神经网络求6-SPS并联机器人正运动学精确解  

An accurate solution for forward kinematics of 6-SPS Stewart platform based on neural network

作　　者：邢殿福[1] 杜志涛[1] 景旭[2] 

机构地区：[1]齐齐哈尔大学理学院,黑龙江齐齐哈尔161006 [2]西北农林科技大学信息工程学院,陕西西安712100

出　　处：《齐齐哈尔大学学报（自然科学版）》2006年第2期67-70,共4页Journal of Qiqihar University(Natural Science Edition)

摘　　要：采用BP网络,利用位置逆解结果,通过训练学习,实现操作从关节变量空间到工作变量空间的非线性映射,从而求出6-SPS并联机器人的正运动学解。为提高正解结果精度,引进辅助网络,运用Kalman滤波器进行误差补偿,并给出了并联机器人的仿真实例。仿真结果表明,该法计算精度高且计算速度满足实时控制的要求。The nonlinear mapping from the joint-variable-space to the platform is accomplished with BP neural network after training, and learning. The training data comes from the inversed kinematical model of the platform. To improve the accuracy of the solution, an associated neural network using Kalman filter is used to compensate for the offset error. Numerical results of the simulation of 6-SPS Stewart platform have shown that the accurate solutions can be obtained by performing only a few iteration steps and the computation speed can meet the requirements of the robot＇ s real time control system.

关 键 词：并联机器人 正运动学 神经网络 KALMAN滤波器 误差补偿 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TH113.2[自动化与计算机技术—控制科学与工程]