并联机构的位置正解是并联机器人机构应用的基础也是研究机构运动学中的难点之一。常用的解法有解析法和数值法两种，但是这两种解法在针对一些复杂机构时求解难度比较大，而且求出的解不唯一。文章研究了BP神经网络在并联机构位置正解中的应用，通过位置逆解结果和训练学习，实现驱动工作空间到机构动平台变量空间的非线性映射，求解出3-RRP并联机构的运动学正解。最后给出一组仿真实例，通过仿真出得的计算结果可以看出该方法的有效性与可行性。%The solution for forward kinematics of the parallel mechanism is the basis of the parallel robot mechanism application and is one of the difficulties in research institutions kinematics. The commonly used method has two kinds of analytical method and numerical method, but these two solutions for some complex institution to solve the difficulty is more difficult, and the solution is not the only. BP neural network in the position of parallel mechanism is studied in the solution for forward kinematics, through the study of position inverse solution results and training, implementing handlers from joint variables space to work space of non-linear mapping, and solving the 3-RRP of kinematics solution of parallel mechanism. A set of simulation in-stance is given, it can be seen that the result calculated by the simulation of the effectiveness of the method and feasibility.
展开▼
