Trajectory Planning for Vehicle Autonomous Driving with Uncertainties

This paper proposes a novel method on dynamic trajectory planning for intelligent vehicle driving under traffic environment with uncertainties. The statistical characteristics of traffic vehicle motion are first analyzed with a traffic vehicle model, in which the inputs are considered to be random variables with certain probability distribution. Therefore the output of the model can be calculated via unscented transformation for probabilistic spread. Then the overall collision probability of the candidate trajectories is assessed with certain confidence level. Finally a trajectory planning method is employed to achieve multiple objectives for lane change maneuver with combined efficiency and comfort. Simulation is conducted with results demonstrating that the proposed method is valid and effective.