Robust Trajectory Tracking of a Quadrotor via Cascade Barrier-Function Super Twisting Control

In multiloop feedback control (MLFC) structure, the appropriate determination of loop gains is critical for achieving satisfactory performance. This work introduces an enhanced MLFC framework equipped with automatic gain tuning for the robust trajectory tracking of quadrotor systems. Key advantages include the application of time-scale separation and singular perturbation techniques (SPTs) to facilitate model decomposition and establish relationships between loop gains. First, the six-degree-of-freedom nonlinear model is decomposed into two singularly perturbed submodels, each representing translational and rotational dynamics operating on distinct time scales. On this basis, the MLFC is designed to enable the quadrotor to track a predefined trajectory in the presence of matched uncertainties. For the inner loops, a barrier-function adaptive super-twisting controller is employed at the fastest time scale to enhance robustness against system uncertainties. Controller gains are set as the functions of the reciprocal of the singular perturbation parameter (SPP). For the outer loops, the slow–fast decomposition is performed through SPP-dependent proportional–integral–retarded (PIR) controllers, which manage positions and linear velocities in two separate time scales. Tracking bandwidths are achieved using the pole-placement method. The effectiveness of the proposed methodology is validated through experimental testing on a quadrotor platform, focusing on trajectory tracking accuracy and robustness subject to system uncertainties.