Autonomous Exploration and Mapping of Structured Spaces using a Small Unmanned Aircraft System (sUAS)
Adam Schultz and Maarten Uijt de Haag, Ohio University
In recent years, the number of potential military and, especially, commercial applications for small Unmanned Aircraft Systems (sUAS), a.k.a. drones, has significantly increased. Example applications include environmental monitoring, surveillance, mapping, agriculture, aerial photography, search and rescue, law enforcement, to name a few. In particular, multi-copter platforms have emerged as effective platforms for lower-altitude operations and could be very capable for cost-effective missions in remote or difficult locations for applications such as large building maintenance, search and rescue, indoor mapping, etc. These environments are typically challenging in terms of obstacles that must be avoided, trajectories that must be planned and available navigation capabilities. The increased complexity of these environments furthermore drives the requirement for increased autonomy as manual decisions on navigation, collision avoidance and mission planning are too slow. In previous work, we developed a reliable navigation method for both outdoor and indoor structured environments that uses multiple platform laser scanners, an inertial measurement unit, barometric height and (where available) GNSS. In this paper, we have modified the navigation method, and have added a path planning, real-time mapping and collision avoidance function so it can be used to autonomously explore structured environments including indoor environments, urban environments and forest environments.
A data collection platform has been developed using a Blackout quad-copter racing frame equipped with laser scanners, a camera, GNSS and inertial sensors to observe the environment and as an input to the navigation methods. All navigation, obstacle avoidance, and flight planning algorithms have been implemented in the Robotic Operating Systems (ROS) on an onboard-embedded processor and integrated with a ArduPilot flight controller for autonomous flight.
This paper will discuss the various aspects of autonomy on a small-size multi-copter sUAS for the challenging environments identified, as well as address in detail the additional path planning, real-time mapping and collision avoidance modules, its integration with the flight controller for autonomous flight and the actual implementation on the multi-copter platform. Furthermore, the paper and presentation will include flight test results and videos of our multi-copter UAS operating in the Russ Research Center, RRC (indoor industrial facility with various room sizes), in parts of the city of Athens (“urban”), and in a forest environment (near RRC).