Translational Model Identification and Robust Control for the Parrot Mambo UAS Multicopter

RUBIO SCOLA, IGNACIO; JOAO HESPANHA; GABRIEL ALEXIS GUIJARRO REYES; JUNFEI XIE; LUIS RODOLFO GARCIA CARRILLO

Waikoloa

Workshop; 2019 IEEE Globecom Workshops (GC Wkshps); 2019

IEEE

Two main problems are addressed in this paper.The first one is the model identification for a commercialunmanned aircraft system (UAS): the Parrot Mambo multi-copter. The second one aims at synthesizing a robust controllerfor guaranteeing the stability of the X and Y translationaldynamics of this UAS. To accomplish these goals, we first collectinput-output data from a set of real-time flight experiments.Next, by applying an extended least square (ELS) algorithmto the data, a group of dynamic models are identified. Dueto uncertainties, the obtained models are similar in naturebut exhibit parametrical variations. For this reason, fromthe set of identified models, a unique nominal (i.e., average)parameter-dependent linear model is built, which also takesinto account the minimum and maximum values defining themodel parametrical variations. Finally, a static linear controlleris synthesized for the dynamics of interest, guaranteeing globalstability for every model. The identification results and theperformance of the closed-loop controller are validated in a setof numerical simulations, demonstrating the effectiveness of theproposed modeling and control approaches.