Multi-sensor data fusion algorithm based on the improved weighting factor

Abstract

Abstract Aiming at the decrease of the accuracy of fusion data caused by the abnormal value and noise interference in the multi-sensor observations, this paper proposes a multi-sensor data fusion algorithm based on improved weighting factors. Firstly, the Dixon criterion is used to eliminate outliers in observations to avoid data containing gross errors. Then the Kalman filter algorithm is used to effectively reduce the noise impact caused by various reasons and provides the optimal data for weighted data fusion. Finally, an improved weighted fusion algorithm is used to comprehensively consider the nature of the sensor and the influence of various factors in the measurement process to obtain the best fusion data. The simulation analysis of the soil humidity in the greenhouse shows that the error of the multi-sensor data fusion algorithm based on the improved weighting factor is maintained at 0.04%-0.18%. Compared with the adaptive weighted fusion algorithm, the error of this algorithm is reduced by 0.12%, which verifies the algorithm’s effectiveness.