Sistema de posicionament mitjançant balises LED infraroges
Tutor / director / evaluatorCasas Piedrafita, Óscar
Document typeBachelor thesis
Rights accessRestricted access - author's decision
Currently most of the civilian available unmanned aerial vehicles (UAVs) use only satellite systems (GPS, GLONASS and BeiDou) to obtain its position. It's a good solution for open spaces but not for indoors because the building walls block the satellite signals. The aim of this project is to create a positional system for indoors and the needed tools to show the results in real time (web interface). It has to require low processing power to be able to run in embedded platforms. The system has to have an update rate equal or higher than the satellite systems in order to warrant a seamless navigation. The precision also has to be superior, in the order of few centimeters because in closed spaces distances are critical. The chosen solution uses groups of infrared LED beacons (waypoints) spread on the ground as a reference for the positioning. Each group has a characteristic LED distribution in order to be differentiable from the rest. The autonomous vehicles detect them using a camera that captures infrared light and blocks the visible light. Then the system uses the orientation sensors to obtain the pitch/yaw of the vehicle in order to be able to project each beacon to a horizontal plane. In this way it's not necessary to keep the camera perpendicularly pointing to the ground. Once the system obtains the 2D projection of the beacons, it organizes them in different groups, it identifies them using stored patterns and finally it obtains the position and the heading. The obtained results are quite satisfactory, the precision and the update rate are inside the desired margins and the computational cost of the algorithms is low enough to be able to run in the chosen platform. Nonetheless in some situations the system error can exceed the desired levels.