Affine Formation Control of Multiple Quadcopters

This paper considers the distributed time-varying formation tracking control problem of multi-quadcopter systems using affine formation control strategies with multiple virtual leaders. A novel two-layer (formation layer and local control layer) affine formation control structure is established to account for the underactuated nature of the quadcopter dynamics. In the formation layer, a quadcopter is abstracted as a virtual double-integrator agent and affine formation controllers are then designed based on the networked double-integrator dynamics. The resultant virtual affine formation control inputs from the formation layer are converted to the desired attitudes based on the quadcopter dynamics, and a sliding mode controller is then proposed to ensure flnite-time tracking convergence to the desired attitude in the local control layer. Numerical simulations were carried out using a group of six quadcopters in the XY-plane to demonstrate and validate the effectiveness of the developed controllers.

I. Introduction

Quadcopters are mechanically simple, highly maneuverable and capable of vertically taking off and landing. In recent years, multiple-quadcopter systems have been applied in various fields to accomplish tasks, such as forest fire monitoring and load transportation, that would be impossible or inefficient for a single qaudcopter system. Formation control design, where the entire group is driven to achieve and maintain a certain geometric pattern, is often required to enable these applications. Although some formation control approaches, such as virtual structure and behavior-based, are effective and easy to implement for many systems with complex dynamics, the overall system stability analysis is mathematically challenging and the convergence is often not guaranteed. Recently, the consensus-based [1] formation control methods that can guarantee the overall system convergence have prevailed.