论文部分内容阅读
针对无人车车辆路口通行中动态目标的跟踪预测,分析比较基于SICK与Velodyne两种激光雷达数据的车辆提取方法,参照Velodyne的提取方法,提出合适的激光雷达布局对路口环境中的动态障碍物(主要是车辆、行人)信息进行了提取.选取交互式多模型(IMM)算法对动态目标运动趋势进行预测,并对IMM算法进行优化,提出将局部路径规划的三次曲率多项式算法抽象为路径规划模型,作为IMM算法的滤波模型以替代常规的车辆运动模型作为滤波模型.验证实验结果表明基于路径规划模型的IMM算法在无人车车辆运动趋势的预测上具有更好的超前性与更高的预测精度.
Aiming at the tracking and predicting of the dynamic target in the traffic jam of unmanned vehicles, a vehicle extraction method based on SICK and Velodyne laser data is analyzed and compared. According to the Velodyne extraction method, a suitable lidar layout is proposed for dynamic obstacles (Mainly vehicles, pedestrians) were extracted.The IMM algorithm was used to predict the moving tendency of the dynamic target and the IMM algorithm was optimized, and the cubic curvature polynomial algorithm of the local path planning was proposed as a path planning As a filtering model of the IMM algorithm to replace the conventional vehicle motion model as the filtering model.The experimental results show that the IMM algorithm based on the path planning model has better advance and higher predictive value for the movement tendency of the unmanned vehicle Prediction accuracy.