Noncoherent equalization algorithms based on sequence estimation

Two noncoherent equalization algorithms are derived, which are based on sequence estimation. Better performance can be obtained than for noncoherent symbol-by-symbol equalizers known in literature. In the first scheme, the current received signal sample is multiplied by the previous complex-conjugated one like in a conventional differential detector prior to sequence estimation. Although this results in nonlinear intersymbol interference, the Viterbi algorithm (VA) can be employed for reconstruction of the data. The second algorithm is derived by modifying the optimum noncoherent sequence estimator, which is too complex for an implementation. The resulting scheme, which again applies the VA, can be interpreted as a generalization of multiple-symbol differential detection for the AWGN channel.