Wi-Fi based radio map for location sensing by hypothesizing existence of barriers

摘要

Location-aware services using mobile devices have become popular and many WiFi-based location systems have been developed. Position estimation algorithms in radio waves often use radio maps in order to achieve accurate positioning. Widespread use of WiFi-based location systems requires radio maps precisely representing the actual radio conditions at low cost. We propose an algorithm for making precise radio maps by hypothesizing the existence of barriers. The algorithm we propose can adopt training data collected by many common training methods. This algorithm combines training data and location information of access points in order to infer accurate radio conditions. We explain how the algorithm works when used with training data obtained from an Autonomous Acquision Mechanism. We evaluated and confirmed the utility of our radio map using a particle filter method.