A Method of UAV Navigation Planning Based on ROS and Improved A-star Algorithm
Junlin Li, Xingzhong Xiong, Yunhong Yang · 2023
The traditional A-star algorithm searches numerous nodes, paths are more costly, and the obtained result is too close to the obstacle at the inflection point, which may easily cause flight safety hazards. Thus, an improved method is proposed to change the above situations. Firstly, the evaluation function is revised by using dynamic weighting; secondly, use azimuth to change the search neighborhood, and adjust the search method adaptively according to different map areas; then, considering the influence of the actual size of the UAV, set the UAV and the safety radius of obstacles; Finally, inflection points of the generated path are smoothed using Bezier curve. The efficiency and path security are improved while ensuring optimality. The MATPLOT simulation results show that the cost is effectively shortened, both search amount and inflection points are reduced, the path is smoother and safer. Based on ROS experiments, the revised algorithm has excellent effects on obstacle avoidance and smoothness, facilitates UAV operations in indoor scenarios.