Infrastructure-free indoor navigation based on smartphone sensors in smart buildings

Abstract

One of the main elements of location-based services (LBS) is the awareness and knowledge of the user’s location information inside the smart buildings. In this study, a smartphone sensor-based indoor positioning system (IPS) is proposed to track a person’s location in Texting and Pocket carrying modes in a smart building. The gravity, gyroscope, and magnetometer sensors data were combined using a gradient descent algorithm (GDA) to estimate the heading angle. This system was implemented in three straight, complex, and rectangular paths. The mean (M) and standard deviation (SD) of the absolute heading error of each step were obtained as (1.68°, 1.97°) in the Texting mode and (4.39°, 5.22°) in the Pocket mode, respectively. Acceleration and angle-based models were employed to estimate the step length in the Texting and Pocket modes, respectively. The mean relative error (MRE) of the distance in the Texting and Pocket modes were obtained as %4.8 and %4.37, respectively. Experimental results indicated the MRE of the final position along the three paths in the two carrying modes of Texting and Pocket by magnetic and proposed methods reduced from %3.75 to %2.66 and %7.02 to %4.24, respectively.