Poster: SaveAlert: an efficient and scalable sensor-driven danger detection system

Abstract

SaveAlert is an adaptive framework for crowd-monitoring and danger-detection using off-the-shelf smartphones and other peripherals such as smartwatches. It is a system that provides users with an increased awareness of their surroundings by detecting and notifying them of impending danger, by relying only on sensor data collected from the users. Our framework's novelty is in how it performs efficient sensor data collection from potentially a large number of people by limiting the disturbance and stress on the existing Wi-Fi and cellular infrastructure. To the best of our knowledge, this is the first crowd-monitoring framework that takes advantage of peer-to-peer connections to perform local aggregation to alleviate the stress on existing infrastructures for better scalability and efficiency.