Multimodal Indoor Localization: An Audio-Wireless-Based Approach

Location-based services on mobile devices have become a key element in today's wireless and mobile phone infrastructure, which due to their potential for precise personalization offer interesting opportunities for Semantic Computing. However, location information is mostly only available outdoors and current indoor localization schemes are not very accurate. In this paper, we therefore present a novel approach for indoor localization using multiple modalities of information that are easily available indoors on handheld devices. We use the microphones plus the various wireless signals that are sensed by smartphones to serve as input for a novel localization approach. Our proposed approach is computationally lightweight and, by making use of recent machine learning techniques for integrating modalities, achieves greater accuracy than current work in the area.