Cooperative Control and RTK Navigation System for Multirotors

Abstract

This thesis considers the implementation of cooperative control on small, unmanned, multirotor systems. More specifically, the problem of distributed formation control is handled, as well as the necessary high precision navigation. The control of multiple Unmanned Aerial Vehicles (UAV) has seen a lot of development in the last years, but most advanced implementations use central processing and rely on motion capture systems for precision navigation, or simply cope with the low accuracy of a Global Navigation Satellite System (GNSS). This thesis uses an established passivity-based framework to achieve decentralized formation control, and investigates the use of Real-Time Kinematics (RTK) to augment the accuracy of a GNSS to the centimetre level. The development of a suitable multirotor platform is presented, and control is implemented with the combination of the open-source APM:Copter platform on a commercial autopilot and the DUNE: Uniform Navigational Environment on a payload computer. The resulting distributed system of UAVs is capable of fulfilling a flexible group objective while sustaining a desired formation, which is demonstrated in both simulations and field experiments.