Decentralized multi-exchange scheduling of sensor networks for parameter estimation of distributed systems

An efficient approach to determine an activation policy for scanning sensor network monitoring a distributed process over some spatial domain is proposed. Given a finite number of potential sensor the scheduling problem is defined so as to maximize a criterion defined on the Fisher information matrix associated with the estimated parameters. Then, applying an efficient continuous approximation of the the related combinatorial problem, the multi-exchange procedure adopting pairwise communication schemes is developed, which distributes the configuration process between the network nodes. Finally, the delineated approach is illustrated via simulations on a sensor network scheduling problem for a two-dimensional convective diffusion process.