Unsourced Random Access for Distributed State Monitoring in Internet of Things.

Distributed state monitoring is mission-critical in Internet of Things (IoT), in which multiple sensors independently observe the system outputs at different locations and then send them to the fusion center to estimate the global system state vector. As is typical in a linear system, the observation of each sensor is only determined by a subset of components of the global state vector, while the fusion center only cares about the state information embedded in that observation rather than which sensor sends it out. This inspires us that the whole transmission and estimation process can be unsourced, i.e., during which the identities of the transmitter are not needed. Based on this finding, in this paper, we propose a new and efficient approach to accomplish the distributed state monitoring under an unsourced random access(URA) framework. Specifically, all active IoT sensors send their observations to the fusion center using a URA protocol, then the fusion center estimates the global system state and broadcasts a reliability indicator of the estimated state, based on which each sensor decides its activation probability in the next time slot. The procedure stops until all the estimates of the state components are reliable enough. Simulation results verify the excellent performance of the proposed scheme in terms of estimation error and convergence rate.