Kosunalp S.Mitchell P.D.Grace D.Clarke T.20.04.20192019-04-2020.04.20192019-04-2020161225-6463https://dx.doi.org/10.4218/etrij.16.0115.1030https://hdl.handle.net/20.500.12403/582This paper presents the description, practical implementation, and stability analysis of a recently proposed, energy-efficient, medium access control protocol for wireless sensor networks, ALOHA-Q, which employs a reinforcement-learning framework as an intelligent transmission strategy. The channel performance is evaluated through a simulation and experiments conducted using a real-world test-bed. The stability of the system against possible changes in the environment and changing channel conditions is studied with a discussion on the resilience level of the system. A Markov model is derived to represent the system behavior and estimate the time in which the system loses its operation. A novel scheme is also proposed to protect the lifetime of the system when the environment and channel conditions do not sufficiently maintain the system operation. © 2016 ETRI.eninfo:eu-repo/semantics/closedAccessMedium access controlQ-learningResilience levelWireless sensor networksAccess controlEnergy efficiencyMarkov processesMedium access controlReinforcement learningSystem stabilityTelecommunication networksChannel conditionsEnergy efficientMedium access control protocolsQ-learningReal-world testsResilience levelStability analysisTransmission strategiesWireless sensor networksMedium access controlQ-learningResilience levelWireless sensor networksAccess controlEnergy efficiencyMarkov processesMedium access controlReinforcement learningSystem stabilityTelecommunication networksChannel conditionsEnergy efficientMedium access control protocolsQ-learningReal-world testsResilience levelStability analysisTransmission strategiesWireless sensor networksArticle38591192110.4218/etrij.16.0115.10302-s2.0-84989926636Q2WOS:000384959600013Q3