Low power IoT communication signals (e.g., Bluetooth and ZigBee) may seriously suffer from the presence of high-power co-channel interference like wireless local area network (WLAN) signal. They may effectively avoid WLAN interference by exploiting dynamic characteristics of WLAN traffic. Representing the arrival/departure of WLAN users using an M/M/m/m queueing structure, we consider the characterization of large-scale dynamics of WLAN channel occupancy. We also consider the characterization of small-scale dynamics of WLAN channel occupancy by generating WLAN signal using a two-state semi-Markovian process. Simulation results show that the proposed model generates WLAN signal having similar statistical characteristics to those of real WLAN signal.