Path planning and autonomous obstacle avoidance for work vehicles based on DWA and motion characteristics

As science and technology advance, the need for autonomous navigation mobile robots and work vehicles rises. Traditional path planning has shortcomings such as excessive reliance on prior environmental information, weak real-time performance of algorithm-based path planning, difficulties in adapting to complex environments, and a lack of consideration for the compatibility with the robot’s own underlying motion information. In this paper, the Dynamic Window Approach (DWA) is used to address these issues by combining it with the work vehicles’ own kinematic model. Firstly, constraints are applied to the speed space of the DWA algorithm to ensure the requirements of dynamic models and obstacle avoidance. At the same time, a function for trajectory prediction selection, weight evaluation of velocity functions, and an added target navigation function are designed to improve the algorithm’s guiding capability toward the target point. Finally, MATLAB is used for simulation and field experiments to verify that the work vehicle can quickly and safely reach the destination. The simulation experiment highlights that the DWA algorithm, fused with the vehicle’s motion data, excels in local path planning in complex environments, enabling efficient obstacle avoidance. The field test further validates these findings, confirming the algorithm’s capabilities for path planning and obstacle avoidance. Therefore, the algorithm proposed in this paper can better address the problems of the traditional DWA algorithm, such as getting stuck in local optima and its inability to effectively handle complex obstacle environments, thereby significantly improving the path planning capability and real-time performance of the algorithm.