The Multi-sensor Information Fusion Technology Research And Application

In recent years multi-sensor information fusion technology in military and daily life has been widely used, Multi-sensor information fusion technology is from different sensors through comprehensive measurement information to processing Data pretreatment, data association, data decision-making and data fusion, etc. Make a correct judgment and decision process is the purpose of the fusion technology. This thesis mainly research data decisions and data fusion process involving algorithm.First of all, introduced the multi-sensor information fusion methods of management, and introduced multi-sensor of the function model, hierarchical model, and the mathematical model.Secondly, the paper mainly studies the Kalman filter, and proposed the algorithm based on the information entropy of multi-sensor Kalman filtering direct at the characteristics of Kalman filtering. The simulation results show that this method can effectively solve the single sensor information loss and the filtering problems under Gaussian white noise.Again, due to the characteristics of neural network is very suitable for characteristic layer fusion, this paper studied the BP neural network and RBF. And direct at the characteristics of neural network training slow, this paper presents an improved BP neural network algorithm.Finally, this paper emphatically studied DS theory of evidence, evidence theory is mainly used for processing information fusion of decision-making process, and have a wide range of applications. But the DS evidence theory also have some defects, Aiming at the problem of one ticket veto and excessive conflict, this paper propose the improved algorithm. By comparing the simulation concluded that this paper proposed method has good judgment effect, and has good robustness.