On the asymptotic performance of TBMA with multichannel diversity over fading channels

In distributed detection for wireless sensor networks, multiple access schemes and channel conditions affect the performance of data fusion at a fusion center together with communication constraints (bandwidth and transmission energy). For an efficient channel use, the type-based multiple access (TBMA) has been proposed by Mergen et al. Since the performance of TBMA is not satisfactory for zero-mean fading channels, it could be extended to exploit multichannel diversity for better performance, which is called TBMA with multichannel diversity (TBMA-MD) in this paper. We study the performance of TBMA-MD over Rayleigh fading channels and show that the error probability decays exponentially in the multichannel diversity gain (not in the number of sensors). Error exponents are also derived for binary hypotheses with closed-form expressions for some cases.