Dynamic Regressor Extension and Mixing Parameter Estimators: A Comparative Simulation Study of the Quadrotor System

This paper presents a comparative numerical simulation study of three dynamic regressor extension and mixing estimators applied to the quadrotor aircraft system. These estimators are particularly relevant in the context of quadrotor dynamic models due to their ability to handle these vehicles’ complex and nonlinear dynamics and their significantly strong asymptotic stability properties. Namely, they ensure exponential convergence with the weakest assumption of interval excitation. Our main objective is to compare the transient performance of these three parameter estimators, a topic that plays a significant role in parameter estimation tasks. Our study goes beyond verifying theoretically validated (asymptotic) properties. We delve into aspects not previously explored, such as sensitivity to tuning parameters, focusing specifically on a quadrotor dynamic model. The ultimate goal of this study is to provide practical design guidelines for the unmanned aerial vehicles community interested in utilizing these estimators.