摘要
本文是以用于高精度测量的磁致伸缩液位传感器为研究对象,通过该种传感器的工作原理,分析了滑油温度的变化对传感器测量精度的影响,并提出了一种基于蚁群聚类算法的RBF神经网络系统对传感器进行温度补偿的方法。该系统使用蚁群算法对输入样本数据进行聚类计算。隐含层到输出层的权值通常使用伪逆算法进行调节。计算各个隐含层单元的输出,并计算其对网络输出的贡献量。通过裁剪法对网络结构进行简化,但要求误差控制在允许误差范围内。
This paper takes high precision measurement of magnetostrictive liquid level sensor as the research object, through the working principle of the sensor, analyzes the sliding oil temperature changes on the measurement precision of the sensor effects, and proposes a method based on ant colony clustering algorithm of RBF neural network system of sensor to compensate temperature. The system uses ant colony algorithm to calculate the data of the input sample. The weights of the hidden layer to the output layer are usually regulated by the pseudo inverse algorithm. calculates the output of each hidden layer unit and its contribution to the network output. The network structure is simplified by cutting method, but it requires that the error is controlled within the permissible error range.
出处
《自动化技术与应用》
2015年第12期5-8,33,共5页
Techniques of Automation and Applications
基金
2013年华东理工大学校级本科教育教学重点改革建设项目(编号YR0126103)
关键词
磁致伸缩液位传感器
RBF神经网络
蚁群聚类算法
温度补偿
BP神经网络
magnetostrictive liquid level sensor
RBF neural network
ant colony clustering
temperature compensation
BP neural network
