An Improved Hybrid Path Planning Algorithm Combining A* and Dynamic Window Approach for Navigation and Optimization in Dynamic Environments

In recent years, path planning has become a critical research area in robotics, particularly for navigating dynamic environments in autonomous navigation tasks. This study investigates the application of a hybrid path planning algorithm that combines A* and Dynamic Window Approach (DWA) for navigation in dynamic environments. The hybrid A* and DWA approach leverages the global path planning of A*, while DWA enhances local adaptability for obstacle avoidance. This combination enables the robot to adjust its path dynamically, providing smoother navigation in environments with frequent changes. Although there has been extensive research on individual A* and DWA algorithms and their hybrid forms, much of the focus has been on optimizing A* for efficiency, with limited attention to addressing local optimum issues. A deadlock recovery mechanism is introduced to help the robot escape local optima, ensuring continuous progress towards the target even in complex surroundings. Additionally, this paper also introduces a novel turning angle evaluation function and an adaptive function to adjust the coefficients of the evaluation function. The former encourages the robot to choose smaller-angle turns, thereby reducing unnecessary detours and improving path smoothness. The latter will keep the robot away from obstacles by evaluating the weight based on the distance between the robot and the obstacle. Simulation results show that the proposed method significantly improves global path efficiency compared with traditional methods. The hybrid algorithm effectively reduces sharp turns and achieves excellent adaptability in complex dynamic environments.