基于D-S推理算法的智能车辆轨迹预测研究  

Research on Trajectory Prediction for Intelligent Vehicles Based on D-S Reasoning

作　　者：黄建根[1] 陈祯福 裴晓飞[2] 董兴智 张杰[1] HUANG Jiangen;CHEN Zhenfu;PEI Xiaofei;DONG Xingzhi;ZHANG Jie(Wanxiang Group,Hangzhou 311215,China;School of Automobile Engineering,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]万向集团公司,杭州311215 [2]武汉理工大学汽车工程学院,武汉430070

出　　处：《汽车工程学报》2018年第1期24-30,共7页Chinese Journal of Automotive Engineering

基　　金：国家自然科学基金青年项目(51505354)

摘　　要：介绍了D-S证据推理的基本原理以及用于轨迹预测不同车辆运动模型。建立基于D-S证据推理的轨迹预测系统的总体结构框架,并且构造了不同传感器的基本概率分配函数。利用Carsim与LabView软件联合仿真并设计实车试验进行轨迹预测算法验证。仿真及试验结果表明,D-S推理算法能在每个预测周期内选择最佳的汽车运动模型,使任意时刻的轨迹预测结果获得最优。Trajectory prediction is one of the hot topics on intelligent vehicles.Firstly,the basic principles of D-S reasoning theory and vehicle motion models for trajectory prediction were introduced.Then,the framework of trajectory prediction system based on D-S theory was established,and the basic probability assignment(BPA)function values for each sensor were constructed.Finally,the co-simulation of Carsim and LabView and vehicle experiment were implemented to verify the proposed algorithm.The simulation and experimental results show that by using the D-S reasoning the best motion model can be selected and therefore the optimal trajectory prediction can be obtained in each prediction period.

关 键 词：智能汽车 轨迹预测 D-S推理 汽车试验 

分 类 号：U461.3[机械工程—车辆工程]