Robust Path Planning of Obstacle Avoidance for Unmanned Delivery Robots

Aiming at complex road conditions and difficult path planning of semi-structured roads in a park, a robust path planning method based on sequential quadratic programming (SQP) algorithm optimization is proposed. Firstly, an improved vector field histogram (VFH) algorithm is used to determine an optimal passable area, and then the determination method of the target state is given based on the optimal passable area; Secondly, according to a starting point and the target state, a description method of an obstacle avoidance path is given based on a piecewise quadratic Bezier curve. Then, the problem of curve parameter optimization is established based on a robot running curvature constraint and a direction change margin constraint of the target point, and the SQP algorithm is used to optimize the curve parameters; Finally, the effectiveness and practicability of the proposed method are verified by simulation experiments and experiments of a real robot. Compared with other methods, the proposed method has the shortest path and better robustness.