MABOM: A MAB-based Opportunistic MAC Protocol for WiFi/M2M Coexistence

With the development of Internet-of-Things (IoTs) in last decade, Machine-to-Machine (M2M) devices have explosively increased to billions. However, the scarcity of spectrum resources brings great challenge to the popularity of M2M communication. Although unlicensed spectrum gives great choice for M2M communication, the coexistence with WiFi which is popular in this spectrum band must be addressed. Fortunately, among the WiFi communication, there are plenty of white spaces which can be utilized by M2M communication. In this paper, considering the uncertainty of white space duration, we propose a MAB-based opportunistic MAC protocol (MABOM) for M2M communication to coexist with WiFi communication. In MABOM, we model estimation of white space duration into a non-stochastic multi-armed bandit (MAB) problem. Then a scheduling scheme is designed that M2M communication can be scheduled according to the estimated white space duration. Extensive simulations show that MABOM always approaches the optimal scheme, increases the channel utilization effectively as well as achieves higher M2M communication throughout compared with another representative protocol and other two scheduling schemes.