Autonomous Localization and Landing of MAVs
Burnaby South Secondary
Floor Location : S 015 N
Multi-rotor UAVs are taking on a growing role in society. They have great appeal in fields like search and rescue or firefighting because their high maneuverability allows them to navigate in tight environments. However, in order to achieve optimal efficiency in these applications, they must lose their dependency on humans, i.e. become fully autonomous. The goal of this project was to design a new type of positioning system for the autonomous landing of multi-rotors, which is an essential component to the sought after full-automation of drones. Optimal solutions to the problem should minimize power-draw and payload, and offer potential for down-scaling in size, all while maintaining a practical level of accuracy. Most prior solutions require an impractical level of computing power due to their utilization of cameras with computationally expensive image processing. Drawing inspiration from the basic visual systems of phototactic organisms, a new type of optical targeting system was devised. Following geometrical analysis and derivation of equations for determining position relative to a given light source, a sensor module comprising only 4 photo-sensitive devices--as opposed to the many millions present in a camera sensor--was constructed. Early testing shows promising results, with the system exhibiting an acceptable level of accuracy and speed.