Conference Publication
Resources:
- Paper link: https://doi.org/10.1007/978-3-319-03653-3_5
- Bibtex citation: Bibtex
Abstract:
This paper presents a low-cost framework for visual quadrotor swarm prototyping which will be utilized to participate in the 2013 International Micro Air Vehicle Indoor Flight Competition. The testbed facilitates the swarm design problem by utilizing a cost-efficient quadrotor platform, the Parrot AR Drone 2.0; by using markers to simplify the visual localization problem, and by broadcoasting the estimated location of the swarm members to obviate the partner dectection problem. The development team can then focus their attention on the design of a succesful swarming behaviour for the problem at hand. ArUco Codes [2] are used to sense and map obstacles and to improve the pose estimation based on the IMU data and optical flow by means of an Extended Kalman Filter localization and mapping method. A free-collision trajectory for each drone is generated by using a combination of well-known trajectory planning algorithms: probabilistic road maps, the potential field map algorithm and the A-Star algorithm. The control loop of each drone of the swarm is closed by a robust mid-level controller. A very modular design for integration within the Robot Operating System (ROS) [13] is proposed.