自主车辆避障安全路径的可达集建模  

Modeling Method Based on Reachable Set for Safety Path in Autonomous Vehicle Obstacle Avoidance

作　　者：曹凯[1,2] 黄肖肖[1] 于云[1] 刘春[1] 

机构地区：[1]山东理工大学交通学院,山东淄博255000 [2]山东英才学院计算机电子信息工程学院,山东济南250104

出　　处：《系统仿真学报》2016年第3期526-533,共8页Journal of System Simulation

基　　金：国家自然科学基金资助项目(61573009);山东省自然科学基金(ZR2010FM007)

摘　　要：针对传统的路径规划算法在车辆不确定行为状态方面存在的问题,提出将运动车辆视为在连续和离散模式间动态切换的混合系统,采用单一安全目标位置建模车辆避障最优轨迹,并基于多目标轨迹束建立安全状态可达集的建模方法。借此分析了车辆碰撞不可避免的条件,提出了车辆避障最优控制问题的松弛约束。仿真表明,该方法不仅可以规划车辆避障轨迹,还可以推断避障过程中车辆安全状态的可达范围,增强了系统对未来某一时间间隔内车辆状态安全的预测能力。In view of problem in all possible uncertain behavior of vehicle unable existed by the traditional path planning algorithm, a modeling method was proposed which deemed a moving vehicle as a hybrid system switching dynamically between continuous and discrete mode, modeled an optimal trajectory for vehicle obstacle avoidance by using a single security target location, and built a safe state reachable set based on the trajectory beam of multiple security target locations. On this basis, the condition of the inevitable collision of vehicle was analyzed, and the optimal control problem with loose constraints for the vehicle obstacle avoidance was proposed. The simulation results show that the method can not only plan trajectories for vehicle avoiding obstacle, but also infer the reachable range of security state of vehicle, and the predictive power of vehicle state security is enhanced within a future time interval.

关 键 词：安全路径 道路交通 混合系统 可达集 轨迹规划 

分 类 号：X913.3[环境科学与工程—安全科学] U471.15[机械工程—车辆工程]