Consensus based distributed concentration-weighted summation algorithm for gas-leakage source localization using a wireless sensor network

A distributed estimation of the gas-leakage source location using a wireless sensor network is proposed. The cumulative objective function of centralized least square approach imposes difficulties on its distributed implementation, while the sequential updating of posterior probability density function in the Bayesian inference makes it straightforward to form a distributed algorithm. The objective function of centralized least square method is decomposed to single summands. Then, local estimates aimed at minimizing these summands are calculated in the perspective of sequential Bayesian inference. Finally, the weighted summation of these local estimates forms the global estimates, and the measurements are regarded as the adaptive weights. Consensus algorithm is applied to fuse the local estimates. The computational complexity of local estimation is dramatically decreased, and numerical examples show the fast convergence and considerable accuracy of our algorithm.