Multi Sensor Data Fusion Technology Based on Linear Mean Square Estimation and Least Square Method

Abstract

A unified linear data fusion model is proposed to describe the unified data fusion model of measured data, prior information and forecast information; The concept of the information amount of data is put forward, and it points out that the greater the information amount is, the higher the accuracy will be; The optimal data fusion theorem and the information decomposition theorem based on the unified linear fusion model are presented and proved; The essential law of data linear fusion is revealed. In this paper, the least squares algorithm based on optimal weighting, the least squares algorithm based on finite window weighting and the self-learning weighted least squares algorithm are adopted to conduct a fusion processing on the measured data of multiple sensors, respectively. After fusion, the variance of the data is greatly reduced and the estimation accuracy is significantly improved. The simulation results show that the fusion precision of the least square algorithm is higher than that of the traditional least square algorithm.