Stationary collectors reading wireless, battery powered smart meters, often operate in harsh channel conditions to cut network installation cost to a minimum, challenging the individual link to each meter. The desired performance measure is reliable reception of at least some data from as many as possible meters, rather than maximizing the number of received packets from one meter. We consider a method for improving the reliable reception in a metering system that operates under the constraints of the popular Wireless M-Bus protocol. We develop a framework for reliable reception in which we use the deterministic protocol structure to obtain side information and group the packets from the same meter. We derive the probability of falsely pairing packets from different senders in the simple case of no channel errors, and show through simulation and data from an experimental deployment the probability of false pairing with channel errors. The pairing is an essential step towards recovery of metering data from as many as possible meters under harsh channel conditions. From the experiment we find that more than 15% of all conducted pairings are between two erroneous packets, which sets an upper bound on the number of additional meters that can be reliably recovered.

The technology of Radio Frequency IDentification (RFID) enables many applications that rely on passive, battery-less wireless devices. If a RFID reader needs to gather the ID from multiple tags in its range, then it needs to run an anticollision protocol. Due to errors on the wireless link, a single reader session, which contains one full execution of the anticollision protocol, may not be sufficient to retrieve the ID of all tags. This problem can be mitigated by running multiple, redundant reader sessions and use the statistical relationship between these sessions. On the other hand, each session is time consuming and therefore the number of sessions should be kept minimal. We optimize the process of running multiple reader sessions, by allowing only some of the tags already discovered to reply in subsequent reader sessions. The estimation procedure is integrated with an actual tree-based anticollision protocol, and numerical results show that the reliable tag resolution algorithm attain high speed of protocol execution, while not sacrificing the reliability of the estimators used to assess the probability of missing tags.