Abstract

The primary goal of our team for the Spring quarter was to build a prototype UAV that can survey a delimited area (either autonomously or controlled by a pilot), detect the target objects, perform sensor fusion and combine the readings to detect the precise positions of the targets. To achieve this, we assembled a prototype drone using off-the-shelf components, integrated it with a companion computer, implemented our algorithms on this unit and combined all the modules together to achieve good results. Throughout the quarter, we successfully deployed a working prototype, and tested our algorithms (finding some weaknesses which we were able to improve), ending up with a solid solution that can be adapted by the TritonAI team for their competition drone.

Video Presentation