Performance analysis of a low-cost solution to vision-based obstacle detection

Since one of the main requirements for road following and platooning functionalities is the robust detection of obstacles and other vehicles on the path, a detailed analysis of the performance of obstacle detection is imperative. This paper presents a critical analysis of the obstacle detection functionality integrated onto the ARGO experimental vehicle, whose main characteristics are low cost and the use of visual information only. Many tests have been performed on different obstacles, with varying shape and size, located in front of the vehicle in a number of different positions. The algorithm has been executed and the results collected and analyzed. This paper surveys the sets of results obtained so far, highlights its characteristics, and discusses the main advantages and problems of such a solution.