论文部分内容阅读
从Velodyne雷达构建障碍物地图入手,分析了触须构建方法、自主驾驶和避障策略。考虑车辆质心侧偏角对触须重构的影响,运用卡尔曼滤波对惯导数据进行处理,得出车辆纵向和侧向实时速度,从而对质心侧偏角实时辨识,并利用质心侧偏角对不同分组的触须进行修正。计算结果表明:在中低速时,轨迹误差由0.40 m减小到0.20 m;在高速时,通过采用性能优良的控制器和合理的融合参数使轨迹误差由1.00 m减小到0.75 m。可见,采用修正的触须算法可以较好实现车辆自主驾驶和合理避障。
Starting with the construction of the obstacle map by Velodyne radar, the method of building the tentacles, the autonomous driving and obstacle avoidance strategies are analyzed. Considering the influence of the vehicle center of mass roll angle on the tentacle reconstruction, the Kalman filter was used to process the INS data to obtain the vehicle longitudinal and lateral real-time velocities, so that the center of mass roll angle could be identified in real time. Different groups of tentacles to be amended. The calculation results show that the trajectory error decreases from 0.40 m to 0.20 m at low and medium speeds, and decreases from 1.00 m to 0.75 m at high speed by adopting a good controller and reasonable fusion parameters. Can be seen, the modified tentacles algorithm can be better achieved autonomous driving vehicles and reasonable obstacle avoidance.