A multi-level behavior network-based dangerous situation recognition method in cloud computing environments

Abstract

There are a variety of dangerous situations that elementary school students encounter when they commute to their school. Given that smart phone is the one of devises utilized a lot by the elementary school students, it is possible to develop and utilize the safety apps of the smart phone for the elementary school students to protect them from the variety of the dangerous situations. For an example, when the elementary school students encounter dangerous situations in cloud computing environments, the urgent situations can be notified to their parents automatically and the smart phones of theirs can inter-perform with diverse kinds of deployed sensors and actuators. One of research introduces an app that utilizes behavior networks. By applying behavior network two times for recognizing dangerous situations, two different situations can be recognized and handled separately. However, given that the dangerous situations are more complicated, further research is required to improve the processes of the app. This paper proposes a multi-level behavior network-designed method to automatically determine dangerous situations. Behavior network is applicable to the circumstances by using measured values of smart phone sensors. To recognize dangerous situations by utilizing behavior networks, Bayesian probability is also utilized. By learning dangerous situations iteratively, multiple dangerous situations were recognized and handled accurately, which increases the safety of elementary school students.