A Real-time Robust Indoor Tracking System in Smartphones

Nowadays, a growing number of ubiquitous mobile applications has increased the interest in indoor location-based services. Some indoor localization solutions for smartphones exploit radio information or data from Inertial Measurement Units (IMUs), which are embedded in most of the modern smartphones. In this work, we propose to fuse WiFi Receiving Signal Strength Indicator (RSSI) readings, IMUs, and floor plan information in an enhanced particle filter to achieve high accuracy and stable performance in the tracking process. Compared to our previous work, the improved stochastic model for location estimation is formulated in a discretized graph-based representation of the indoor environment. Additionally, we propose an efficient filtering approach for improving the IMU
measurements, which is able to mitigate errors caused by inaccurate off-the-shelf IMUs and magnetic field disturbances. Moreover, we also provide a simple and efficient solution for localization failures like the kidnapped-robot problem. The tracking algorithms are designed in a terminal-based system, which consists of commercial smartphones and WiFi access points. We evaluate our system in a complex indoor environment. Results show that our tracking approach can automatically recover from localization failures, and it could achieve the average tracking error of 1.15 meters and a 90% accuracy of 1.8 meters.