A Delta-Smoothed Path Algorithm: An Improved A* Algorithm for Optimal and Safe Trajectory Planning

Path planning plays a crucial role in the real-world applications of mobile robots in cluttered environments. A* algorithm is considered one of the most important algorithms used in path planning. However, it only works with a grid map, and implementing it in real-world applications is challenging. This paper introduces a hybrid algorithm called Delta-Smoothed Path (DSP), a novel algorithm that generates a smooth, safe, and short path. The algorithm technique involves dividing the desired path into straight, long shortcuts in open areas and small, smooth arcs when approaching sharp turns. The suggested method distinguishes itself from previous path smoothing algorithms by employing a critical angle detection mechanism to ensure the avoidance of obstacles and safety zones. Simulations were carried out in three distinct environments—a maze, a complicated environment, and a structured array—to confirm the algorithm's validity. The result shows a significant decrease in path length (up to 12.83%) and a highly reduced number of turning points (by up to 81.81%) compared to the classical A* path. The simulation results demonstrate the effectiveness of the proposed algorithm without any computational complexity, what shows its potential use in real applications.