Neural network control of quadrotor UAV formations

Abstract

In this paper, a novel framework for leader-follower formation control is developed for the control of multiple quadrotor unmanned aerial vehicles (UAVs) based on spherical coordinates. The control objective for the follower UAV is to track its leader at a desired-separation, angle of incidence, and a bearing by using an auxiliary velocity control. Then, a novel neural network (NN) control law for the dynamical system is introduced to learn the complete dynamics of the UAV including unmodeled dynamics like aerodynamic friction. Additionally, the interconnection dynamic errors between the leader and its followers are explicitly considered, and the stability of the entire formation is demonstrated using Lyapunov theory. Numerical results verify the theoretical conjectures.