Performance of joint data and channel estimation using tap variable step-size (TVSS) LMS for multipath fast fading channel

Maximum-likelihood joint data and channel estimation with the so-called tap-variable step-size least mean square (TVSS-LMS) algorithm is proposed for the emerging Trans European Trunked RAdio (TETRA). It is shown that the TVSS-LMS is suggested by a simplification of the more sophisticated Kalman channel predictor. The different step sizes can be coarsely determined depending on the respective average channel tap power. Simulation results of a 16-state joint estimator show that for comparable complexity the TVSS-LMS considerably outperforms the conventional LMS in frequency-selective fast fading. Also in flat fading, where differential demodulation is near optimum for TETRA, the TVSS-LMS achieves almost the same bit error rate as a differential demodulation, while the conventional LMS experiences severe degradation