【作者】 何银铜；

【导师】 赵克定； 董彦良；

【作者基本信息】 哈尔滨工业大学 ， 机械电子工程， 2008， 硕士

【摘要】 实现机器人的柔顺性控制的方法主要有两种:力/位混合控制和阻抗控制。其中,阻抗控制得到了广泛的应用。本文所研究的就是阻抗控制,它是通过调节机器人末端的阻抗(包括刚度、阻尼和惯量)使力和位置满足期望的动力学关系。其中,按照目标阻抗的实现方式不同,阻抗控制可分为基于位置的阻抗控制和基于力的阻抗控制。本文是基于位置的阻抗控制系统的研究。本文在原有气液联控伺服控制系统基础上,模拟设计了气液联控机械手臂,引入负载气缸模拟变化的外界环境,购置了可变电阻式位移传感器,从而进行阻抗控制的研究。介绍了气液伺服阻抗控制系统的结构、原理和特点。对该阻抗控制系统进行了理论分析,建立了气液联控机械手臂的数学模型,分析了力跟踪的实现方法。为了实现精确的力跟踪,采用两种策略。一种是通过适当设计方法更新期望跟踪轨迹动力学模型;另一种是估计环境动力学模型(或阻抗)。本文针对以上两种策略研究了三种方法:离线环境参数估计法、在线环境参数估计法和基于神经网络阻抗控制法,并进行了计算机仿真。其中,离线环境参数估计法和在线环境参数估计法是针对第一种策略提出的,基于神经网络阻抗控制是针对第二种策略提出的。用一个弹簧和不计质量的薄板和负载气缸来模拟外界环境,购置了可变电阻式位移传感器,搭建了能够利用计算机进行数据采集和实时控制的气液伺服阻抗控制系统试验台,并开发了基于Windows操作系统具有存储功能的数据采集软件。通过大量试验研究,得到了力跟踪曲线和位置跟踪曲线,证明了基于位置的气液伺服阻抗控制系统实现的可能性。它能够在短时间内同时实现准确的力跟踪和位置跟踪,试验结果是满意的。更多还原

【Abstract】 There are mainly two ways to meet the robot’s compliance control: force /position mixed control and impedance control. Especially, impedance control, which has been widely used, is studied by adjusting the end of the robot impedance (including stiffness, damping and inertia) to make force and location meet the expectations of the dynamic relationship. In accordance with the different objectives, impedance control can be divided into location-based and force-based. This article is based on the location of the impedance control systems.Introduce a gas-liquid servo impedance control system structure, principles and characteristics. Analysis the impedance control system theory and the method of tracking of a Pneumatic Hydraulic Combination Control (PHCC) mathematical model for the dexterous hands. In order to achieve accurate tracking force, there are usually two types of strategies. One is appropriately to update expectations of the dynamics model tracking trajectory, the other is to build a dynamic model of the estimated environment (or impedance). In this paper, there are three methods: offline control, on-line control and impedance control based on neural network. Among them, offline and online control are in response for the first strategy, and impedance control based on neural network is for the second strategy.Design the test platform structures to use the computer for data acquisition and real-time control of the gas-liquid servo impedance control system with a spring and the plate and the load cylinder simulating the external environment. Get force tracking curve and position tracking curve by doing lots of experiments. It proves the possibility of the PHCC-Impedance Control servo system based on Location. Experimental results of the system are satisfactory.更多还原