The effects of filtering and nonlinear amplification on the performance of DS-CDMA with OQPSK-type modulations

This paper investigates the effects of channel filtering and nonlinear amplification on the signal-to-noise ratio (SNR) performance of direct-sequence code-division multiple-access systems employing offset quadrature modulation formats with various chip waveforms. An average SNR analysis is developed and followed by numerical examples to illustrate relative performance merits among modulation formats, including offset quaternary phase-shift keying (OQPSK), minimum shift keying (MSK), sinusoidal frequency shift keying (SFSK), and time-domain raised-cosine shaped QPSK (TDRC). The channel filtering considered is based on Butterworth-type filters with varying 3-dB cutoff bandwidth, and the nonlinear amplification is modeled by amplitude hard limiting. The results obtained show little SNR performance differences between the cases of filtering alone and combined filtering and hard limiting. It is also found that for tighter filtering (one-sided 3-dB bandwidth on the order of half the chip rate), all modulations have comparable performance, but that for wider filter bandwidth values, differences become more noticeable, with SFSK and TDRC slightly better than MSK (on the order of 0.5 dB) and clearly outperforming OQPSK (by up to 1.5 dB). On the other hand, examples with strictly bandlimited signals employing frequency-domain Nyquist raised cosine (FDRC) filtering with a rolloff factor varying from zero to one showed a clear degradation in performance (by about 1 dB) due to the introduction of hard limiting. These schemes were also outperformed by all of OQPSK, MSK, SFSK, and TDRC modulations when combined with strict "brick wall" filtering, with SNR differences ranging from 1 to 2 dB.