Design and Flight Demonstration of a Quadrotor for Urban Mapping and Target Tracking Research

TL;DR

The paper addresses urban mapping and target-tracking challenges for small UAVs by introducing a low-cost quadrotor platform equipped with five cameras, including two pairs of fisheye stereo cameras and a two-axis gimbal, all powered by an onboard NVIDIA Jetson Orin Nano and ROS. It details the integrated hardware (QAV500 V2 frame, Cube Orange, Here GPS, dual batteries) and software stack (MAVLink via MAVROS, ROS, rosbag) used to enable perception, tracking, and mapping, with Direct Sparse Odometry applied to captured imagery. Key contributions include a robust autonomous tracking capability toward GPS-tagged targets, extensive characterization of endurance ($ ext{≈}11$ min), range ($ ext{≈}3.8$ km), and top speed ($ ext{≈}20$ m/s), as well as acoustic and communication-range analyses relevant to urban operation. The demonstrated tracking flight and mapping experiments, plus the open hardware-and-software design, offer a practical, reproducible platform for researchers studying urban UAV perception, hazard avoidance, and real-time mapping and tracking.

Abstract

This paper describes the hardware design and flight demonstration of a small quadrotor with imaging sensors for urban mapping, hazard avoidance, and target tracking research. The vehicle is equipped with five cameras, including two pairs of fisheye stereo cameras that enable a nearly omnidirectional view and a two-axis gimbaled camera. An onboard NVIDIA Jetson Orin Nano computer running the Robot Operating System software is used for data collection. An autonomous tracking behavior was implemented to coordinate the motion of the quadrotor and gimbaled camera to track a moving GPS coordinate. The data collection system was demonstrated through a flight test that tracked a moving GPS-tagged vehicle through a series of roads and parking lots. A map of the environment was reconstructed from the collected images using the Direct Sparse Odometry (DSO) algorithm. The performance of the quadrotor was also characterized by acoustic noise, communication range, battery voltage in hover, and maximum speed tests.

Figures (10)

• Figure 1: Example gimbaled camera data collected by the UNC Charlotte QAV 500 Drone while tracking a gray car along a straight road and through a parking lot.
• Figure 2: The UNC Charlotte QAV500 drone is equipped with a Cube Orange flight controller and carries a NVIDIA Jetson Orin Nano, a gimbaled camera, and two pairs of stereo fisheye cameras.
• Figure 3: The front (top) and rear (bottom) stereo fisheye images during the tracking test. The horizontal aberrations are due to the presence of the propeller in the field of view.
• Figure 4: A diagram of the hardware connections and ROS nodes running on the NVIDIA Orin, the gray lines are software connections, the black lines are the hardware connections, and the dashed black line is a future hardware connection.
• Figure 5: Battery voltage while the UAV hovers in place. The blue line was a hover test with GPS hold and manual altitude control with a flight time of 11min 27s, while the orange line was a hover test with GPS hold and automatic altitude control with a flight time of 12min 24s. The low voltage cutoff was 21.6V corresponding to 3.6V per cell. The green line is the predicted flight time of 10min 18s.