【摘要】 利用工业用6-DOF并联机器人位置反解易于获得这一特性,把较难的六自由度并联机器人位置正解问题转化为应用位置反解结果作为训练样本进行学习,从而实现机械手从关节变量空间到工作变量空间的非线性映射。文中以工业用6-DOF并联机器人为例,采用BP算法对其正解进行了求解,并利用Matlab进行仿真,结果表明该方法计算精度高,克服了数值解法的求解精度受初值影响较大的缺点。经过验证,BP网络的求解精度满足了控制精确的要求。更多还原

【Abstract】 By using the feature that the inverse position solution of 6-DOF industrial parallel robot is easily obtained,we can turn the difficult problem of forward position solution into the result of inverse position solution as the training sample of learning,so as to achieve the non-linear mapping of machine hand from joint variable space to work variable space.Taking a 6-DOF industrial parallel robot as an example,we use BP algorithm to solve its forward solution,and use Matlab to simulate it,the results indicate that the method is of high computational precision and can overcome the disadvantage of numerical method that its accuracy is influenced greatly by the initial value.Through verifying,the solving precision of BP network can meet the requirement of control accuracy.更多还原