With the wide application of intelligent vehicles to scenarios with frequent changes of dynamic obstacles, the shortcomings of the traditional path planning algorithms used in intelligent vehicles have gradually emerged, these algorithms are originally designed for holonomic robots, such as the lack of dynamic obstacle avoidance performance, the algorithm does not consider the kinematic constraints of Ackermann-steering. In addition, there is a lack of a method for quantifying the dynamic complexity of scenes in dynamic path planning. To overcome these shortcomings, we propose a dynamic window approach tailored for large scene change rate (SCR). Our algorithm integrates Ackermann-steering constraints, dynamic obstacle assessment, and predictive analysis to optimize path planning. By considering vehicle motion and dynamic obstacle data, we dynamically adjust window size and incorporate obstacle trajectory prediction into the cost function. Furthermore, a method for quantifying the dynamic complexity of scenes has been proposed. The proposed algorithm and three other path planning algorithms, DWA, DCP-DWA, and AVO, are simulated and compared in different SCRs, and the physical intelligent vehicle is used for verification. The simulation results demonstrate our method’s superiority in achieving higher target arrival rates and improved smoothness of operation. Additionally, experiments with two Ackermann-steering cars demonstrate the efficiency of our proposed planner.
Zhang et al. (Mon,) studied this question.