摘要
提出了一种基于径向基神经网络误差补偿的磁编码器细分算法.首先对双路正余弦信号进行差值运算并分区,然后采用最小二乘法的线性回归分析获得角度计算函数,最后运用径向基神经网络建立误差模型并进行误差补偿.仿真结果表明,该算法有效提高了磁编码器的输出精度与分辨率,经过补偿后精度为0.09°,分辨率可达到4 096p/r.
This paper presents a magnetic encoder subdivision algorithm based on radial basis neural network error compensation.At first,to calculate out the difference between the dual cosine signal and partition,then using least squares linear regression analysis to obtain the output angle calculation function.Finally,using the radical basis neural network to establish the error model and compensate the error.The simulation results show that the proposed algorithm can effectively improve the output precision and resolution of the magnetic encoder,after compensation,the accuracy is 0.09°and the resolution can reach 4 096p/r.
出处
《杭州电子科技大学学报（自然科学版）》
2016年第2期52-56,61,共6页
Journal of Hangzhou Dianzi University：Natural Sciences
关键词
磁编码器
最小二乘法
径向基神经网络
误差补偿
magnetic encoder
least square method
radical basis neural network
error compensation
