based on image processing, has become a powerful impetus for the promotion of research on intelligent vehicles.

(2) Session 1, Session 4, and Session 5

   There were many reports on systems intended to improve performance by the fusion of different kinds of sensor. For example, radar (laser or millimeter wave) that has superior range measurement performance is integrated functionally with optical sensors using conventional stereo cameras or with infrared cameras. Many such systems were also introduced in the demonstrations.

(3) Session 6

   The reports described below were given on future transportation systems. The report on practical application of the low-speed automated driving system for use in a limited area, known as CyberCar, was of great interest.

1) AWAKE Project

   Development of a driver monitoring system began in 2001 with Information Society Technologies (IST) funding.
   Sixteen companies, including DC and Fiat, in eight countries are taking part.

2) Minnesota Snowplow

   The details were introduced in the last AHSRA report.
   Testing on State Highway 7 was completed on April 1, 2002. The program is being extended for one year, however, because the weather conditions were not suitable.

3) CarTALK 2000

   This is a three-year project, initiated in August 2001, to provide danger warnings, ACC, and merge support using inter-vehicle communication. DC, Fiat, and TNO Automotive are each scheduled to provide two demonstration vehicles. In the past year, the project produced communication system requirements.
   When asked whether there was any negative impact, the response was that the project provides driver support, and that this does not extend to responsibility for driver actions.

4) Cybernetic Transport System

   A participant from TNO Automotive of the Netherlands reported on results from the CyberCar program. A pilot People Movers program (FROG Navigation Systems) is being operated at Schiphol Airport among other locations.
   The system concept is low-speed automated on-demand, door-to-door transport with central management.
   Questionnaire surveys of users have brought up points for improvement in the human-machine interface (HMI), and so on. The greatest issue, however, is the legal problem, and there is no legal framework for operation of the system in place at present.
   The CyberCar Consortium is led by INRIA, with participation by 14 organizations (including seven private enterprises) such as TNO Automotive, FROG, YAMAHA, and so on.

(4) Session 7

   Participants from Ohio State University reported on inter-vehicle communication at various ranges.

   Communication at ranges of 1 – 2 km in urban areas and up to 5 km in suburban areas was achieved using 220 MHz (4 kHz band width). Road surface conditions, visibility, road closure (detours), and so on are being considered for message content.

(5) Session 9

   This session was held in the auditorium at the demonstration site, and several of the presentations showed live images (images from escort vehicles, sensor output imagery, etc.) relayed from the test course. Details of the presentations will be given in the next section. This made for extremely effective presentations, and also provided useful information on demonstration methods.

3.Overview of DEMOS 2002 Demonstration

   The test course of GIAT Industry in nearby Versailles–Satory was the venue for 23 demonstrations conducted during the two-day period from Session 9, described above, to the following day (June 21).

1. Vehicle environment monitoring and driver warning
   
   • Lane departure warning for trucks
     
   • Obstacle detection and risk assessment
     
   • Safety zone detection near the front of the vehicle
     
   • Lane edge and obstacle detection and localization
     
   • Surveillance of environmental conditions for driving and vehicle maneuvers
     
   • Lane edge detection and inter-vehicle distances measurement
     
   • Night vision and vocal control
     
2. Cooperation with the driver
   
   • Inter-vehicle distance management
     
   • Lateral and longitudinal control
     
   • Obstacle detection and low-speed automated driving
     
   • Lateral and longitudinal control (overtaking)
     
3. Guiding the driver
   
   • Navigation and speed advice
     
4. Monitoring the intentions and ability of the driver
   
   • Emergency breaking
     
   • Driver gaze monitoring
     
   • Driver awareness
     
5. CyberCars
   
   • Automated shuttles
     
   • Capsule
     
   • Laser guidance for CyberCars
     
   • Platooning
     
   • Automated driving
     
   
   
   

   CyberCar (INRIA: automated driving by a laser guidance)

   
   
6. Research aids
   
   • Intelligent robots
     
   • Driving simulators
     
7. Already commercialized technology
   ACC (speed and distance control)
   

1. France
   ENSMP, LIVIC, LASMEA, Valeo, PSA, Renault, MIAM, INRETS/LESCOT, INNRIA, Robosoft/LASMEA, INRETS/CIR, Velsatis/Renault
   
   
2. Netherlands
   AVV, FROG, TNO Automotive
   
   
3. Others
   TRW (United Kingdom), University of Southampton (United Kingdom), DaimlerChrysler (Germany), SERPENTINE, IAI&UAH

1. We took test rides of some systems, including the lane departure warning system and the FROG CyberCar, and received the impression that they are at the level of practical application. Many of the test cars were equipped with laser radar from IBEO. The cost was not given, but the system seemed to be close to practical application.
   
   
2. The CyberCar from FROG Navigation Systems used transponders as lane markers. They explained that due to the large size and cost of the in-vehicle equipment, their new model to be introduced next year is slated to use magnetic lane markers.
   
   

   Demonstration Truck for Lane Departure Warning (white lane recognition)

   
   

   on-board sensor (IBEO laser radar.etc.)

   
   

   CyberCar (FROG Navigation Systems)

   
   
   
3. There were demonstrations of independent vehicle systems except use of lane maker, and no demonstrations for use in intersections in this time. Therefore, the demonstrations themselves were very different from the field operation tests in Japan, and superficially the tests appeared not to present anything out of the ordinary.
   
   
4. In technical terms, these systems have reached the level of practical application. The market needs are unknown, however, and the participating members took the basic stance that demonstrations were also the first step to take for marketing purposes (which of course includes social and political acceptance).

1. The number of participants and applications to present papers was higher this time than at past symposiums, partly because of the attraction of Versailles as the venue. As observed earlier, however, one reason that activities by this scientific society aimed to achieve greater sophistication of intelligent systems for dedicated use on motor vehicles have grown so much more intense than it was is probably the major advances made in microcomputer performance. These have resulted in competitive development of algorithms and software that have brought image processing to the level of practical application.
   
   
2. There were no presentations on cooperative vehicle-highway at the symposium, no doubt in part because of the nature of this scientific society. The EU, however, has set the objective of a 50% reduction in fatal accidents by 2010, and meeting this target will require measures for intersections, which have a high incidence of accidents. There are limits to the intelligence of independent vehicle systems for dealing with complex intersections, and research on cooperative vehicle-highway is crucial. Consequently, we expect that intersection support will be taken up in detail as a common topic for cooperative vehicle-highway that will be important at the next symposium in Ohio and thereafter.
   
   
3. This field has two main movements.
      One is in the direction of intelligent application, including for general motor vehicles, in order to provide information (including warnings) and some degree of control of driving by means of sensing, positioning, communication, and so on. Venture enterprises are becoming major business players in this area, and they are gradually making moves toward the stage of practical application according to market trends.
      The other movement is toward a new transport system, as symbolized by the CyberCar, by deployment of vehicles that are also capable of automated driving under a limited range of circumstances. Practical application has begun at Schiphol Airport and Rotterdam in the Netherlands, although the scale, of course, is small, and the repercussions are expected to spill over onto the highway, transport, and vehicle fields in the future.
   
   
4. In terms of the technology, this field is also gradually reaching the stage of practical application, and it is starting to come up against the bottleneck of social and political constraints. Industry is now at the stage where it must make a political commitment. The field appears to be entering the time to engage the social and political issues in a serious manner.