Obstacle avoidance for autonomous Ground Vehicles based on moving horizon optimization

Obstacle avoidance is a critical problem to solve for Unmanned Ground Vehicles (UGVs). The schemes of obstacle avoidance for UGVs are first discussed according to the structured environment road. The braking algorithm is discussed using moving horizon optimization based on differential flatness, where 3 DOF vehicle model is used to follow the planning trajectory for the obstacle avoidance of UGVs. In addition, the cruise control and changing lane control algorithm are given using 2 DOF vehicle model. The optimal problem of avoiding obstacles is formulated in terms of minimizing the bias of the lateral displacement considering the lateral acceleration as constraints. In order to verify the effectiveness of the proposed method, simulation under multi-vehicle environment is carried out.