The cross-correlation between the ocean ambient noise (OAN) fields received at two hydrophones gives an estimate of the Green's function (GF) between the hydrophones' positions, which may be used to passively estimate the bottom profile of the ocean, if the hydrophones are placed vertically. A device using that technique for that purpose is called a passive fathometer. OANs generated at the surface end-fire region due to wind action on the surface provide useful contributions to the passive fathometer, whereas other noises are considered overheads. The estimate can be improved through improving both the signal to noise ratio (SNR) and resolution of the ambient noise cross-correlation function (NCF). The improvement of the SNR and the resolution of the cross-correlation function (CCF) depends on the number of coherent signals extracted from the noise fields where coherent signal is that signal whose propagation path passes through both the hydrophones, which in turn requires more data to be collected. In a given environment, the amount of data obtained can be increased through longer observations, more observation points or collections over broader bandwidths. Long-term averaging has been demonstrated, but requires that the channel be stationary over the averaging time, which can be challenging given that the presence of surface waves often makes the channel non-stationary even over short periods. Large hydrophone arrays are commonly used, but result in increased cost and complexity. According to the literature, using broader bandwidths is limited by the inter-element spacing of the array because of spatial aliasing. However, in this thesis, it is shown theoretically and numerically that estimations of the bottom profile of the ocean can be improved using a broadband hydrophone array which is not affected by spatial aliasing. Moreover, it is shown experimentally that, in bottom profiling, the useful bandwidth of the OAN field is limited by the broadband biological noise in Australian waters not the hydrophone spacing in the array. The effect of broadband shrimp noise on the passive fathometer is demonstrated using experimental data which shows that, in the presence of strong biological activities, the lower band of the ambient noise is useful. The effect of power equalisation on the OAN field, which significantly improves estimates, is also demonstrated. Experimental analyses are performed using both delay and sum (DS) and minimum variance distortionless response (MVDR) beamforming techniques in which the latter gives superior results to the former. A relationship between the array size and bandwidth of the noise field that gives a trade-off between them is also derived.
展开▼
