Iterative correction of ISI via equalization and decoding with priors

An iterative algorithm is presented for joint equalization and decoding of data that has been transmitted over intersymbol interference (ISI) channels. This differs from well-known “turbo equalization” (TEQ) methods, in that the ISI is removed with a soft-input soft-output (SISO) equalizer via linear or decision feedback equalization (DFE). The data is encoded with a convolutional code and interleaved prior to transmission over the channel. At the receiver, symbol estimates are successively refined by passing extrinsic information, in the form of priors over the symbols, between the SISO equalizer and a SISO decoder based on maximum-aposteriori-probability (MAP) symbol estimation. The low complexity of this algorithm make it a practical alternative to existing methods, without sacrificing bit error rate (BER) performance