单神经元自适应PID的机器人恒力控制研究  

Controlling Constant Force of Robot using Single Neuron Adaptive PID

作　　者：饶巍林 彭晋民 阮玉镇 王粲 Rao Weilin;Peng Jinmin;Ruan Yuzhen;Wang Can(School of Mechanical&Automotive Engineering,Fujian University of Technology,Fuzhou 350118,China)

机构地区：[1]福建工程学院机械与汽车工程学院,福州350118

出　　处：《机械科学与技术》2020年第10期1593-1599,共7页Mechanical Science and Technology for Aerospace Engineering

基　　金：福建省自然科学基金项目(2018J05099);福建省教育厅中青年教师教育科研项目(JAT170375)资助。

摘　　要：针对传统基于位置控制的机器人因末端执行器缺少力负反馈环节,难以实现对接触力精确控制问题,建立了一种基于Windows平台和RSI(Robot sensor interface)应用程序包的工业机器人开放式控制系统,在此基础上提出单神经元自适应PID的机器人恒力控制自适应算法。通过在KUKA工业机器人平台实验验证,该力控算法可在未知环境参数情况下实现机器人末端执行器与工件之间恒力接触,并且易于实现;最后通过实验提出了基于所搭建实验平台的单神经元系数K自调整的单神经元自适应PID的机器人恒力控制算法,进一步提高了控制器的自适应性和鲁棒性。A traditional robot that uses position control has difficultyin precisely controlling the contact force because the endeffector lacks the closed-loop control system formed by the negative feedback link of force.Therefore,an open industrial robot control system based on Windows platform and RSI(Robot Sensor Interface)application package is established.On this basis,a single neuron adaptive PID adaptive algorithm for controlling theconstant forceof the robot is proposed.Experiments on KUKA industrial robot platform show that the force control algorithm can realize the constant force contact between the end-effector and the workpiece under unknown environmental parameters,thusbeing easy to implement.Finally,the constant force control algorithm based on the coefficient K self-adjusting single neuron adaptive PID is proposed.Experiments on it are carried out.The proposed algorithm further improves the adaptability and robustness of the control system.

关 键 词：力负反馈 工业机器人 单神经元自适应PID 恒力控制 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]