Research On Deterministic Signal Detection Based On Vector Acoustic Sensor

In the underwater acoustic engineering, sometimes we need to make detection ofdeterministic signal, for exemple underwater communication system or active sonar. Whenacoustic signal become very weak, sonar system will not find the signal which buried in thenoise. So making a research on deterministic weak signal detection is meaningful to theunderwater acoustic engineering.In this paper I fixates on the single-frequency continuous signal and LFM pulse signaldetection, and also take advantage of acoustic vector sensor. For the single-frequency Iresearch three methods: multi-layer autocorrelation, the mothod originating from lock-inamplifier, stochastic resonance. I finish formula derivation of the methods, computersimulation for the methods, processing experimental data using the methods and developrapid algorithm for the three methods. The study has proved that: performance of multi-layerautocorrelation is mainly affected by threee parameters: input SNR, length of signal andnumber of iteration, and this method is really suitable for single-frequency continuous signaldetection. The method of Lock-in amplifier is sensitive to frequency shift, only a littlefrequency shift will affect the performance seriously. As to stochastic resonance, parametersare key of its performance. Comprehensive study of the three methods showed that insingle-frequency continuous signal detection performance of multiple auto-correlation methodis best, followed by stochastic resonance method, and the last is Lock-in amplifier.I also do research in the broadband signal detection. In this paper I use LFM pulse, it canbe processed by a classic method called pulse compression of which the rapid algorithm isvery efficient. The performance of pulse compression is determinated by the duration of LFMpulse, the longer, the performance is better.