Multi-channel DFE equalization with waveguide constraints for underwater acoustic communication

Although the multichannel decision feedback equalizer (DFE) has been shown to be nearly optimal and very effective for handling the difficulties of the underwater communications channel, this technique has been slow to be implemented in operational systems due to its high computational complexity. In this paper, we propose using measurements commonly available to oceanographic systems, such as depth, range, and speed of sound, to create a model of the arrival structure at a receiver with multiple elements. Three structures are presented which take advantage of this model to constrain the complexity of the multichannel DFE: a beamspace approach, a time-aligned beamspace approach, and an approach which uses discrete prolate spheroidal functions. Each of these approaches is integrated into a multichannel, direct adaptation DFE which is implemented using a recursive least squares (RLS) algorithm. The proposed structures are tested using the SPACE08 data set across a range of environmental conditions and using several exponential forgetting factors. It is found that these constrained approaches provide significant computational advantages over the full sensor-space approach and performance advantages over other computationally similar algorithms.