A Markov-based model for performance evaluation in multimedia CDMA wireless transmission

This paper shows that block errors in multicode DS-CDMA (MC-CDMA) transmission systems operating over slow fading channels are accurately modeled by a Markov process with 2M states, where M is the number of multicodes. Although earlier contributions have shown that packet transmission over radio channels can be modeled by Markov processes, error models for DS-CDMA systems using multicode techniques have not been reported so far. Our model includes the effects of the convolutional code, the Viterbi decoder with soft decision decoding, and the RAKE diversity fingers with maximal ratio combining, which have not been considered in previous work. We also describe a new method to estimate the parameters of the Markov model The validity and accuracy of the proposed model are established using information-theoretic arguments and confirmed by computer simulation. As an application, the proposed Markov model is used to analyze the throughput of the GBN and SR ARQ protocols over MC-CDMA systems. Comparisons with results of detailed simulations of the entire system confirm that this is an excellent packet-level error model for MC-CDMA