Nonlinear Model Predictive Control of Single-Actuator Monocopters Using Hybrid Rotational Dynamics and INDI

Abstract

In this article, we present a trajectory tracking framework for samara seed-inspired single-actuator monocopters (SAM) by reformulating a nonlinear model predictive control (NMPC) architecture that leverages a hybrid rotational dynamic model and is integrated with incremental nonlinear dynamic inversion (INDI). Unlike conventional approaches for the SAM that decouple disk and body rotational dynamics, we formulate a hybrid representation that incorporates both disk and body angular rates for the NMPC scheme. Our architecture enables previously unachievable levels of linear velocity and robustness on the SAM, enabling it to execute aggressive maneuvers for dynamically infeasible trajectories. Finally, extensive flight trials on SAM variants (short, long, and ultralight-wing) were conducted on different trajectories under nominal and perturbed conditions, with NMPC+INDI outperforming a differential flatness-based controller with INDI in empirical benchmarking. The results showed NMPC+INDI reduced Euclidean positional root-mean-square-error (RMSE) by up to 39.5% on the long-wing and 37.2% on the short-wing, along with significant velocity RMSE reductions as high as 43.5% on the ultralight-wing. These significantly reduced error margins from our proposed NMPC+INDI method not only validate the hybrid model but also demonstrate the approach's effectiveness and scalability, as shown by experiments on SAMs of diverse form factors.