WiFi White Spaces for Opportunistic LTE-U

In the coexistence problem of LTE-U and WiFi in unlicensed spectrum, interference can result in serious degradation in the performance of the two networks, especially WiFi. Conventionally, LTE is designed as the exclusive operator of the spectrum, whereas WiFi nodes contend in a distributed fashion based on CSMA-CA to access the spectrum. Consequently, WiFi network performance will be pushed to the limit as a result of backing-off for continuous LTE-U transmissions. To mitigate WiFi bandwidth starvation, in this paper we introduce opportunistic utilization of white spaces in the WiFi channel for LTE-U transmissions. Recognizing the dynamic and bursty nature of WiFi traffic, we develop an analytical framework based on Markov modulated batch Poisson processes (MMBPP) to stochastically model the temporal behavior and the idle gaps, or white spaces, in the unlicensed bands. Closed-form expressions are derived and approximations are provided to characterize the duration and the frequency of occurrence of WiFi white spaces. Noticing the considerable availability of idle resources, we discuss the possibilities of exploiting these white spaces for LTE-U transmissions without disturbing and delaying WiFi traffic. Next, under our MMBPP framework, we propose an opportunistic coexistence mechanism which enables the LTE-U base station to dynamically schedule transmissions by estimating the duration of WiFi white spaces, so as to maximize its utilization while limiting the disturbance introduced to the WiFi system. The latency experienced by the coexisting WiFi network is reduced to 90% of the case with ABS, while LTE-U utilization remains comparable.