摘要
针对影响智能稳定平台中的倾角仪测量精度的误差,分析了其产生原因.提出了一种基于最小二乘法和RBF神经网络的误差多步补偿方法,首先对倾角仪的测量误差通过最小二乘法对误差进行第一步粗校准补偿;其次,再用RBF神经网络进行第二步精校准补偿.结果表明:经过多步补偿方法之后,倾角仪的测量误差得到明显的提高,误差精度由测量之前的最大误差5.61°基本稳定在±0.1°之内.
Focused on the errors affecting the measurement accuracy of angle gauge used in stabilized platform,the reason which cause errors is analyzed.The multi-step error compensating method of least squared method and RBF networks is proposed.Based on it,an integrated multi-step error compensation method is presented.Least squared method is used to coarsely calibrate the angle gauge output,and then the method of RBF networks is applied to finely calibrate the error.The results show that,after the method of multi-step compensation,the error of the angle gauge is obviously improved,errors with a maximum 5.61° before compensated is controlled with in ±0.1°
出处
《测试技术学报》
2015年第6期534-539,共6页
Journal of Test and Measurement Technology
基金
高等学校博士学科点专项科研基金资助项目(20133219110027)
关键词
稳定平台
多步补偿
最小二乘法
RBF神经网络
intellective stabilized platform
multi-step compensation
least squared method
RBF networks
