Reduced-complexity simultaneous beamforming and equalization for underwater acoustic communications

Multichannel spatial signal processing of high-speed underwater acoustic communication signals is associated with computationally intensive receiver algorithms. Besides requirements in computational time, large adaptive filters operating under computationally efficient algorithms imply increased sensitivity to numerical errors and large noise enhancement, which limit their performance at high signalling rates. To overcome these difficulties, a receiver is proposed which consists of a multi-input multi-output combiner in which many input channels are coherently combined into a smaller number, suitable for subsequent multichannel decision-feedback equalization. Receiver operations are optimized jointly to ensure minimum mean-squared error detector performance. Experimental results obtained on a long-range shallow water channel with data rates up to 2000 bits per second demonstrate the receiver capability to fully exploit the spatial diversity of underwater multipath while keeping the receiver complexity at a minimum, thus allowing both fast convergence and little noise enhancement