Outdoor flocking of quadcopter drones with decentralized model predictive control

Quan Yuan; Jingyuan Zhan; Xiang Li

doi:10.1016/j.isatra.2017.07.005

Outdoor flocking of quadcopter drones with decentralized model predictive control

ISA Trans. 2017 Nov;71(Pt 1):84-92. doi: 10.1016/j.isatra.2017.07.005. Epub 2017 Jul 11.

Authors

Quan Yuan¹, Jingyuan Zhan², Xiang Li³

Affiliations

¹ Adaptive Networks and Control Lab, Department of Electronics Engineering, Fudan University, Shanghai 200433, China; Research Center of Smart Networks and Systems, School of Information Science and Engineering, Fudan University, Shanghai 200433, China. Electronic address: qyuan15@fudan.edu.cn.
² Adaptive Networks and Control Lab, Department of Electronics Engineering, Fudan University, Shanghai 200433, China; Research Center of Smart Networks and Systems, School of Information Science and Engineering, Fudan University, Shanghai 200433, China. Electronic address: 13110720042@fudan.edu.cn.
³ Adaptive Networks and Control Lab, Department of Electronics Engineering, Fudan University, Shanghai 200433, China; Research Center of Smart Networks and Systems, School of Information Science and Engineering, Fudan University, Shanghai 200433, China. Electronic address: lix@fudan.edu.cn.

PMID: 28709654
DOI: 10.1016/j.isatra.2017.07.005

Abstract

In this paper, we present a multi-drone system featured with a decentralized model predictive control (DMPC) flocking algorithm. The drones gather localized information from neighbors and update their velocities using the DMPC flocking algorithm. In the multi-drone system, data packages are transmitted through XBee^® wireless modules in broadcast mode, yielding such an anonymous and decentralized system where all the calculations and controls are completed on an onboard minicomputer of each drone. Each drone is a double-layered agent system with the coordination layer running multi-drone flocking algorithms and the flight control layer navigating the drone, and the final formation of the flock relies on both the communication range and the desired inter-drone distance. We give both numerical simulations and field tests with a flock of five drones, showing that the DMPC flocking algorithm performs well on the presented multi-drone system in both the convergence rate and the ability of tracking a desired path.

Keywords: Decentralized model predictive control; Flocking; Multi-agent system; Quadcopter drone.