A new approach to vision-based road following for autonomous vehicles

A new approach for vision-based autonomous road following is presented. By building on a recently developed optical flow-based theory, the authors suggest that motion commands can be generated directly from a visual feature, or cue, consisting of the projection into the image of the tangent point on the edge of the road, along with the optical flow of this point. Using this cue, fast computationally inexpensive, and simple partial control algorithms are derived for both circular and noncircular roads. Other inputs to these algorithms may be the current steering angle, the current vehicle speed, etc. The output of these control algorithms is the change in steering angle for the next instant of time. Only the kinematics of the vehicle and the camera are considered