A ROS driven platform for optimizing radiomap management in fingerprinting based indoor positioning

An electromagnetic beacons infrastructure is commonly used in posi- tioning applications within buildings where the GPS signal is not present. Through techniques of multilateration and fingerprinting an average accuracy of about 2 meters can be reached, but the accuracy is limited by multiple reflec- tions, obstacles and signal dispersion that make it unreliable analytic field mod- eling. Dense field sampling field allows a reconstruction more detailed but is costly and clever uneven sampling is appropriate. This work describes the progress of an interactive robotic platform under de- velopment at ICAR-CNR for supporting field modeling through intelligent iter- ative strategies of data acquisition and for optimization of the positioning of bea- cons. The platform integrates a MATLAB®-based control and simulation software, a robot equipped with distance sensors and able to perform autonomous naviga- tion in a known environment and a data logger module hosted in an Android mobile device, all connected via a ROS framework. The first module (developed in MATLAB®) enables the identification of measurement positions and the optimal settings for the robot's motion tasks (real- time collision avoidance, path planning, and mission planning). The second sys- tem (robot with Android devices) enables the interfacing with BLE beacons placed in the selected environment. Robot assisted field sampling is here proposed and used to reduce costs of radiomap construction and update. In particular, this technology is proposed for complex environments as museums and exhibitions.