Cowlagi, Raghvendra V.
Autonomous obstacle avoidance is a fundamental desirable feature of all unmanned aerial vehicles (UAVs). Due to severe payload weight restrictions, small UAVs are limited in the number and type of sensors that can be carried on-board for detecting and avoiding the environment. This project describes an obstacle avoidance system based on two lightweight off-the-shelf cameras, and a small Raspberry Pi microcontroller. The system was designed to receive telemetry and sensor data from the UAV’s basic autopilot and to return command guidance for to avoid obstacles in the desired path. To implement this system, algorithms for object tracking and depth mapping using monocular and stereo camera vision were adapted. Workbench tests and flight tests on the IRIS quadrotor UAV were performed.
Worcester Polytechnic Institute
Major Qualifying Project
All authors have granted to WPI a nonexclusive royalty-free license to distribute copies of the work, subject to other agreements. Copyright is held by the author or authors, with all rights reserved, unless otherwise noted.