Impact of Channel Statistics and Correlation on Underwater Acoustic Communication Systems
MISSOURI UNIV OF SCIENCE AND TECHNOLOGY ROLLA
Pagination or Media Count:
Several statistical properties of underwater acoustic channels gathered from experiment data are analyzed. The baseband channel impulse response CIR is estimated using a time domain least squares technique with a sliding window applied to the probing sequences. From the CIR estimation, the probability distribution functions PDFs of the magnitude, real part, imaginary part, and phase of the CIR are calculated. Gamma Rayleigh, and compound k distributions are fitted to the magnitude PDF and the fitness of the distributions are calculated with a two-sample Kolmogorov-Smirnov test. Other statistics such as the autocorrelation function, coherence time, and scattering function are evaluated. The results show that the underwater acoustics channels are worse than the Rayleigh fading commonly seen as the worst case radio channel. Furthermore, the spatial and intertap correlation matrices of multiple input multiple output MIMO systems are estimated using experimental data. It is shown that underwater acoustic MIMO channels exhibit high spatial and temporal correlation. The bit error rate BER of the receiver using Frequency-domain turbo equalization is also evaluated in different channel correlation setups, demonstrating strong effects of the spatial-temporal correlation function on the performance.