Turbo equalization of convolutional coded and concatenated space time trellis coded systems using radial basis function aided equalizers

A reduced-complexity radial basis function (RBF) aided neural-network based turbo equalization (TEQ) scheme is proposed for employment in a serially concatenated convolutional coded and systematic space time trellis coded (CC-SSTTC) arrangement. A two-path Rayleigh fading channel having a normalised Doppler frequency of 3.3615/spl times/10/sup -5/ was used. The BER performance of the RBF-CC-SSTTC(4,4) scheme employing a transmission burst consisting of 100 symbols using a space-time-trellis (STT) interleaver of at least 400 symbols and eight turbo equalization iterations was found to be similar to that of the CC-SSTTC system using a trellis-based TEQ, which attains the optimum performance. However, the Jacobian RBF based TEQ provided a complexity reduction factor of 14.