A Smartphone Management Method to Save Battery Using Bayesian Networks and Fuzzy System

Conventional method for determining the policy to control devices have problems of the race condition of two contexts. It can be solved by combining the probabilities of contexts. In this paper, we propose a method to determine the device control policy from the contexts, which infers the necessity of devices using modular Bayesian networks, and decides the device control policy using fuzzy inference system. To evaluate the proposed method, we collected data with nine college students in two weeks and used the data to evaluate the accuracy of the proposed method. Also we conducted scenario based experiments with three graduate students in four days to confirm the battery saving effect. The proposed method shows 94.2% accuracy, and it saved 490mA more battery than the other method that use support vector machine in a day.