MIL: A mobile indoor localization scheme based on matrix completion

Jie Cheng, Zeqi Song, Qiang Ye, Hongwei Du

Abstract: Mobile indoor localization is the foundation for the location-based features of many pervasive computing applications. Due to the popularity of WiFi networks in indoor environments, WiFi-based indoor localization has been considered to be a promising approach. Despite the feasibility of WiFi-based localization, the existing WiFi-based schemes suffer from the serious problem of low precision. In this paper, we propose a high-precision indoor localization scheme for mobile networks, Mobile Indoor Localization (MIL). Technically, MIL adopts a matrix completion approach which efficiently utilizes the collected information to achieve high localization precision at low computation cost. Our experimental results indicate that MIL outperforms the state-of-the-art mobile indoor localization schemes in terms of localization precision.