Indoor navigation system using image and sensor data processing on a smartphone

The navigation and mobile positioning paradigm has been an active research topic in the past two decades. The proliferation of mobile devices equipped with inertial sensors able to detect phone motion lead to an increase in research and product development in the field of personal locating technologies. Within this scenario, this paper proposes a combined technique for the estimation of user's position in indoor environments, which only relies on sensors of a commercial smartphone, without the active aid of external infrastructures. The smartphone's video camera is used to identify known keypoints, named anchors previously identified and geo-referenced in the building map. A SURF algorithm is employed for quick and effective detection of image features independently from the viewpoint. The accelerometers and the compass of the smartphone are used to estimate the amount of steps and the direction in which the user is moving by performing statistical data processing. Heading data is compensated taking into account the relative position of the smartphone with respect to the user. Experiments have been carried out in a real scenario showing the effectiveness of the proposed solutions.