Tracking Vehicles with Autonomous Flying Robots

Micro aerial vehicles (MAVs) have become a major trend in robotics research. While controllers of MAVs already show an amazing performance, most systems lack autonomy and depend on global localization based on external sensors. Our goal is to eliminate this dependency and to enable MAVs to navigate through the environment using only onboard hardware. We developed a method for pose estimation based on a pattern of four orange table tennis balls, which has moderate computational cost and runs completely on-board [4].

Quadrocopter following an Outdoor Buggy
Fig. 1: The quadrocopter autonomously following the outdoor buggy. The system works even in difficult lighting conditions

We examined our method for self-localization and evaluated its use for controlling the pose of MAVs [3] and a Follow- the-Leader scenario [2]. We added several outdoor experiments to examine the robustness of the methods in the presence of different lighting conditions. In a further experiment we examine how a quadrotor can fly autonomously using the method [1]. Currently the MAV is able to take-off, follow and land autono- mously on a robot outside in the presence of sunlight and shadows.

Markers

Hardware Platforms

AscTec Hummingbird
  • 3-axis gyroscope
  • 3-axis accelleration sensor
  • 3-axis magnetometer
  • PointGrey FireFly Color Camera
  • Custom built Pan-/Tilt-Unit
  • Gumstix Overo Fire Single-board-computer with 800 MHz ARM CPU

Outdoor Buggy
  • Based on E-Maxx model truck
  • Equipped with markers and landing pad
  • Speeds up to 15 km/h
Fig. 2: The pattern recognized within the image from the on-board camera

Video

This demonstration video shows autonomous flights using the proposed method: Video

Publications

[1] Andreas Masselli and Andreas Zell. A new method for solving the perspective-three-point problem. In International Conference on Pattern Recognition (ICPR 2014), Stockholm, Sweden, August 2014. Accepted for publication. [ details ]
[2] Andreas Masselli, Shaowu Yang, Karl Engelbert Wenzel, and Andreas Zell. A cross-platform comparison of visual marker based approaches for autonomous flight of quadrocopters. Journal of Intelligent & Robotic Systems, 73(1-4):349-359, 2014. [ DOI | details | pdf ]
[3] Jacobo Jimenez Lugo, Andreas Masselli, and Andreas Zell. Following a quadrotor with another quadrotor using computer vision. In European Conference on Mobile Robots (ECMR 2013), Barcelona, Catalonia, Spain, September 2013. [ details | pdf ]
[4] Andreas Masselli, Shaowu Yang, Karl Engelbert Wenzel, and Andreas Zell. A cross-platform comparison of visual marker based approaches for autonomous flight of quadrocopters. In Proceedings of International Conference on Unmanned Aircraft Systems, pages 1-9, Atlanta, Georgia, USA, May 2013. [ details | pdf ]
[5] Andreas Masselli and Andreas Zell. A Novel Marker Based Tracking Method for Position and Attitude Control of MAVs. In Proceedings of International Micro Air Vehicle Conference and Flight Competition, pages 1-6, Braunschweig, Germany, 2012. DGON. [ details | pdf ]

Contact

, Phone: (07071) 29 - , Email: