A Low-cost Neuroadaptive Control Approach for Unmanned Aerial Vehicle under Time-Varying Asymmetric Motion Constraints

Abstract

This paper presents a neuroadaptive tracking control scheme for uncertain Unmanned Aerial Vehicle (UAV) subject to asymmetric yet time-varying (ATV) full-state constraints without involving feasibility conditions. By blending a nonlinear state-dependent transformation into each step of backstepping design, a neural network-based adaptive control scheme is developed, which, as compared with most existing methods, exhibits several attractive features: 1) it is robust and adaptive to parametric/non-parametric uncertainties; 2) it not only directly accommodates ATV motion (position and velocity) constraints but also removes the feasibility conditions on virtual controllers; and 3) it only involves one lumped-parameter adaptation, thus is structurally simpler, computationally less expensive, and easier in implementation. Neural network (NN) unit accounting for system uncertainties is included in the loop during the entire system operational envelope in which the precondition on the NN training inputs is always ensured. The effectiveness of the proposed control strategy for UAV is confirmed by systematic stability analysis and numerical simulation.