Successive Interference Cancellation for Underwater Acoustic Communications

In this paper, we introduce the addition of an iterative, successive interference cancellation (SIC) process to improve on a multiuser, single-input-multiple-output (SIMO) communications receiver using passive time reversal as a space-time preprocessor. Time reversal has been shown to apply the spatial degrees of freedom to enhance the signal-to-noise ratio (SNR) and suppress interference for a target user. With the introduction of SIC, the receiver can remove the residual interference experienced by each user while preserving the SNR gain achieved by time-reversal preprocessing. The SIC process is a decision-directed approach for removing multiuser interference at the receiver and is similar to the decision-feedback equalizer (DFE) for intersymbol interference (ISI) channels. The interference experienced by each user is estimated at the receiver using previously decoded symbols from interfering users. This estimate is scaled and synchronized before subtraction from the target user's signal after time-reversal combining. Since SIC is applied before symbol decoding, symbol estimates are improved as the process is allowed to iterate until a stationary point is reached. Following time-reversal combining and SIC, a DFE can mitigate any remaining ISI before symbol decisions are made. Data collected from two Focused Acoustic Fields experiments (FAF-05 and FAF-06) are used to demonstrate the performance of the proposed interference cancellation scheme. During the FAF-05 experiment, three users transmitted 16-quadrature amplitude modulation (QAM) symbols simultaneously over the 3-4-kHz frequency band to a 20-element receiving array deployed in 120-m-deep water at a range of 20 km. The FAF-06 experiment included the simultaneous transmissions of 8-QAM symbols from two users over the 9-21-kHz band to a 16-element receiving array in 92-m-deep water at a range of 2.2 km. For both of the examples, SIC is shown to improve the output SNR in the presence of strong interference over time-reversal processing alone. This translates to a significant bit error rate (BER) reduction from 1.53 x 10(-2) to 8.80 x 10(-4) for the FAF-05 data and from 1.77 x 10(-3) to error-free decoding for the FAF-06 data.