Dynamic clock-model of Wi-Fi access-points to help indoors localisation of smartphones

Smartphones contain both Wi-Fi and GPS technologies. When outdoors, the GPS receiver onboard the Smartphone can define its location and provides accurate GPS time. This capability is lost when the Smartphone is indoors. In another vain, Wi-Fi Access Points (WAPs) can be used to locate Smartphones indoors, using Trilateration techniques. However, WAPs contains low accuracy clocks [1] that would produce large position errors when used to locate any Smartphone indoors [2]. By using GPS time, obtained from when the Smartphone was insight of GPS signals, the Smartphone can virtually synchronise in time with any available WAPs. This time delay/offset knowledge will then be used to determine the Smartphone's position indoors more accurately, and also to provide seamless outdoors-indoors positioning. To help develop localization hypothesis and simulate various scenarios, based on multiple Smartphones and WAPs, this paper proposes an accurate WAPs clock model that can be used for such simulations. The model uses GPS time as the reference clock, together with WAP clock offset and accurate clock drift model based on real noise factors such as white noise, flicker noise and random walk noise. Our clock model is designed to be dynamic to accommodate various WAPs hardware and their position. Simulation results, using MATLAB, show that our model achieves higher location accuracy of up to 41.33 nanosecond than existing approaches; resulting in average of 12.4 meter position error improvements. Furthermore, our clock synchronization experiments show that our WAP clock model achieves 11.8 nanosecond of clock time error delay improvement over any existing clock model.