Collision-aware distributed estimation in WSNs using sensor-censoring random access

We consider a single-hop, wireless sensor network (WSN) performing distributed estimation where the fusion center (FC) will collect local binary estimates within a limited collection time over a single transmission channel. We propose a transmission protocol called sensor-censoring random access (SCRA), in which the collection time is divided into frames and only local binary estimates of the observations in a specific range will be sent in a specific frame by using slotted ALOHA. Since we study a WSN with a single transmission channel, during the collection time, the FC will observe idle time slots, successful time slots, and collision time slots. As a result, we derive a collision-aware maximum likelihood estimator and collision-aware type-based estimators at the FC such that the global estimate is computed from not only the successfully received estimates but also idle time slots and collision time slots. The Cramer-Rao lower bound and mean square error of these estimators are evaluated.