Remote Monitoring of Two-State Markov Sources in Random Access Channels: Joint Model and State Estimation

Abstract

This paper addresses the problem of remote monitoring of two-state Markov sources via a slotted ALOHA random access channel, where the source statistics are not known a priori to the receiver. We develop a joint model and state estimation method using the Baum-Welch algorithm for two different transmission strategies. In the first strategy, the nodes transmissions are independent of the underlying state evolution process (random policy). In the second strategy, the nodes transmit an update only upon observing a state transition (reactive policy). We show that the reactive approach is beneficial not only in terms of reducing the state estimation error probability (a result that was recently established under perfect knowledge of the source statistics), but that it allows a faster learning of the source statistics.