Adaptive Formation Control for Multiple Quadrotors with Nonlinear Uncertainties Using Lipschitz Neural Network

This paper proposes a new distributed formation control scheme for multiple quadrotors with uncertainties, which is more practical for real implementation. The proposed design takes into account the inherent nonlinear and coupled dynamics of quadrotor and incorporates the Lipschitz model uncertainty. The multi-quadrotor formation control system is divided into the high-level distributed formation controller for position control, and the lowlevel attitude controller for desired acceleration tracking. To improve the performance of the lowlevel system, we employ the Lipschitz neural network (LipNet) adaptation. LipNet adaptation learns the model uncertainty and provides a significant improvement in acceleration tracking by modifying the reference signal of the low-level system. Consequently, the performance of the whole system is enhanced. We provide a stability analysis of the proposed design and validate the performance by some simulation examples. Copyright (c) 2023 The Authors.