A Sensor System for Autonomous UAV Landing

Abstract

Unmanned Aerial Vehicles (UAVs) have recently surged in popularity and are now seen as viable tools for use in commercial delivery, search and rescue operations, and planetary exploration. A challenge in utilizing UAVs for these purposes is the limited battery capacity: an average-sized drone has 25 minutes of flight time, which results in limited flight range and payload capacity. Implementing an array of charging stations would extend UAV flight range, but GPS navigation would not provide the resolution necessary to reliably find and land on these stations. This project focuses on developing a vision-based sensor system that would provide onboard sensing mechanisms and path planning in order to accurately detect and land on a charging pad. This solution aims to be lightweight, unobtrusive, self-contained, and more accurate than relying on GPS and IMU alone. This proof of concept sensor system can be adapted to the form factor of mission-specific UAVs.