Modeling and compensation control of asymmetric hysteresis in a pneumatic artificial muscle
Modeling and compensation control of asymmetric hysteresis in a pneumatic artificial muscle
复制标题
DOI：
10.1177/1045389x17698588
复制
发表时间：
2017-03
影响因子：
2.7
通讯作者：
Lina Hao;Hui Yang;Zhiyong Sun;Chaoqun Xiang;Bangcan Xue
中科院分区：
文献类型：
--
作者：
Lina Hao;Hui Yang;Zhiyong Sun;Chaoqun Xiang;Bangcan Xue
气动人工肌肉是一种新型的柔性驱动器，具有高功率密度、安全性和柔性等优良的驱动特性。然而，气动人工肌肉具有很强的非线性和非对称迟滞特性，这给气动人工肌肉的精确轨迹控制带来了很大的困难。本文通过等张试验分析了气动人工肌肉的压力/长度滞后特性。然后，采用扩展的非并行Prandtl-Ishlinskii模型进行描述，并采用带变异部分的自适应权重粒子群算法对模型参数进行辨识。为了进行比较，还考虑了经典的Prandtl-Ishlinskii模型，并用最小二乘法对其参数进行了辨识。基于推广的非平行Prandtl-Ishlinskii模型建立的滞环模型，提出了一种积分逆补偿器，并设计了一种带积分逆补偿器的比例-积分-微分控制器（积分逆-比例-积分-微分）。仿真和实验验证了积分反比例积分微分控制器具有良好的动态性能。与不带滞环补偿器的传统比例积分微分控制器相比，积分反比例积分微分控制器的控制精度提高了43.86%。
Pneumatic artificial muscle is a novel compliance actuator, and it has many excellent actuator characteristics, such as high power density, safety, and compliance. However, it also has strong nonlinear and asymmetric hysteresis, which makes the accurate trajectory control for a pneumatic artificial muscle very difficult. In this article, the pressure/length hysteresis of a pneumatic artificial muscle was analyzed via an isotonic test. And then, it was described using extended unparallel Prandtl–Ishlinskii model, and the model parameters were identified by an adaptive weight particle swarm optimization with a mutation portion algorithm. For the comparison, the classical Prandtl–Ishlinskii was also considered, and its parameters were identified by least square method. Based on the hysteresis model built by extended unparallel Prandtl–Ishlinskii model, an integral inverse compensator was proposed, and then a proportional–integral–derivative controller with the integral inverse compensator (integral inverse-proportional–integral–derivative) was designed. The simulations and experiments validated that the integral inverse-proportional–integral–derivative controller has good dynamic performance. Compared with conventional proportional–integral–derivative controller without a hysteresis compensator, the control precision of integral inverse-proportional–integral–derivative controller is improved by 43.86%.