Target Tracking Technology Research Based On Multi-sensor Information Fusion

Multi-sensor information fusion is an important research direction in the field of information science. After thirty years’ development, there have been a lot of academic research results in the information fusion field, and it has been widely used in the military and civilian fields, and target tracking is one of the important application field of information fusion technology. In view of many advantages of multi-sensor fusion system over single sensor system, the dissertation studies the theory and practice related to multi-sensor multi-target tracking technology, the dissertation’s main works are as follows:1. Firstly, the two key technologies, data correlation and state estimation in the single sensor multi-target tracking, are mainly analysed, the dissertation also compares the advantages and disadvantages of various algorithms. On the basis of convenient for engineering application, the dissertation chooses the nearest neighbor domain correlation algorithm, - filter and other related technologies, writes a radar terminal track processing and display software based on vc++ on the Windows platform, and tests the performance of the software. Test results show that the software has good multi-target tracking performance.2. Because there are deviation and measurement error between multiple sensors, the dissertation studies the data preprocessing technique of multi-sensor fusion system. Least squares time registration algorithm, cubic spline curve fitting time registration algorithm, polynomial curve fitting time registration algorithm and spatial registration algorithm based on earth-centered earth-fixed(ECEF) coordinate are simulated and analysed. The simulation results show that the data preprocessing techniques can effectively make the measurement of each sensor synchronization on time, and compensate measurement deviation of each sensor on the same target, which contributes to effective track fusion in fusion center.3. On the basis of the fusion structure of sensor to sensor, the distributed track fusion algorithm is studied. In order to verify the validity of the distributed track fusion system,a distributed track fusion simulation system with two sensors is designed. By a lot of system tests in the case of different track fusion algorithm, different target motion model and different sensor measurement error, it shows that the distributed track fusion system can effectively improve the target tracking performance of the system, the performance of the fusion center is the same in both cases of using the optimal distributed fusion algorithm with feedback or without feedback, the feedback can’t improve the target tracking performance of the fusion center, but it can effectively improve the target tracking performance of the local sensors.