径向基函数神经网络在多维力传感器标定中的应用

The Calibration of the Multidimensional Force Sensor Based on RBF Neural Network

维间耦合是制约多维力传感器测量精度的主要因素,为了克服传统线性标定方法的局限性,利用径向基函数(RBF)神经网络强非线性逼近能力进行了多维腕力传感器的静态标定,并将其与最小二乘法和BP神经网络标定法作了比较。以研制的六维腕力传感器为对象进行了实验,结果表明,采用RBF神经网络对多维腕力传感器标定比用最小二乘线性标定有更高的标定精度,网络训练速度则大大快于BP神经网络。这种新方法具有一定的实用价值。

The couple of multidimensional force sensor is one major factor to limit the measurement precision. A new method of multidimensional wrist force sensor calibration based on the radial basic function （RBF） neural network is described, Apart from the RBF neural network based multidimensional wrist force sensor calibration methodology, a comparison of calibration results is also presented. These results show that the proposed RBF neural network method has higher precision than that of the least square method and has faster training speed than that of the BP neural network.