Research And Simulation Of Data Fusion System Based On Radar And Infrared Sensor

With the rapid growth of the technology of sensor and information, data fusion is of great significance for improving the detection, recognition and tracking of the target. Data fusion is an integrated science for multi-source information. Compared to the information from any single source,multi-data sources can get much more accuracy in estimation. This paper aims at the research of data fusion based on radar and infrared sensor.First, an overview of the basic principles of single target tracking is presented. The main elements of single target tracking are summed up, including target motion models and target tracking algorithms. The simulation result proves the IMM algorithm based on EKF is effective. Next, a centralized data fusion algorithm is applied to radar and infrared network target tracking. Emulation experiment validates that tracking effect of radar and infrared network is increased rapidly.Then, problems about multi-target tracking of radar and infrared networking are discussed including the formation of tracking gate and data association algorithms. And two kinds of data association algorithm are analyzed by simulation.The track association of radar and infrared is an important part in multi-sensor data fusion.Recently people usually realize it by using the two angles in measurement information. This paper presents a novel method by using filtering information instead of measurement information. The simulation proves the algorithm achieves higher correct rate of track-to-track association and reduces the influence of measurement accuracy.Tracking performance is bad to anti-jamming and unstable in Cartesian coordinate using EKF because of the bearing-only problem existing in the IR sensor. A distributed fusion model with radar revised is presented. The results indicate the precision of the infrared tracking is better when using the model proposed. Meanwhile, the anti-jamming ability of the system is strengthened.Finally, a simulation software platform is designed and developed in MATLAB environment in order to simulate, test and evaluate the radar and infrared network data fusion. Next the structures and the functions of the platform are introduced. The effectiveness of the algorithms and practicality of the platform are tested by an instance running in the platform.