Abstract

In many countries, the population is either declining or rapidly concentrating in big cities, which causes problems in the form of vacant houses in many local communities. It is often challenging to keep track of the locations and the conditions of vacant houses, and for example in Japan, costly manual field studies are employed to map the occupancy situation. In this paper, we propose a technique to infer the locations of occupied houses based on ambient WiFi signals. Our technique collects RSSI (Received Signal Strength Indicator) data based on opportunistic smartphone sensing, constructs hybrid networks of WiFi access points, and analyzes their geospatial patterns based on statistical shape modeling. We show that the technique can successfully infer occupied houses in a suburban residential community, and argue that it can substantially reduce the cost of field surveys to find vacant houses as the number of potential houses to be inspected decreases.