基于观测偏差校正的无人地面车级联自抗扰跟踪策略

Observation Bias Rectification-based Cascaded Active Disturbance Rejection Tracking Strategy for Unmanned Ground Vehicle

  • 摘要: 无人地面车辆(UGV)在复杂外部环境下难以保持其良好的动态性能。为此,本文提出了级联形式的偏差校正自抗扰控制策略,通过提高观测器对扰动的估计精度来增强UGV控制系统的扰动抑制能力。首先,将广义外部扰动与系统内部扰动重构为总扰动,并设置校正项以扩展初级扰动观测器的观测阶次。之后,为提高观测精度,在次级扰动观测器中对未被及时观测出的扰动残余值进行估计。基于反馈线性化理论,设计了包含总扰动估计值的轨迹跟踪控制律,从而提高UGV的轨迹跟踪误差收敛速度和跟踪精度。为了提高控制器对不同UGV模型的适应性,针对控制器关键参数与闭环模型的对应关系设计了参数配置规则。通过理论分析证明了系统的稳定性和抗扰性能,并在多种情况下进行对比实验以测试所提方法的性能。实验结果表明,所提策略可有效改善UGV在不确定环境下的轨迹跟踪效果。

     

    Abstract: It is difficult for unmanned ground vehicle (UGV) to maintain good dynamic performance in complex external environments. For this problem, a bias rectification-based cascaded active disturbance rejection control (ADRC) strategy is proposed, which enhances the disturbance suppression performance of UGV control system by improving disturbance estimation accuracy of the observer. Firstly, generalized external disturbance and system internal disturbance are reconstructed as total disturbance, and rectification items are set to expand observation order of the primary disturbance observer. Then, residual disturbance unobserved in time is estimated in the secondary disturbance observer, to improve the observation accuracy. Based on feedback linearization theories, a trajectory tracking control law containing total disturbance estimation is designed, so as to improve the tracking error convergence speed and tracking accuracy of UGV. To improve controller adaptability to different UGV models, a parameter configuration strategy is designed according to correspondence between controller key parameters and the closed-loop model. Stability and anti-disturbance performance of the UGV system are demonstrated by theoretical analysis, and comparative experiments are conducted to test the performance of the proposed control method in various situations. The experimental results show that the proposed strategy can effectively improve the trajectory tracking effect of UGV in uncertain environments.

     

/

返回文章
返回