液压驱动机械手的神经网络控制

针对液压驱动机械手的轨迹跟踪控制问题，在分析研究机械手动力学特点的基础上，推导出机械手的数学模型，分析了神经网络自适应控制的特点，提出了基于函数连接神经网络液压驱动机械手的自学习控制结构与控制算法。其控制结构着重智能知识的加强；控制算法以PD形成学习规则为基础，运算结果不是直接参与控制；而是根据控制器作用于系统之后所产生的误差及其微分对控制器作出评价和修正，仿真和实验研究证明了数学模型的有效性及控制方法的可行性。

The Control of Hydraulic Robot with Neural Networks

The robot driven by hydraulic is a typical electr-hydraulic servo system. The non-linearity,character varing are commen in this kind of sysytem. This paper makes a mathemaic model with neural networks and introduces a new kind of control method which can be used in robot control system. This control method using neural networks jointed with fanction,and has self-learning ability. It is suitable for the curve tracing problem. The construction of this control emphasizes the ability of the knowledge; the learning rule of the networks is based on PD. The result of the calculation is not used to the control system directly. It correct the controller accoding to the error and the differential of the system output. Simulation and experiment show that the mathematic model is right and the control method is effective.