Delta并联机器人运动学性能分析与结构参数优化

Kinematic Performance Analysis and Structural Parameter Optimization of Delta Parallel Robot

摘要:

Delta并联机器人具有速度快、结构简单、承载力强等优点，在钵苗移栽、产品分拣与包装中应用广泛。针对目前Delta并联机器人各结构参数对运动学性能的影响及系统性结构参数优化设计缺乏完整理论体系的问题，本文分析可达工作空间雅可比矩阵条件数分布规律、结构参数约束关系、运动学性能随结构参数的变化规律及相关性，获得条件数分布特性和失真约束条件，在此基础上得出动静平台半径差和主动臂长度增加和从动杆长度减小能够使得机构性能较优。给定设计工作空间，对原结构进行参数优化设计，通过建立包络惩罚函数，采用多元非线性拟合与线性加权组合法得到运动学性能评价函数，结合条件数分布特性和失真约束条件建立优化模型，利用遗传算法进行优化。相较于优化前，优化后可达工作空间体积减小14.26%，设计工作空间的全局条件数均值和全局条件数标准差分别减小31.20%和11.78%，且设计工作空间各截面条件数分布规律验证了条件数分布特性的可靠性。

Abstract:

Delta parallel robot has the advantages of fast speed, simple structure and strong bearing capacity, and is widely used in pot seedling transplanting, product sorting and packaging. Aiming at the problems of the analysis of the influence of structural parameters on the kinematics performance of Delta parallel robot and the complete theoretical system for the optimization design of systematic structural parameters, the distribution of condition numbers and the distortion constraint conditions were obtained by analyzing the distribution law of the Jacobian matrix condition in accessible workspace, the constraint relation of structural parameters, and the change law and correlation of kinematic performance with structural parameters. On this basis, it was concluded that the increase of the radius difference of dynamic and static platform, the increase of the length of the driving arm and the decrease of the length of the driven rod can make the mechanism performance better. Then, given the design workspace, the parameters of the original structure were optimized. The optimization model was established combining the condition number distribution characteristics and distortion constraints, the envelope penalty function and the kinematic performance evaluation function that was obtained by using multivariate nonlinear fitting and linear weighted grouping method, and the genetic algorithm was used to optimize the structure. Compared with before optimization, while the volume of the accessible workspace was reduced by 14.26%, the mean value and standard deviation of the global condition number in the design workspace were reduced by 31.20% and 11.78% respectively after optimization. Moreover, the distribution law of the condition number in each section of the design workspace verified the reliability of the condition number distribution characteristics. It provided a reference for the structure optimization and performance analysis of Delta parallel robot.