Enhanced Sparse A☆ Search for UAV Path Planning Using Dubins Path Estimation

Sparse A* search (SAS) approach has been widely used in robot motion planning and flight vehicle path planning. However, considering the kinematics constraint of UAV, the estimated cost of SAS based on the straight line distance is not accurate which probably gives rise to abundant useless search. Dubins path is the shortest one to connect two poses under the kinematics constraint, thus the estimated cost using Dubins path is guaranteed to satisfy the admissible condition and it is beneficial to resolving accurate cost for the path planning objective. Therefore, the search efficiency of SAS is expected to be improved by introducing the Dubins path. Meanwhile, SAS employs the error of straight line distance from the current node to the terminal as the stopping criterion, which cannot guarantee the direct flyability of terminal path and leads to a farther real flight path. In this paper, the Dubins length is also used in the length evaluation for termination condition to ensure the final planned path could be flied and to increase the the accuracy of the result. Simulations are conducted to test the proposed algorithm, and the results show that the number of expanded nodes for exploring the optimal path is dramatically decreased and the planning time is reduced.