NEWS RELEASE, 12/16/96 Honda, UC Berkeley building automated car for national demonstration next August

by Bill Stone, PATH Publications

Berkeley -- Engineers from Honda R & D North America are joining researchers from the University of California's Partners for Advanced Transit and Highways program (PATH) in an effort to build a computer-controlled car that uses video cameras to avoid obstacles and keep to its own lane. The goal of the project, now underway at UC Berkeley's Richmond Field Station, is to produce a demonstration of some Automated Highway System (AHS) capabilities by August 1997. In the demonstration, two test cars will follow each other down a stretch of highway completely under computer control, "seeing" the lane markings using video cameras and magnetometers, and using the cameras to detect an obstacle and safely steer around it. The test drivers will sit behind the wheel, hands folded, legs crossed, as the cars drive themselves. Two Honda Accords have now arrived at the Richmond facility, the throttle, steering and braking systems are being tested, and the cameras, computers, magnetometers and software are being installed. Honda R & D North America is supplying three 1996 Accords, each with a laser rangefinder and automatically controlled steering, throttle and braking systems. The third car will be a backup vehicle. The project is being led by PATH project manager Daniel M. Empey and principal engineer Damon Delorenzis of the Honda research team in Torrance, Calif. Equipment being added to the cars includes machine vision systems (each consisting of three Sony CCD video cameras linked by computer), magnetometers and vehicle control computers. PATH is supplying the vision system software, as well as the PATH-developed magnetic sensing system that will enable each car to know its velocity and location -- to the inch -- using magnetometers that sense magnetic markers embedded in the roadway. PATH also will provide the cars' control software and communication system. The Honda-PATH Automated Highway System demonstration planned for August 1997 -- one of several demonstrations by teams from across the country -- will comprise four maneuvers: 1) The vehicles will operate independently of human control, using the PATH-developed stereo vision system for lane tracking, and the vision system and Honda Laser radar for longitudinal control. The vehicles will at first be more than 100 meters (330 feet) apart. 2) Using the same sensors as above, and still operating independently, the vehicles will individually see and avoid an obstacle. 3) The vehicles will close to a shorter gap and demonstrate an adaptive cruise control type of operation, in which the cars will stay at the same distance from each other and maintain speed, without intervehicular communication. 4) The vehicles will close to 20 meters (65 feet) or less and commence vehicle-to-vehicle communication using PATH developed software with commercially available wireless local area network (LAN) technology. Lane tracking will be accomplished by the vision system in conjunction with PATH's magnetic-sensor technology, and longitudinal control by the Honda laser rangefinder, with PATH stereo vision system as a backup. Two engineers from Honda R & D, assistant chief engineer for AHS Program Development Masayuki Habaguchi and AHS vehicle systems engineer Mahito Ishiyama, are working together at the Richmond Field Station with Empey; assistant development engineer for vehicle control systems Hung Pham; postdoctoral researcher for vision system development Camillo J. Taylor; and UC Berkeley electrical engineering and computer science department professor Jitendra Malik, with postdoctoral researchers Jana Kosecka (Vehicle Control Systems) and Philip Mclauchlan (Vision System Software). Also on the research team at Honda is chief engineer Toyohei Nakajima in Tochigi, Japan. The California PATH program, founded in 1986, is a joint venture of the University of California, Caltrans, private industry, and private academic institutions, with the mission of applying advanced technology to increase highway capacity and to reduce traffic congestion, air pollution, crash rates, and energy consumption.