Autonomous Landing of UAVs under Unknown Disturbances using NDI Autopilot with L1 Adaptive Augmentation

Abstract

This paper presents autonomous landing of Unmanned Aerial Vehicles (UAVs) under unknown external disturbances and internal plant parameter uncertainties. The external disturbances such as wind shear, wind gust and ground effects are considered. The plant parameters uncertainties are also considered due to aerodynamic force and moment coefficients random perturbation. L1 adaptive controller with piece-wise-constant adaptation law is augmented over Nonlinear Dynamic Inversion (NDI) autopilot and implemented on Six Degree of Freedom (Six-DOF) model of a UAV. The NDI autopilot with two time scale separation is designed for the nominal plant. It tracks a reference trajectory computed from a path planning and guidance algorithm under undisturbed plant model with normal environment conditions. The L1 adaptive controller takes into account the disturbances and computes the adaptive control command which enables the plant trajectory tracking closer to the desired reference trajectory. The state predictor is designed to track the plant states smoothly by incorporating proportional and integral error terms in the state predictor model. The piece-wise-constant adaptive law is designed to estimate the unknown disturbances. The unknown disturbance estimates are used to design the adaptive control law to nullify the effect of the disturbance on the plant performance. The algorithm is simulated to show the auto landing performance on a Six-DOF UAV model.