Distributed gas-source localization based on MMSE estimation with cooperative sensor networks

Sensor network based gas source localization (GSL) has great application prospect in the fields such as environmental monitoring, security protection and pollution control. A novel GSL algorithm based on distributed minimum mean squared error (MMSE) estimation with sensor node cooperation is proposed. In the proposed distributed GSL algorithm, an information fusion objective function which combines the information utility measure and the communication cost between sensor nodes is constructed. The sensor-node scheduling scheme is designed by optimizing the information fusion objective function. For each scheduling sensor node, the MMSE is updated based on its own observation and the estimated result of the previous node. The proposed distributed GSL algorithm is implemented via global and local searches. Simulation results show that both the global search and local search could localize the gas source, however, the local-search scheme is suggested because it can get a significant reduction in the required number of processing nodes to achieve the desired performance. In other words, the local search shows potential advantages in decreasing the congestion, energy consumption and time latency in communication.