Multi-sensor Data Fusion Algorithm For Target Tracking System

Fast development of science and technology with constantly increasing complexity of technology in both military and unmilitary field drives researchers to use new technologies to assimilate, explain and evaluate excessive information. In that case, people gradually realize the importance of data fusion. Target tracking has broad application in both military and unmilitary area, while sensor has become an essential part of target tracking system. In order to improve the performance of target tracking, it is necessary to make further research in the algorithm of multisensor data fusion.This paper first gives the summarization of the development history, research status quo and developing trend of data fusion algorithm, then introduces some basic algorithms in target tracking and some basic fusion algorithms, and also constructs an algorithm performance evaluation system for data fusion with Matlab software.In the tracking system, the measurement noise of the system might change as the ambience changes, at the same time, the acceleration of the target also might change fast in a short time. Although a lot of algorithms have been introduced to solve one of the problems, none can be used to solve both of them. In order to solve these two problems, this paper presents a new algorithm through combining different methods into the same fusion system. It is a hybrid algorithm combining neural network with fuzzy logic. On the purpose of verifying the performance of this new algorithm, several simulation results are shown in this paper, and from the results, we can tell this algorithm can almost solve the two problems we referred, however, it still has some faults to be solved.