基于径向基神经网络的加速度计随机误差处理

神经网络应用于加速度计的随机误差处理，更接近真实值，因而比线性方法更优越近年来，BP神经网络受到了广泛重视，但径向基神经网络尚未得到重视径向基神经网络具有比BP神经网络更快的收敛速度，但是径向基神经网络能否达到全局最优解尚没有理论上的判别方法对BP神经网络和径向基神经网络在加速度计输出数据处理方面的优劣势进行了分析，分析时既考虑了数据量的增加，也考虑了优化性未来，对数据以最优化的方式进行大量处理将成为发展的趋势，也是走向实用化的切实可行的发展路径

Stochastic Error Correction of Accelerometer on the Basis of Radial Basis Function Neural Network

When applied to handle the stochastic error of accelerometer, neural network is capable to obtain closer values to the real, which is superior to the linear method. BP neural network has been attached much attention while the radial basis function （RBF） neural network not. Actually, RBF neural network converges faster than BP neural network. But theoretically their still lacks a method to judge whether RBF neural network can find the global optimal solution. The BP and RBF neural networks were applied to handle the data produced by the accelerometer. And the results were compared. Not only the amount of data was increased but also the optimization was considered. In the future, handling mega data in an optimized way will be a development trend and also be a way to practicality.