Performance comparison of iterative/integral equalizer/decoder structures for underwater acoustic channels

The use of an adaptive decision feedback equalizer (DFE) in the demodulation of high speed data transmitted through underwater acoustic channels has been well established. In many channels, however, the performance obtained with the conventional DFE and decoder is not adequate for particular applications. This paper considers four different iterative equalization/decoder techniques for improving the performance of the receiver. One technique uses the hard decisions from the decoder output to feed back to the DFE for making additional passes through the data. The second technique uses the soft outputs from the decoder output to feed back to the DFE. The third technique, termed an integral iterative equalization scheme, is designed to mitigate and correct the errors being fed back to the DFE in a block fashion within the data packet. Finally, the fourth technique, called a turbo equalizer, is an iterative scheme which employs a MAP equalizer and a MAP decoder. These iterative/integral equalization/decoding techniques are applied to convolutionally encoded BPSK and QPSK data received during several field tests. The performance of the iterative equalizer/decoder algorithms is compared on the basis of bit error rate and packet statistics.