一种新的实时智能汽车轨迹规划方法

针对智能汽车的驾驶决策和轨迹规划问题, 将轨迹表示为轨迹曲线和加速度变化两部分, 以优化轨迹的行驶效率、安全性、舒适性和经济性为目标建立非线性规划模型. 基于序优化思想, 提出混合智能优化算法OODE, 分内、外两层分别优化加速度变化和轨迹曲线, 通过“粗糙” 评价轨迹曲线实现轨迹曲线的快速择优. 仿真结果表明, 所提出的方法能够处理包含多动态障碍物的复杂交通场景, 且具备实时应用能力, 模型的精度和求解速度均优于传统方法.

A novel real-time trajectory planning algorithm for intelligent vehicles

With the trajectory modeled in two parts: trajectory curve and acceleration profile, the problems of decisionmaking and trajectory planning for intelligent vehicles are formulated as a non-linear programming(NLP) model to optimize the efficiency, safety, comfort and economy of trajectory. To solve this model, a hybrid intelligent optimization algorithm OODE is developed. With a two-layer framework applied, OODE optimizes the acceleration profile and trajectory curve in the inner and outer layers, respectively. By “roughly” evaluating the candidate trajectory curves, the optimal curve is determined very efficiently. The simulation results show that, the proposed method is capable of handling complicated traffic scenarios with multiple dynamic obstacles, and also can meet the demands of real-time applications. Compared with traditional methods, the model accuracy of the proposed method is higher, and the planning speed is obviously faster.