Inner-outer feedback linearization for quadrotor control: two-step design and validation

This paper proposes a novel control architecture for quadrotors that relies twice on the Feedback Linearization technique. The solution comprises a tracking inner-loop resulting from applying the mentioned method to the attitude and altitude dynamics. The horizontal movement, and, thereby, the zero dynamics, are stabilized without linearizing nor simplifying it by resorting to the same nonlinear technique. Linear quadratic controllers with integral action are implemented to the resulting chain of integrators of the inner and outer loops. As a result, the inner-loop dynamics asymptotically track the desired attitude and altitude over a broad region of the state-space, and the outer-loop yields a tracking system that is input-to-state stable and exponentially stable in the absence of external inputs. The stability of the proposed inner-outer loop control architecture is studied, leading to the proof of asymptotic stability in an extensive region of the state-space. Trajectory tracking, the capacity to overcome significant deviations on the mass and inertia values, and the robustness to external disturbances are evaluated using a simulation model, in which measurement noise and saturation limits are considered. In addition, comparisons regarding the performance in trajectory tracking of the proposed strategy and the results obtained with similar solutions from the literature are established. Experimental tests were conducted using a commercially available drone, equipped with an Inertial Measurement Unit, a compass, and an altimeter. A motion capture system gives the inertial position of the drone. The results obtained allow the validation of the modeling and control system solution.