A local path planning algorithm based on improved dynamic window approach.

In order to solve the problem that the traditional DWA algorithm cannot have both safety and speed because of the fixed parameters in the complex environment with many obstacles, a parameter adaptive DWA algorithm (PA-DWA) is proposed to improve the robot running speed on the premise of ensuring safety. Firstly, the velocity sampling space is optimized by the current pose of the mobile robot, and a criterion of environment complexity is proposed. Secondly, a parameter-adaptive method is presented to optimize the trajectory evaluation function. When the environment complexity is greater than a certain threshold, the minimum distance between the mobile robot and the obstacle is taken as the input, and the weight of the velocity parameter is adjusted according to the real-time obstacle information dynamically. The current velocity of the mobile robot is used as input to dynamically adjust the weight of the direction angle parameter. In the Matlab simulation, the total time consumption of PA-DWA is reduced by 47.08% in the static obstacle environment and 39.09% in the dynamic obstacle environment. In Gazebo physical simulation experiment, the total time of PA-DWA was reduced by 26.63% in the case of dynamic obstacles. The experimental results show that PA-DWA can significantly reduce the total time of the robot under the premise of ensuring safety. [ABSTRACT FROM AUTHOR]