BP神经网络在加速度计误差补偿中的应用

Application of BP Neural Network in Accelerometer Error Compensation

加速度计受其零偏、温度等影响明显,直接影响导航系统的精度,需要研究补偿方法,提高加速度计的测量精度。鉴于神经网络具有高效的曲线拟合功能和优越的逼近复杂非线性函数的特点,提出了基于BP神经网络的加速度计误差补偿方法。仿真结果表明,经BP神经网络补偿后,加速度计误差补偿后的输出能良好地逼近其期望输出,输出最大误差的绝对值由0.4575g减少到0.07014g,误差降低了一个数量级,很好地抑制了加速度计的误差,提高了加速度计的精度。

Accuracy of accelerometers is badly influenced by zero offset, temperature changing and so on, which affects the accuracy of navigation system directly. It is necessary to research a compensation method to improve the measurement accuracy of accelerometers. Considering characteristics of neural network namely the effective curve-fitting ability and the superior approximation complicated nonlinear function, the accelerometers error compensated algorithm with the BP neural network is proposed. The simulation results show that after the error compensation of the BP neural network, the output of accelerometer can approximate its expected output effec- tively, and the maximum error of absolute value is decreased from 0.4575g to 0.07014g, which is reduced by one order of magnitude. This compensation method has good performance in accelerometers error drift, and the accuracy of accelerometer is improved by using BP neural network.