An Adaptive Power Management Scheme for WLANs using Reinforcement Learning

Nowadays, most mobile devices are adopting WLAN(Wireless Local Area Network) technology due to its low cost and high performance. Since those devices are usually battery-operated, their power consumption has been a very important problem. The IEEE 802.11 standard provides specifications on PSM(Power Save Mode) which enables the mobile devices to save their energy. However, regardless of its traffic environment or energy status, each device is supposed to use a constant value of LI(Listen Interval), which determines how long it sleeps before it awakes to communicate. In this paper, in order to optimize the sleep/awake period of a mobile station in various environments, we propose an intelligent power management scheme which uses a reinforcement learning algorithm. By dynamically adjusting the LIs of mobile stations, their energy consumption can be optimized, while the trade-off between the energy consumption and the transmission delay is efficiently managed. Using the implementation of our power management scheme based on the NS-3 environment. The simulation results show that our proposed scheme can improve both power consumption and delay performance.