Autonomous Drone Charging System

Summary

With the advent of autonomous drones, and their growth into a multi-billion dollar market, there is no end in sight to what they can accomplish. However, a severe drawback to these flying vehicles is their limited flight time. Battery technologies have been improving, but any battery still needs to be charged. An autonomous drone is not truly autonomous if a human operator needs to change or charge its batteries.

This limitation was the drive behind the development of the Autonomous Drone Charging System by myself and my Capstone team. We created an end-to-end solution—from compute stack and software architecture to vision system and charging hardware—to allow for autonomous recharging of the drone.

My part included design of the local path planning for precision landing, as well as the rail design for the final mechanical alignment with the charging contacts. I developed a ROS node to control the navigation of the drone in C++. Its responsibilities included the high level state management, as well as generating polynomial paths limited by constant accelerations. It integrated with the vision system to plan based on monocular camera capture of a unique marker on the landing pad that encoded the drone's XYZ displacement, as well as yaw error.

Awards

• First place out of all Electrical, Computer and Biomedical Engineering Capstone projects

Gallery