Application of Artificial Potential Field Algorithm Based on Gauss Function in Obstacle Avoidance of Autonomous Vehicle 2019-01-0692
An improved APF (artificial potential field) algorithm is proposed for obstacle avoidance in autonomous vehicle. With the advantage of high real-time, the traditional artificial potential field and its improved algorithm aiming at the shortcoming of local minimum are widely used in the field of robot research. Compared with the robot as the controlled object, the autonomous vehicle driving in the structured environment is subjected to more constraints of dynamics, kinematics and environment, and requires smoother control. According to the problems of the sharp change of traditional artificial potential field and the constraints of vehicle, the artificial potential field algorithm based on Gauss function is presented in this paper, Artificial potential field model based on Gauss function is built, this model can express the constraints of vehicle, the environmental constraints of vehicle in the structured environment can be constructed by factor µ,the dynamic constraints of vehicle can be satisfied by factor δ. Compared to traditional artificial potential field, Gauss artificial potential field algorithm can achieve more smooth control, Compared to others algorithm for avoid obstacle of autonomous vehicle such as curve fitting, Gauss artificial potential field algorithm does not need to establish the model of complex environment, according to the environmental characteristics can be effective construct of environmental constraints, and combining the kinematic and dynamic constraints of the vehicle achieve the lateral control. The experimental results show that the Gauss function algorithm is effective in the obstacle avoidance of autonomous vehicle, and compared with other algorithms, it has higher real-time performance.
Caijing Xiu, Chao Qu
Guangzhou Automobile Industry Group, University of Pennsylvania