摘要
基于 BP神经网络算法生成误差补偿单元 ,提出一种可行的采样学习实施方法和高效的算法迭代格式 ,对轴系、齿隙、传动、检测元件和光机电匹配的误差引起的天线伺服机械的测向误差可进行有效校正 ,实现系统高精度的智能化创成。方法简单可行 ,显著提高系统精度 ,测角误差由 2毫弧度减小为 0 .2毫弧度 。
On the basis of BP neural net algorithm that creates error compensation units, we propose a feasible sampling learning method and an efficient iterative format. It makes an efficient correction to direction measured error of antenna servo machanism which is caused by the error of shafting, dentition, transmission, inspect element and photic mechanical electrical match, and realizes intelligent system creation with high precision. This method is practical and simple, which can remarkably improve the system precision, and the angle measured error is reduced from 2 to 0.2 milliradian.
出处
《机械科学与技术》
CSCD
北大核心
2002年第6期876-877,880,共3页
Mechanical Science and Technology for Aerospace Engineering
