Improved Symbol to Symbol Detection of Inter-Symbol Interference in Channel Equalization

Abstract

Inter-symbol interference (ISI) is a common practical impairment found in many transmission and storage systems, including voice-band modems, digital subscriber loop data transmission, storage disks, digital mobile radio channels, digital microwave channels, and fiber-optic cables. In a receiver for detection of a succession of messages, output sample is the input to the same one-shot detector that would be used on an AWGN (Adaptive White Gaussian Noiseless) channel without ISI. Interference between successive transmissions, or inter-symbol interference, can degrade the performance of symbol-by-symbol detection. This performance degradation increases as the symbol rate increases in most communication channels. Communication engineers use equalization methods to mitigate the effects of the inter-symbol interference. There are several inter-symbol interference and equalization methods, which amount to different structures for the receiver. An alternative (suboptimal) receiver can detect each of the successive K messages independently as presented in this paper. The symbol-by-symbol (SBS) detector, while optimum for the AWGN channel, will not be a maximum-likelihood estimator for the sequence of messages. The bank of matched filters, found by Gram-Schmitt decomposition of the set of channel output waveforms, precedes a block detector that determines the K-dimensional vector symbol transmitted. The complexity would become large or infinite as K becomes large or infinite for the block detector.