WiFi based indoor localization with adaptive motion model using smartphone motion sensors

We present an adaptive motion model for tracking the movement of smartphone user by using the motion sensors (accelerometer, gyroscope and magnetometer) embedded in the smartphone. A particle filter based estimator is used to seamlessly fuse the adaptive motion model with a WiFi based indoor localization system. The system applies Gaussian process regression to train the collected WiFi received signal strength (RSS) dataset, and particle filter for the estimation of the smartphone user's location and movement. Simulations were conducted in MATLAB to provide more insights of the proposed approach. The experiments carried out with an iOS device in typical library environment illustrate that our system is an accurate, real-time, highly integrated system.