No abstract

No abstract

Underwater acoustic channels are band-limited and reverberant, posing many obstacles to reliable, phase-coherent acoustic communications. While many high frequency communication experiments have been conducted in shallow water, few have carried out systematic studies on the channel properties at a time scale relevant for communications. To aid communication system design, this paper analyzes at-sea data collected in shallow water under various conditions to illustrate how the ocean environments (sea surface waves and random ocean medium) can affect the signal properties. Channel properties studied include amplitude and phase variations, and temporal coherence of individual paths as well as the temporal and spatial coherence of multipaths at different time scales. Reasons for the coherence loss are hypothesized.

No abstract

No abstract

Direct-sequence spread-spectrum signals collected from the TREX04 experiment are analyzed to determine the bit-error-rate (BER) as a function of the input signal-to-noise ratio (SNR) for a single receiver. A total of 1160 packets of data are generated by adding ambient noise data collected at sea to the signal data (in postprocessing) to create signals with different input-SNR, some as low as -15 dB. Two methods are analyzed in detail, both using a time-updated channel impulse-response estimate as a (matched) filter to mitigate the multipath-induced interferences. The first method requires an independent estimate of the time-varying channel impulse-response function; the second method uses the channel impulse-response estimated from the previous symbol as the matched filter. The first method yields an average BER <10(-2) for input-SNR as low as -12 dB and the second method yields a similar performance for input-SNR as low as -8 dB. The measured BERs are modeled using the measured signal amplitude fluctuation statistics and processing gain obtained by de-spreading the received signal with the transmitted code sequence. Performance losses caused by imprecise symbol synchronization at low input-SNR, uncertainty in channel estimation, and signal fading are quantitatively modeled and compared with data.

In this paper we report the measurements of temporal coherence of acoustic signals propagating through shallow water using data from three experiments in three different parts of the world, with sound speed standard deviation (STD) varying from 0.3 to 5m∕s near the layer depth. Temporal coherence is estimated from the autocorrelations of broadband channel impulse functions, the latter are deduced from broadband signals transmitted through the ocean during the experiments. The measurements covered three frequency bands: low frequencies below 1.2kHz, midfrequencies between 2 and 5kHz, and high frequencies between 18 and 22kHz. The source-receiver range covers 3, 5, 10, and 42km. The signal coherence-time is defined and deduced from the data. Motivated by previous theoretical work in deep water on the signal coherence-time as a function of the signal frequency, the source-receiver range, and sound speed STD, a similar but empirical analysis is applied to the measured data in shallow water. While the range dependence agrees with the theory, the data exhibit a different dependence on the signal frequency than the theoretical prediction for deep water.

Normal mode amplitudes are definite functions of depth and have a characteristic phase as a function of source range rs [i.e., exp(−ikirs), where ki is the ith mode wavenumber]. The range and depth of an acoustic source in the ocean can then be determined by decomposing array data and beamforming on the mode amplitudes. In particular, the product of the normal mode amplitudes with the steering vector Ui [where Ui=exp(ikir) ] is maximum for the true source range (r=rs). Similarly, the correlation of the measured (decomposed) mode amplitudes with the theoretically calculated mode amplitudes is maximum at the source depth. Accurate range and depth estimation with this approach, however, requires reliable estimates of the mode amplitudes. In this article, an eigenvector decomposition technique is used to extract the mode amplitudes for data received on a long (1-km) vertical array. Range and depth are successfully estimated both for simulated data and for data from the 1982 FRAM IV experiment in the Arctic Ocean. The effect of the number of modes on the estimates is also illustrated.