Indoors Smartphone Positioning Enhancement Using Wi-Fi and Magnetometer

Abstract

Recent embedded technologies on today’s smartphone make the smartphones more comfortable to run a large number of applications which are used for people’s daily activities. Among all these applications, the location-based services (LBS) are frequently used. The LBS applications utilized location information via Smartphones technologies. For example, when outdoors, Global Navigation System (GPS) or generally Global navigation satellite system (GNSS) signals are used to retrieve the location information within enough positioning accuracy. However, when smartphone holders are entering to the urbane area or indoors, the performance of GPS service will be degraded or sometime cannot retrieve location information due to blocking the GPS signals through the roofs or walls of the buildings. Beside this, many onboard smartphones wireless chipsets or sensors’ readings can be used as alternate technologies to provide location information including Wi-Fi, Bluetooth, cellular, and inertial sensors. However, these technologies during positioning process will faced its own limitations such as: none-line-of-sight signals, multipath signals, and sensor drift or accumulated error. For these reasons, it is very difficult to provide a good positioning accuracy, when only a single technology is utilized alone. Therefore, this study proposes a new positioning solution based on hybridize two different technologies measurements including received signal strength (RSS) of the Wi-Fi access points and onboard smartphone magnetometer readings within fingerprinting positioning technique. The hybridization of these technologies is based on taking their advantaged and mitigating their drawbacks. In addition to that, this study also provided an improved version of matching algorithm of the fingerprinting technique by applying the concept of boosting-dataset records. A set of real trial experiments are conducted to prove the validity of the proposed solution. The obtained results of the experiments show that the proposed positioning solution can provide an enough positioning accuracy, up to 0.13 meter.