BP网络算法在平面运动测量误差补偿中的应用

在航天对接试验平台开发项目中,针对平面运动物体非接触高精度测量的难点,提出一种针对物体三自由度平面运动的创新性测量方法.利用激光测距传感器的并联组合,实现对物体位置和姿态的非接触式测量.以传感器测量值为输入样本,运用BP神经网络算法建立误差模型,通过调整权值因子使建立的网络模型能够拟合出测量误差,从而修正测量的数学模型.基于BP神经网络的误差补偿方法对平面运动物体测量误差的补偿具有明显的效果,能够满足位姿测量的精度要求.

Application of BP network algorithm in error compensation of planar motion measurement

In the development project of space docking test platform, an innovative measurement method for the planar movement of object with three degrees of freedom was proposed aiming at the difficulties in the non contact high precision measurement of object in planar movement. With the parallel combination of laser ranging sensors, the non contact measurement for the position and posture of object could be realized. Through taking the measured values of the sensors as the input sample, the error model was established with BP neutral network algorithm. Through adjusting the weight factor, the network model could fit the measurement error. Thereby, the mathematical model for measurement could be corrected. The results verify that the error compensation method based on BP neural network has a significant effect on the compensation of measurement error for the object in planar motion and can meet the accuracy requirements in position and posture measurement.