DOPPLER CORRECTION FOR DIGITAL MODULATION IN A FADING CHANNEL WITH SINGLE-INPUT SINGLE-OUTPUT

In a high mobility wireless channel, the Doppler effect is compounded and must be corrected by using pilot-aided symbols. The pilot symbol rate depends on the severity of the Doppler effect. There are existing algorithms such as differential decision-feedback (D-DF) and double-differential decision-feedback (DD-DF) for single-input single output (SISO) systems to improve channel tap estimation. Such algorithms improve the bit error rate (BER) performance of pilot symbol aided decision feedback demodulation. In this thesis, the use of minimum mean square error (MMSE) estimation was implemented to further improve channel tap estimation for the D-DF and DD-DF algorithms. BER performance showed significant improvement for higher order modulation schemes. On the other hand, implementation of the new algorithm on quadrature phase-shift keying (QPSK) showed negligible improvement. Also, the MMSE algorithm is not very effective for cases with very high Doppler frequency shifts. Finally, it was observed that the performance improvement due to MMSE estimation decreases as the signal to-noise ratio increases. To counter this, adaptive tolerances, tied to the variance of channel tap estimation, were found to provide better channel estimation surge detection capabilities. Such implementation dramatically improved the bit error rate performances for D-DF with the MMSE algorithm but was not effective for the DD-DF algorithm.Military Expert 5, Republic of Singapore Air ForceApproved for public release; distribution is unlimited