An unobtrusive interaction interface for multiple co-located mobile devices

In this paper, we propose a practical interaction interface for a local set of devices to connect to each other without requiring a lot of effort from their users. We find that highly time-correlated Wi-Fi signal measures experience less variation in their received signal strength (RSS), which alleviates the impediment of the RSS-based solution. We also notice that each room observes significantly discrete signal signature even when they are adjacent to each other. We leverage these observations and rely on the RSS measure to detect and unobtrusively pair devices in the same space. To achieve our goal, we devise a new similarity metric for RSS comparison and propose a means of selecting the cluster threshold. We implement a prototype named Flock for Android devices. By relying on our client-server system architecture, mobile users can interface with others in a few seconds without having to search-and-select nearby devices.