Abstract
In this work, a joint distributed estimation problem for multiple source wireless sensor network is considered. A wireless sensor network involves numbers of sensors and a fusion center. We design the estimation algorithm in the fusion center, and minimize the quantization effect to improve the performance of estimator. Let each sensor observe some combined signal from different source, quantize these row data into a finite symbol set, and transmit a finite number of bits to the fusion center. We consider the observation noise distribution is Gaussian, and ideal backward communication channels. First, we design a feasible joint maximum-likelihood (ML) estimator for centralized sensor network to avoid searching estimation result in infinite space. Second, we apply this estimator to our decentralized environment. In conclusion, we modify our algorithm to reduce the bad effect of quantization. The proposed method has better performance, or we say less mean square error (MSE), than the conventional one.