Title | ||
---|---|---|
Smart Measurement Systems Exploiting Adaptive LoRaWAN Under Power Consumption Constraints: a RL Approach |
Abstract | ||
---|---|---|
Nowadays, the development of smart, distributed and possibly wirelessly connected sensors networks is acquiring a greater importance. Indeed, the need to connect mobile or even unreachable devices (sensors, controllers and actuators) is mandatory in the novel smart factory context. Actually, the novel industrial measurement systems must guarantee certain levels of measurement accuracy, while suitably handling time-critical measurement and non - measurement data. Several commonly-used wireless standards can be employed, after careful protocol modifications, aiming at increasing both deterministic and real-time behaviour. Low Power Wide Area Networks (LPWANs) can be suitably adapted to cope with the stringent requirements of the industrial measurement scenario, due to the long communication range and low energy consumption. In particular, this paper focuses on one of them, LoRaWAN, that can be suitably configured through a set of communication parameters. For this reason, the paper aim is to propose and analyse a novel Reinforcement Learning - based adaptation strategy, to maximize the correctly received packets while lowering the energy consumption. The proposed adaptation policy has been compared with the Adaptive Data Rate (ADR) one, that is described by the standard. Results are encouraging, since the proposed methodology achieves better results in terms of correctly received packets compared to ADR, while maintaining similar levels of energy consumption. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1109/MetroInd4.0IoT54413.2022.9831487 | 2022 IEEE International Workshop on Metrology for Industry 4.0 & IoT (MetroInd4.0&IoT) |
Keywords | DocType | ISBN |
Reinforcement Learning,LoRa,LPWANs,ADR,Machine Learning,Artificial Intelligence | Conference | 978-1-6654-1094-6 |
Citations | PageRank | References |
0 | 0.34 | 5 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tommaso Fedullo | 1 | 0 | 1.35 |
Alberto Morato | 2 | 0 | 1.35 |
Federico Tramarin | 3 | 0 | 1.35 |
Paolo Ferrari | 4 | 1 | 2.73 |
Emiliano Sisinni | 5 | 4 | 2.78 |