Vision-based Convoy Following by Recursive Filtering

The problem of convoy following includes automatic speed and steering control of an intelligent vehicle following the motion of a lead vehicle. In this paper, we describe a vision-based scheme to achieve convoy following in real-time. Our approach has been based on Kalman-Bucy filtering (KBF) requiring a simple vision cue. KBF allows recursive estimation of speed and steering as new image measurements are made available from a stereo vision system mounted on the intelligent vehicle. Simulation results demonstrate the robustness of our scheme in the presence of noise. Ongoing research is directed towards testing our module on BART, a testbed vehicle which has been developed at Texas A&M-University for autonomous mobility.