ARGO and the MilleMiglia in Automatico Tour

A RGO IS THE EXPERIMENTAL autonomous vehicle developed at the Department of Information Engineering of the University of Parma, Italy. It is a passenger car with a vision-based system for extracting road and environmental information from the acquired images, using different output devices to test the automatic features. ARGO integrates the main results of our last few years’ research on algorithms and architectures for vision-based automatic road vehicle guidance. (Its name is based on the name of the ship of the mythical Greek hero Jason and on Argus, the many-eyed Greek god.) Thanks to ARGO’s availability, we’ve developed, tested, and tuned several different approaches to autonomous navigation, particularly for obstacle and lane detection. We’ve integrated the most promising approaches into the GOLD (Generic Obstacle and Lane Detection) system,1 which acts as ARGO’s automatic driver. GOLD’s primary task is road following. It uses stereo vision to detect and localize obstacles on the road, and processes a single monocular image to extract the road geometry in front of the vehicle. This article focuses on lane detection; we describe obstacle detection elsewhere.1 To extensively test ARGO and GOLD under different traffic situations, road environments, and weather conditions, we carried out the MilleMiglia in Automatico tour, a 2,000-km journey from 1 to 6 June 1998. The tour demonstrated GOLD’s robustness in different terrain, weather, and traffic conditions, and helped to detect situations in which the system has problems, with the aim of enhancing its abilities.