On Recognition Of Target Signal For CW Doppler Radio Proximity Fuze

Complex electromagnetic jamming environment is a prominent feature of the modernbattlefield. The radio fuze faces severe electronic background noise interference and themanmade electronic jamming. Under the modern battlefield environment, the accuraterecognition of target signals and jamming signals has important theoretical and practicalvalue for the continuous wave Doppler fuze which is the mainstream of ammunition toimprove its survivability. Based on Fuze channel leakage theory, three channel protectionlevels of continuous wave Doppler fuze and target recognition methods were systematicallystudied in this thesis.First, the three channel protection levels of continuous wave Doppler radio fuze wereanalyzed based fuze channel leakage theory, and the second channel protection level,namely, the protection of correlation between the transmitted signal and received signal,was studied in depth. The processing gain was proposed as a parameter to characterize theanti-jamming performance of the second level channel protection. The quantitativeexpressions of processing gain under the RF noise jamming, noise-AM jamming, noise-FMjamming, square wave-AM jamming and Sinusoidal AM jamming were derivedtheoretically. The simulation model of fuze under the action of target echo signal andjamming signal was established. The fire performance under different jamming modes wastested with typical continuous wave Doppler fuze. The results showed that the theoreticalanalysis, the simulation analysis, the experimental results are consistent with each other.The results shows the following rules: in term of breaking the fuze second protection level,the AM jamming signal is more effective compared with the FM jamming signal and thesquare wave-AM jamming signal and the Sinusoidal AM jamming are more effectiveamong the AM jamming signals.Secondly, the characteristic parameters extraction for the third channel protection level,namely, the fuze signal recognition channel protection level, was studied systematically inthis thesis. The thesis analyzed the lower-level characteristic parameters such as the fuzemagnitude, rate of change of amplitude and frequency etc. The extended characteristicparameter, namely, the ratio of the amplitude of the peak point of the Fourier spectrum, andthe high-level characteristic parameters, namely, the Amplitude Modulation (AM)bandwidth and the Frequency Modulation (FM) bandwidth of IMF through EMD were proposed. The nonparametric statistical analysis method was utilized to analyze theeffectiveness of the proposed parameters in distinguishing the target signal and jammingsignal for fuze. To eliminate the riding waves appearing in the AM/Fm decomposition, anew local normalization method was proposed.Finally, the continuous wave Doppler fuze target signal recognition based on SupportVector Machine (SVM) was studied systematically. Having the priori knowledge ofjamming signals, we utilized the binary SVM to distinguish the target signals fromjamming signals and optimized the kernel functions and the associated parameters throughcross-validation combining with grid-search. Having no priori knowledge of jammingsignals, the single classification SVM method was proposed to determine the characteristicparameter value interval of the fuze target signals. The simulation results show that: thesingle classification SVM has high generalization ability and better goodness of fit.