基于BP神经网络的冗余驱动串并联机构结构参数优化设计

BP Neural Networks Based Redundant Driving of Cartesian Serial-Parallel Manipulator's Optimal Design

对全新结构的万能直角串并联机构进行冗余驱动机构参数进行优化设计,通过建立优化目标函数,以机构的相对工作空间和运动精度为主要指标,通过构建4个BP神经网络对系统进行训练,将已知函数和输入参数通过特定转换后利用神经网络进行分析和逼近运算,得到满足要求的机构参数数据。

Designed the redundant driving structure based on a new type of universal Cartesian serial-parallel manipulator, structural parameters are optimized through objective function, relative workspace and accuracy of motion included as the main inex, four BP neural networks are constructed with reiterative exercise carried out, transferred function and input parameters through the constructed and exercised neural networks, analysis and approach computation completed with pretty optimal data suggested.