Delta机器人拾放操作门型轨迹优化与ANPD控制试验  

Optimal Trajectory Planning and ANPD Control Experiment for Pick-and-Place Operation of Delta Robot

作　　者：陈伟堤 方红根 何文松 杨建国[3] CHEN Weidi;FANG Honggen;HE Wensong;YANG Jianguo(Shanghai Aerospace System Engineering, Shanghai 201109, China;Shanghai Aerospace Equipments Manufacturer, Shanghai 200245, China;College of Mechanical Engineering, Donghua University, Shanghai 201620, China)

机构地区：[1]上海宇航系统工程研究所,上海201109 [2]上海航天设备制造总厂,上海200245 [3]东华大学机械工程学院,上海201620

出　　处：《东华大学学报（自然科学版）》2019年第1期93-101,共9页Journal of Donghua University(Natural Science)

基　　金：国家科技重大专项资助项目(2015ZX04002101)

摘　　要：为实现Delta机器人稳定快速拾放作业,基于次数和长短轴可变的3参数Lamé曲线构造笛卡尔空间门型运动轨迹,采用非对称6次多项式规划门型路径的运动规律,推导离散时间下的位移、速度和加速度公式。为进一步平滑运动轨迹,以关节力矩变化率均值为优化目标,采用遗传算法对Lamé曲线的3个参数进行综合优化,提出Lamé曲线次数的选取方法。在此基础上,通过引入一类饱和非线性函数设计Delta机器人增广非线性PD(ANPD)控制器,对自主设计的机器人开展控制试验。结果表明,该轨迹优化方法能获得平滑运动轨迹,机器人轨迹跟踪误差小,重复定位精度高。In order to achieve stable and quick pick-and-place operation of Delta robot, the trajectory of adept motion in Cartesian space was constructed by using three-parameter Lamé curve. An asymmetrical 6th order polynomial was adopted to generate the motion profiles, and the general formulations of displacement, velocity and acceleration were derived at discrete time. To make the overall trajectory as smooth as possible, a genetic algorithm was applied to optimize the parameters under a cost function which is defined by the average value of joint torque time-derivative. As a result, a technique of selecting Lamé curve order was presented. On this basis, the augmented nonlinear PD (ANPD) controller was designed by introducing a class of saturated nonlinear functions. Experiments were carried out on self-designed Delta robot. The results show that this method can gain smooth trajectory with small tracking errors and high position accuracy.

关 键 词：Delta机器人 Lamé曲线 轨迹规划 参数优化 ANPD控制器 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]