气动人工肌肉关节的迭代反馈整定控制及优化

Iterative Feedback Tuning Control and Optimization of Pneumatic Muscle Actuators

气动人工肌肉关节(PMA)具有低成本、柔顺性和与生物肌肉类似的力学特性等优点,在医疗设备、仿生机器人等领域得到广泛应用.本文针对气动人工肌肉充气变形过程中存在的强非线性、时变性和控制参数难以确定的问题,提出了一种基于迭代反馈整定(IFT)算法的数据驱动优化控制策略,直接基于系统的输入输出数据,定义跟踪性能准则函数并采用Gauss-Newton估计算法实现对PID控制器参数的迭代整定,并通过引入辅助因子获取性能准则函数加权因子的最优值进一步加快了IFT算法的收敛速度.仿真结果表明,该方法相对于Ziegler-Nichols等传统PID参数整定方法可以有效提高控制系统的跟踪性能和鲁棒性.

Pneumatic muscle actuators ( PMA) have advantages such as low costs, flexibility, and mechanical properties similar to those of biological muscles, and these advantages make them widely used as equipment in many fields such as medicine and bionic robotics. Considering the problem of highly nonlinear and time-varying features and the difficulty in determining the control parameters, this paper presents a data-driven optimal control method based on iterative feedback tuning (IFT). Based on the input and output data, the method applies the Gauss-Newton approximation algorithm to iteratively tune PID parameters by defining the tracking performance criterion function, and the optimal value of the weighting factor is obtained by introducing an auxiliary factor to further speed up the convergence of the IFT algorithm. The simulation results show that this method can effectively improve the tracking performance and system robustness, compared with the traditional PID parameter tuning method such as Ziegler-Nichols tuning method.