Abstract: In view of the imperfection of road potential field when the traditional artificial potential field method was applied in driverless vehicle path planning, an improved artificial potential field model was proposed to obtain the obstacle avoidance trajectory. Firstly, combined with the change of potential energy in the road, the gravitational potential field was improved; the elliptical repulsion potential field was increased to replace the traditional repulsion potential field; the road boundary repulsion potential field was increased to ensure that the vehicle would not drive out of the boundary; the speed potential field of obstacles was increased to improve the rationality of the obstacle. At the same time, a three-degree-of-freedom vehicle dynamics model was established, and the model predictive control algorithm was used for tracking control. The research results show that the improved artificial potential has a good obstacle avoidance path planning, and meets the dynamic constraints in all aspects at the same time. The rotation angle change is within ±3°, and the maximum lateral acceleration is 4.54 m/s2<0.67 μg, which ensures the driving stability and comfort of the planned path.

Key words: vehicle engineering, unmanned vehicle, trajectory tracking, artificial potential field, model prediction

摘要： 针对传统人工势场法应用于无人驾驶车辆路径规划存在的道路势场不完善问题,提出了改进的人工势场模型，用于场算法获取避障轨迹。笔者首先结合势能在道路变化来改进引力势场，增加类椭圆形斥力势场代替传统斥力势场，增加道路边界斥力势场保障车辆不会驶出边界，增加障碍物速度势场来完善障碍物的合理性，同时建立3自由度车辆动力学模型，运用模型预测控制算法进行跟踪控制。研究结果表明：改进的人工势具有良好的避障路径规划，同时满足各方面动力学约束，转角变化在±3°以内，最大侧向加速度为4.54 m/s2＜0.67 μg，保证规划路径具有行驶的稳定性和舒适性。

关键词: 车辆工程, 无人驾驶车辆, 轨迹跟踪, 人工势场, 模型预测