摘要
提出一种用最小二乘支持向量机(LS-SVM)构造函数链接型神经网络(FLANN)逆系统的传感器动态补偿新方法。介绍了相关原理和具体算法,并给出了传感器动态逆系统的数学模型。在该方法中,通过在传感器后串接逆系统模型来修正动态测试误差、提高传感器的动态特性。通过典型的传感器动态标定实验数据,该逆系统模型的传递函数可用LS-SVM-FLANN方法辨识得到。实验结果表明,LS-SVM-FLANN辨识逆系统模型的速度是BP-FLANN方法的两倍,而且该逆系统动态补偿误差仅为后者的10%。用LS-SVM构造FLANN的逆系统补偿器精度高、鲁棒性好、实现简单。
A dynamic compensating method for transducers is presented based on functional link artificial neural networks （FLANN） inverse system constructed by least squares-support vector machine （LS-SVM）. The principle and algorithms are introduced and the dynamic inverse system model for transducers is given. An inverse system model is used behind the transducer to correct its dynamic measurement errors and to improve its dynamic performance. According to dynamic calibrated data of typical transducer, LS-SVM-based FLANN is used to obtain the transferfunction of the inverse system model of transducers. Experimental results show that the speed of the inverse system model identified by the LS-SVM-based FLANN is twice than that of BP-based FLANN, and the dynamic compensation errors are about 10% of the latter. As a result, the dynamic compensating method for LS-SVM-based FLANN has the characteristics of high precision, strong robustness, and easy realization.
出处
《数据采集与处理》
CSCD
北大核心
2007年第3期378-383,共6页
Journal of Data Acquisition and Processing
基金
国家自然科学基金(70272032)资助项目
关键词
逆系统
传感器
动态补偿
函数链接型神经网络
最小二乘支持向量机
inverse system
transducer
dynamic compensation
functional link artificial neural networks（FLANN）
least squares support vector machine（LS-SVM）
