Successive interference cancellation for time-varying underwater acoustic channels

In our previous work on multiuser communications, two methods were investigated for user separation in a time-invariant, underwater acoustic environment. The first method employed the conventional and adaptive time-reversal mirrors to either passively or actively remove multiuser interference, and the second introduced iterative, successive interference cancellation (SIC) as another active procedure for interference removal. SIC is the multiuser interference equivalent of the decision-feedback equalizer (DFE) for intersymbol interference channels. Much like the DFE, SIC combines previously decoded symbols with channel estimates to remove interference. Both prior works assumed a time-invariant channel, allowing channel estimates obtained during a training period to be valid for the duration of the packet. This paper generalizes the SIC approach to time-varying channels by carefully restructuring the SIC framework and employing a matching pursuit (MP) algorithm to provide data-driven channel updates. The proposed receiver design was used to decode multiuser data collected during the Focused Acoustic Fields 2006 (FAF-06) experiment. Results from this analysis demonstrate that 2 users can successfully be separated by a 16-element receive array with the restructured approach. At a transmission rate of 5 × 103 symbols per second, the architecture is capable of achieving data rates of up to 40 kbps and spectral efficiencies up to 5.33 b/s/Hz.