基于有限差分法的双臂关节柔性空间机器人智能递阶控制策略

谐波减速器和力矩传感器等柔性元件广泛应用于空间机器人关节系统，以获取高减速比.这些柔性元件为空间机器人系统引入关节柔性，使得对其的稳定控制变得更为复杂.本文讨论研究了参数不确定双臂关节柔性空间机器人基于有限差分法的智能递阶控制及弹性振动抑制.运用递阶系统理论、动量守恒原理及第二类拉格朗日方法推导出系统递阶动力学模型.利用该模型，设计了基于模糊回归神经网络的非奇异Terminal滑模控制算法和基于有限差分法的滑模控制算法.采用模糊回归神经网络（Recurrent Fuzzy Neural Network，RFNN）逼近系统的不确定部分.为避免复杂的求导计算及角加速度可测要求，利用基于有限差分法的滑模控制来抑制柔性关节振动.由于设计控制器过程中未涉及惯常的奇异摄动双时标分解操作，该控制算法理论上具有适合任意大小关节柔性刚度的优点.系统对比仿真试验证明了智能递阶控制算法优于传统基于奇异摄动法的控制方案. 

Influence of Intelligent Hierarchical Control for Flexible-Joint Dual-arm Space Based on Finite Difference Method

Because of their unique properties and good performance, some elastic components, such as the harmonic reducers and the torque sensors, are widely used in the joints of the space robots and manipulators, in order to obtain high reduction ratio. Due to the existence of elastic components, joint flexibility is introduced into the space robot system, which makes its stability control more complex. As a result, the intelligent hierarchical control based on finite difference method and elastic vibration suppression for dual-arm flexible space robot with uncertain parameters is discussed. The hierarchical dynamics model of the system is derived by applying the theory of hierarchical systems, the principle of momentum conservation of the system and the Lagrangian method of the second kind. By the aforementioned model, the non-singular terminal sliding mode control algorithms based on Recurrent Fuzzy Neural Network (RFNN) and the sliding mode control algorithm based on finite difference method are designed. It is found that the uncertain parts of the system are approached by RFNN, and at the same time, in order to avoid the complicated derivative calculation and the requirement of measurable angular acceleration, the sliding mode control based on the finite difference method is used to suppress the vibration of flexible joints. The control algorithm has the significant advantage of being suitable for an arbitrarily sized joint flexible rigidity in theory because it does not involve the common singular perturbation two-time-scale decomposition operation. The system comparison and simulation results show that the proposed intelligent hierarchical control algorithm is superior to the traditional control scheme based on singular perturbation.