Study On Intelligent Data Fusion Of Target Recognition In A Complex Environment

It is important to situation and threat assessment that targets in battle field are identified timely and exactly, which is the key element of victory in the war. However, modern wars have the trait of air-surface integration and electronic counter measures are taken by both hostile sides, which makes battlefield more complex increasingly. As a result, datum got by sensors are imprecise, half-baked and fallibility. In this condition, multi-sensor target recognition is not carried off in a state of normal. So it is significant for improving performance of multilevel weapon system to study multi-sensor target recognition in a complex interference environment.Data fusion technique is a very active field of processing multiple source information, and the important technique in the future intelligent weapon system. However, there are some disadvantages in the study and application of data fusion technique: Most of the studies on data fusion theory have been made in ideal condition, but few have been in actual working conditions; the lack of self-adaptation ability, adaptive learning and reasoning ability in data fusion and the lack of robustness in the system hardly meet the requirements of multi-sensor target recognition system in the complex interference environment. Therefore, in this paper, model and techniques of intelligent data fusion of multi-sensor target recognition in the complex interference environment are studied, and several intelligent techniques of processing information, such as Fuzzy Theory, Evidence Theory and Neural Network, are adopted in data fusion model and combined with the structure and algorithms of data fusion, which enhances environmental self-adaptation ability, adaptive learning and robustness of target recognition system.Firstly, according to the problems of multi-sensor target recognition in a complex interference environment, an intelligent attribute fusion model is brought forward. The model is constituted by three parts: environment analysis, uncertain information processing and classification information fusion. Environment...