Synergic algorithms for the planning and the intelligent following of a trajectory for non-holonomic vehicles

Although there are many studies about two wheels non-holonomic vehicles, in order to simplify the problem the planning and the following of trajectories or paths are generally considered two different phases. In this paper, two algorithms for the synergic working of trajectory planning with intelligent trajectory following are presented. The "trajectory shape" characteristics obtained in the planning phase are used for the efficient trajectory following. From the shape, in order to follow the trajectory, an intelligent controller calculates, without further elaboration, the reference wheels velocity of the vehicle. The knowledge of trajectory shape characteristics and curvature constraints are used for the optimization (by learning) of the intelligent controller. This solution allows optimal trajectory and path following using a controller not dependent of the vehicle parameter and model. Tests results with different applications demonstrate good performances of the proposed algorithms.