In this paper, signal strengths from known WiFi access points are used together with a particle filter to perform indoor navigation. It is shown that more information is obtained by using signals of both 2.4 and 5.0 GHz, compared to using only one frequency. Thus, using both frequencies provides a more accurate positioning. The second contribution is an algorithm where WiFi measurements are combined with pedestrian dead reckoning (PDR), which is based on step counting using an accelerometer and hypotheses of the heading using a gyroscope. This was found to provide further accuracy compared to more conventional methods.
展开▼
