Joint variational Bayesian based localization estimation algorithm on distributed gas source sensor network

According to the non-linear and unknown diffusion distribution model characteristics of gas leakage diffusion in the real environment, a distributed joint optimal estimation algorithm for gas source localization and parameters estimation was proposed in sensor networks. Firstly, the gas source localization framework was constructed based on the compressed sensing theory, Secondly, the joint sparse estimation method of the state distribution and unknown parameters of gas diffusive source is proposed based on the variational Bayesian inference algorithm. In which, an adaptive grid division strategy is given to improve the accuracy and performance of the joint estimation method and to balance the relationship between energy consumption and network resources. Finally, simulation results show that the proposed algorithm could effectively achieve a joint estimation of the diffusive source state and parameters, which could achieve higher estimation accuracy in a shorter time and meet the real-time requirements in complex environment compared with the traditional CS method.