Spectrum Line Detection For Single Vector Hydrophone

Vector hydrophone consists of one acoustic pressure sensor and several particle velocity sensors,measuring both acoustic pressure and particle velocities at one location simultaneously.One kind of vector hydrophones is the optical fiber vector sensor,which transforms the underwater vibration to optical phase,as compared to conventional vector sensors.It has higher sensitivity,greater immunity to electromagnetic interference.Spectrum line detection has always been an important part of underwater target detection.It has wide applications in underwater engineering area and attracts the attention of many sonar designers.This dissertation discusses spectrum line detection techniques basing on a single optical fiber vector hydrophone,aiming at finding new detecting methods using combined information of pressure and particle velocities.The main contents and results of the dissertation are as follows:1.It gives a brief introduction to the sensing principle,structure and technical parameters of optical fiber vector hydrophone.In addition,it also introduces the mathematical description,general process of detection and detector performance indexes.2.It proposes a windowed phase-compensated processor(WPCPR),which is capable of processing real signals.It gives the detailed mathematics of the processor,the expression of the processing gain and the comparison with conventional average power spectrum processor(AVGPR).Theoretical analysis,simulation and experimental data processing reveals that WPCPR has better performance than AVGPR.This paper gives a beamforming interpretation of WPCPR and points out the reason why the SNR improvement is not so good as predicted.This paper also discusses the detection probability,indicating that WPCPR's detection probability is higher than that of AVGPR when SNR and false alarm probability are the same.In short,WPCPR performs better in line-spectrum detection than AVGPR.3.Basing on the signal model of a single vector hydrophone and accounting for self-noise on particle motion channels,this paper derives the expressions of directivity index(DI)for conventional beamformer and Minimum Variance Distortionless Response(MVDR)beamformer.The results reveal that as level of self noise increases,the DI of conventional beamformer decreases to below 0dB and that of MVDR beamformer remains greater than 0dB,which suggests that in spectrum line detection,particle velocity signals should be discarded if conventional beamforming is used and should be kept if MVDR beamforming is used.The correctness of the analysis is confirmed by lake experiment data.4.Based on the circular placement of acoustic pressure sensors in the lake experiment,this paper proposes a function named the Array Gain Loss,which measures the SNR decrease of directly averaging the four acoustic pressure signals over that of the beamforming output.It is found that as frequency increases,the output SNR of averaging four acoustic pressure channels will be lower than that of single pressure channel.5.While the spatial gain of the detector can be enhanced using beamforming techniques,this paper further studies experimentally to increase the temporal gain of the detector by applying WPCPR to the output of MVDR beamformer.Lake experiments indicate that this method can indeed improve the tonal detection performance.