Lane-change manoeuvres for vision-based vehicle

This paper describes algorithms of path generation and lateral control of an automated vehicle with a vision system for autonomous navigation and lane-change manoeuvres. The vision system detects reference lines along which the vehicle is guided. The lateral control consists of steps of reference line detection, path generation, and lateral control calculation. A coefficient of a cubic curve that approximates a series of sampled points on the reference line provides the lateral control. The approximation can be done with a series of sampled points that are detected on the reference line in the field of view with image processing. A path for various manoeuvres including lane-changing is generated by shifting the series of sampled points. Simulation studies and outdoor experiments were conducted to show the feasibility of the algorithms