Skip to Main Content
We present two new low-complexity Bayesian-fusion linear receiver structures for synchronous non-orthogonal DS-CDMA wireless sensor networks. The first receiver treats the underlying hypothesis as a system parameter and estimates it to arrive at a global decision. The second receiver is similar to the MMSE multiuser detector, except that it accounts for the covariance between the sensors' transmitted bits in the system model. Both receivers correctly exploit the covariance among the local decisions to achieve improved performance. We evaluate the receivers' performance in the presence of local-sensor and channel errors due to multiple-access interference (MAI), Rayleigh fading, and noise and show that their performance exceeds that of all of the suboptimal receivers proposed in S.K. Jayaweera (Jan 2005).