Performance Comparisons Among Digital Radio Techniques Subjected to Multipath Fading

A major source of impairment in microwave digital radio is multipath fading, a phenomenon sufficiently prevalent on many paths to frustrate, the attainment of reliability objectives. Performance during multipath fading depends on such system parameters as channel bandwidth and type of modulation, and on details of the receiver processing (e.g., adaptive schemes, synchronization schemes). This paper compares the performance of different digital radio techniques in the presence of multipath fading. The measure used is the expected fraction of time, conditioned on the occurrence of a fading event, that the eye pattern at the receiver detector is closed. This measure is plotted against spectral efficiency (bit rate/channel bandwidth) for various types of modulation (QPSK, offset QPSK, 8-PSK, 16-PSK, 16-QAM, and offset 16-QAM) and for various forms of carrier recovery, timing recovery, transmit/receive filtering, and adaptive equalization. The computations assume a multipath fading frequency response described by a first-order complex polynomial, where the coefficients obey an empirically derived probability law reported previously. The comparisons show clearly the superiority of high-level modulation and the inferiority of offset modulation. The comparisons also demonstrate significant improvements for each of two adaptive equalizers postulated here.