基于改进LPV模型的自动驾驶轻型货车横向控制

Lateral Control of Autonomous Light Truck Based on Improved LPV Model

本文为适应物流自动驾驶轻型货车载荷显著变化的特点,满足低计算负载和高稳定性等需求,提出了一种基于LPV-MPC的路径跟踪控制方法。首先构建线性参变模型,并制定该模型与调度变量-速度和载荷的非线性映射规则,旨在提高行车稳定性并降低系统对参数的灵敏度;然后在滚动优化部分,为解决规划层提供的离散轨迹点稠密程度不匹配控制模块预测层需求的问题,设计了一种轨迹重构的方法,构建了适应预测层时域尺度的平滑轨迹序列,能有效降低预测状态与真实状态的偏差;同时采用了多点状态量偏差预测方式代替单点偏差预测,充分利用了参考轨迹信息从而提高跟踪精度;最后通过联合仿真和实车试验,验证了所提出控制策略的有效性。

For the characteristic of significant load variation in urban logistics autonomous light trucks and to meet the needs of low computational load and high stability,a path-tracking control method based on Linear Parameter-Varying Model Predictive Control(LPV-MPC)is proposed in this paper.Firstly,a linear parameter-varying model is constructed,and nonlinear mapping rules between the model and scheduling variables-speed and load-are established,to improve driving stability and mitigating system sensitivity to parameter fluctuations.Then,for the rolling optimization stage,a trajectory reconstruction method is designed to reconcile disparities between the discrete trajectory points provided by the planning layer and the demand of the control module's prediction layer.A smooth trajectory sequence tailored to the temporal scale of the prediction layer is constructed to effectively decrease the deviation between predicted and actual states.In addition,a multi-point state deviation prediction method is used instead of the traditional single-point prediction,fully leveraging reference trajectory information for improved tracking accuracy.Finally,the effectiveness of the proposed control strategy is verified through combined simulation and empirical vehicle tests.