Measured performance of 4-QAM with 2-branch selection diversity in frequency-selective fading using only one receiver

The authors present the measured performance of a 500 kb/s 4-QAM (quadrature amplitude modulation) radio link with two-branch selection diversity using only one receiver subjected to hardware-simulated slow frequency-selective Rayleigh fading. The single coherent receiver input is switched between two uncorrelated Rayleigh-fading signals and the one with higher power is selected. The experimental radio link uses time-division multiplexing with a coherent receiver. Since the down-link transmitter is always on the received signal measurements and selection can be performed just before the reception of the user's data burst. The measured performance compares favorably with computer simulation. The effect of delay between measurement/selection and the actual data burst demodulation is also presented. It is found from both experiment and computer simulation that selection diversity is effective in reducing the bit-error-rate floor caused by frequency-selective fading. As a result, a higher data rate can be supported by a multipath fading channel without using adaptive equalization. Thus, a low-complexity hardware design incorporating selection diversity is feasible for a portable radio communication system.<<ETX>>