Paroto Project: Infrared and Radar Data Fusion for Obstacle Avoidance

Abstract

PAROTO is a joint project investigating obstacle detection and avoidance by fusion of information from radar and infrared imagery, with the support of the French ministry of Research. The PAROTO consortium includes SAGEM, LASMEA, INRETS and TEAM.

Detection of obstacles by radar and infrared exhibit different properties so their concurrent use allows optimal performance. Innovative image processing algorithms have been set up to detect vehicles and pedestrians (see [6]). Original 77 GHz radar was also specified and realised. Fusion of information exploits a sound mathematical framework getting the best of both alarm sources. Then a filtering step eliminates pointless alarms, when it is clear that the driver has already taken appropriate actions regarding the obstacle. A dedicated driver’s interface was also designed and realised to warn — visually and/or acoustically — without unnecessarily driving the driver’s gaze out of the road ahead.

Both detection schemes already operate in real time. By the end of 2003, the fusion of information, integrated aboard our tested vehicle, will be evaluated together with the driver interface, providing valuable insight into automated perception dedicated to automotive safety. Results strongly incite us to anticipate the design of a low cost hardware dedicated to these functions.