Efficient Power Saving Method for WiFi Direct Devices in IoT based on Hidden Markov Model

摘要

Internet of Things Technology demands interconnection of large number of devices and energy saving is the prime factor in this evolving field due to limited battery energy in the devices. Wi-Fi Direct seems to be an ideal connectivity mechanism for in-vehicular and smart home environment. Wi-Fi Direct has an opportunistic power saving mechanism where in interval between successive doze state is constant and doesn't consider any external factors. Battery constraint and randomness in type of data received from clients demands statistical model to predict an appropriate interval. We have modelled Wi-Fi Direct group owner as a finite automaton which is irreducible, a priori and has well known emission states in the form of data size received and finite hidden state based on battery percentage. Hidden Markov Model (HMM) based approach for our model determines appropriate interval between successive doze state and thereby increasing battery longevity by close to 8%.