无人驾驶铰接式车辆强化学习路径跟踪控制算法

发布时间：2018-06-13 10:06

本文选题：铰接式车辆 + 驾驶　；参考：《农业机械学报》2017年03期

【摘要】：针对无人驾驶铰接式运输车辆无人驾驶智能控制问题,提出了一种强化学习自适应PID路径跟踪控制算法。首先推导了铰接车的运动学模型,根据该模型建立实际行驶路径与参考路径偏差的模型,以PID控制算法为基础,设计了基于强化学习的自适应PID路径跟踪控制器,该控制器以横向位置偏差、航向角偏差、曲率偏差为输入,以转角控制量为输出,通过强化学习算法对PID参数进行在线自适应整定。最后在实车道路试验中验证了控制器的路径跟踪质量并与传统PID控制结果进行了对比。结果表明,相比于传统PID控制器,强化学习自适应PID控制器能够有效减小超调和震荡,实现精确跟踪参考路径,可以较好地实现系统动态性能和稳态误差性能的优化。
[Abstract]:To solve the problem of driverless intelligent control for unmanned articulated vehicles, an enhanced learning adaptive pid path tracking control algorithm is proposed. Firstly, the kinematics model of articulated vehicle is derived. According to the model, the model of deviation between actual driving path and reference path is established. Based on pid control algorithm, an adaptive pid path tracking controller based on reinforcement learning is designed. The controller takes lateral position deviation, heading angle deviation, curvature deviation as input and angle control quantity as output, and adaptively adjusts pid parameters online by reinforcement learning algorithm. Finally, the path tracking quality of the controller is verified in the real vehicle road test and compared with the traditional pid control results. The results show that, compared with the traditional pid controller, the reinforcement learning adaptive pid controller can effectively reduce the overharmonic oscillation, track the reference path accurately, and optimize the dynamic performance and steady-state error performance of the system.
【作者单位】：北京科技大学机械工程学院;北京华为数字技术有限公司;
【基金】：国家高技术研究发展计划(863计划)项目(2011AA060404) 中央高校基本科研业务费专项资金项目(FRF-TP-16-004A1)
【分类号】：TP273;U463.6

【参考文献】