Indoor infrastructure-less solution based on sensor augmented smartphone for pedestrian localisation

The paper deals with the design, the calibration and experimental validation of a novel infrastructure-less solution dedicated to indoor pedestrian localisation issues. The approach involves aerodynamic fluid computation for instantaneous speed estimation of a pedestrian handling a smartphone. For this purpose, a differential hot-wire MEMS anemometer is integrated to an Android smartphone by means of a dedicated ARM 32 bits controller. Pedestrian orientation measurements are ensured by a gyroscope sensor coupled with the smartphone. Consequently, both instantaneous speed and direction measurements are combined to the dead reckoning technique for estimating the 2D relative position of the user. Theoretical modeling is conducted in order to calibrate and quantify the accuracy of the sensor. In situ experiments demonstrate that the sensors coupled with a smartphone achieve pedestrian localisation with average accuracy of 2 to 4 meters for 7 to 49m long trajectories.