Guardian Angel: A Smartphone Based Personal Security System for Emergency Alerting.

Hansong Guo,Liusheng Huang,He Huang,Zehao Sun,Jiade Peng,Zhuolong Yu,Zhenyu Zhu,Hongli Xu,Hengchang Liu

DOI: https://doi.org/10.1109/uic-atc-scalcom-cbdcom-iop.2015.54

2015-01-01

Abstract:In recent years, the number of reports about people assaulted in China and all over the world is significantly increasing, and most of the victims in these assault cases are females, who often lack enough strength to revolt when facing dangerous situations. At the same time smartphones are becoming more and more ubiquitous and being utilized in many fields of our daily lives nowadays, so we argue that smartphones should also play a more important role in the field of personal security. In this paper, we propose a novel system called Guardian Angel for personal security against assaults synthetically leveraging accelerometers, GPS, microphones and vibration motors, which have already been equipped on most off-the-shelf smartphones. As opposed to existing solutions, Guardian Angel follows a more practical, more energy-efficient and more intelligent manner. When confronted with danger, users can ask for assistance by knocking on their smartphones. Guardian Angel recognizes these knock actions with the help of a supervised classifier constructed from thirteen different features. To reduce energy consumption and system overheads ulteriorly, Guardian Angel mines safe locations automatically from GPS positions of users' daily lives utilizing a DBSCAN(Density-Based Spatial Clustering of Applications with Noise)-based algorithm and enters a dormant state when in safe locations. We conduct a series of experiments on Google Nexus 5 smartphones, and evaluation results demonstrate that Guardian Angel achieves 92.8% true positive rate and 0.0% false positive rate about recognizing users' requests and making responses correspondingly.