一种基于IGA-RBF神经网络的传感器动态特性补偿算法  

A Dynamic Compensation Algorithm Based on Improved IGA-RBF Neural Network for Sensor

作　　者：付华[1] 舒丹丹[1] 王新鑫[1] 

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105

出　　处：《计算机测量与控制》2013年第4期1105-1108,共4页Computer Measurement &Control

基　　金：国家自然科学基金(70971059);辽宁省科技攻关项目(2011229011)资助

摘　　要：为了改善传感器的动态特性,减小系统测量误差,对传感器动态性能进行分析,提出了一种基于改进型遗传算法(IGA)和RBF神经网络相结合实现其动态特性补偿的算法,并应用到瓦斯传感器的补偿环节;实验结果表明,经IGA-RBF补偿算法后响应时间仅为0.413s,幅值误差为±5%时的工作频率f≈5.47kHz,不仅响应速度快,精度高,还具有工作频带宽的特点,较大地改善了多项动态特性指标,能够有效地用于传感器的动态特性补偿,在测试领域中有很好的应用前景。To improve the sensor＇s dynamic performance, and reduce errors in systematic measurement, the principle of sensor＇s dynamic performance compensation is analyzed. A kind of improved algorithm （IGA） and the RBF （Radial Basis Function） neural network in the com-pensation algorithm is proposed. The IAGA-RBF compensation algorithm is given and applied to the gas sensor compensation unit. Experi-ments show that the compensation algorithm is of fast response, high accuracy and wide working band, a number of dynamic indicators are al-so largely improved, which can effectively compensate for the dynamic performance of the sensor and has a very good application prospect in the test field.

关 键 词：动态补偿 RBF神经网络 遗传算法 瓦斯传感器 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]