Research On Method Of Radar Emitter Recognition Under The Complex Observation Conditions

In the modern information war, radar plays a very important role. Accurately master the status of enemy radar is the key premise to realize the battlefield transparency. However, with the development of radar technology and electronic countermeasures technology, ESM devices are faced with an increasingly complex electromagnetic signal environment. "Low dimension" 、 "Low resolution" and other weak observation problem has become a normal problem which the ESM equipment has to be faced with. And at the same time, with the explosive growth of the number and types of emitters, the “high aliasing” problem is also raised in the radiant characteristic parameter space. These problems lead to the obvious performance degradation of the traditional emitter recognition system. This paper focusing on the above problems and the real output information characteristics of our current ESM equipment, to study a new emitter recognition framework to promote the recognition system performance under the complex observation. Through knowledge aided processing, feature track matching and multi-observation fusion recognition process to enhance the emitter recognition performance. The research work includes two following aspects:First, we study the knowledge-aided emitter coarse classification technique. The different type emitter has different kinds of electromagnetic parameters; the traditional emitter identification methods can’t process those different dimension observations effectively. In order to make full use of the ESM information, we must find a new method to processing the incomplete observation information in a consistent way. Therefore, this paper draw the knowledge aided processing into the emitter recognition procedure to make full use of ESM information and improve the classification effect of radiation source. The problem includes two aspects in detail, the ontology based radar emitter domain knowledge modeling and knowledge-aided emitter coarse classification procedure, as shown in the third chapter of this article.Second, we study the feature track matching based emitter fine classification technique and multi-observation fusion based recognition technique. Along with the increasing complexity and number of the emitters, the problem of emitter electromagnetic parameter reuse and aliasing are becoming more and more serious. The traditional emitter recognition system, which based on the single feature parameter vector matching method, has already can not adapt to the realistic demand. Therefore, this paper presents a feature trace matching based emitter fine classification process and multi-observation fusion recognition method to solve those problems and promote the emitter system performance finally, as shown in the fourth chapter of this paper.By using the knowledge-aided process, feature track matching and multi-observation fusion technique, we not only fulfilled the consistent processing of the incomplete observation, but also reduce the impact of the observation uncertain under complex observation condition. And we also construct the double layer emitter recognition framework, which include coarse classification and fine recognition, to promote the emitter recognition system performance. Simulations results illustrate the effectiveness of the proposed method.