Research On Comprehensive Target Recognition Based On The Multi-platform Reconnaissance Information

On the modern electronic war, radar's status is proved decisive. Only by getting the information of radar type, radar parameters, radar system and so on, we can judge the plat's characteristics that radar installed on, accordingly offer the reliable decision basis to commanders. So the recognition of radar emitter has become one important function of Electronic Intelligence system (ELINT) and Electronic Support Measure system (ESM).The information which needed by the traditional Electronic Intelligence system mainly depend on radar reconnaissance device of a single platform, but only one of the characteristic of the measured object can be reflected by single transducer, so the recognition of radar type may difficult to acquire better effect. Therefore, Basing on the multi-platform cooperative reconnaissance, this paper mainly studied:(1)The correlation of similar information that from multi-source: Accord to the definition and properties of absolute correlation degree basing on the grey system, correlated signal from the same radar emitter by multi-platforms, established the foundation of comprehensive recognition of radar emitter;(2)Multi-source information fusion: Use the buffering operator upon the groups of radar emitter that correlate correctly, weaken the detecting errors by detect device and battlefield environment, and fuse the correlative signal with least square method, for improving the recognition confidence radar emitter.(3)SVM classification recognizer: With the emergence of many new system and multi-function radars that have complex characteristics in battlefield, making electromagnetic environment become more dense and complex, in addition with the abominable battlefield environment and human disturbance, confirming the greatly uncertainties of characteristics of radar emitter, thus it brings challenge to traditional recognition method. So this paper respectively introduce the establishment of classifier by RBF and SVM algorithm and its application on radar emitter recognition, for showing that SVM classifier is more precise than RBF classifier by simulation and choosing SVM classifier for comprehensive target recognition system design later.(4)Recognition system of radar emitter basing on multi-platforms: Adopt the thought of multi-platform cooperative reconnaissance, share the reconnaissance information provided by other platforms, and integrate the recognition result of each plat-form, for realize the comprehensive recognition of radar emitter. Simulation results indicate that the thought and algorithm presented by this paper can improve the confidence and correct rate of radar emitter greatly, and have some certain feasibilities and engineering values.