Mitigating the antenna orientation effect on indoor Wi-Fi positioning of mobile phones

Due to the limitation of GPS in indoor environment and the rapid growth of Wi-Fi hotspots and mobile devices, indoor Wi-Fi-based positioning has been attracting growing interest. In this paper, we implement a practical and convenient indoor positioning system based on the fingerprint method and Kalman filter on Android mobile devices. This paper not only discusses the positioning algorithms, but also addresses various challenges in practical application, such as the effect of antenna orientation and signal fluctuation. Specifically, an improved mapping algorithm based on k-nearest neighbors (K-NN) is introduced to tackle the orientation effect, and an orientation-based fingerprint database is established through studying the received signal strength patterns in different directions to handle the large fluctuation caused by orientation change. Finally, our experimental result indicates that the proposed indoor positioning system can achieve up to 1.2 meters accuracy in 90 percent of time, which is sufficient for supporting various navigation and infotainment services in large-scale indoor environments (e.g., shopping malls).