Modeling Approaches for Multirotor UAV Dynamics: A Systematic Review of Methods, Fidelity Levels, and Validation Practices

Abstract

The academic landscape of UAV research has seen an exponential growth in modeling methodologies, resulting in a fragmented ecosystem of techniques varying in mathematical rigor and application focus. This systematic literature review (SLR) provides a comprehensive synthesis of the state-of-the-art in multirotor UAV dynamics, following PRISMA guidelines to analyze over 200 core studies. The review categorizes prevailing approaches into four primary clusters: first-principles analytical models (Newton-Euler), data-driven black-box models (neural networks), hybrid grey-box models, and high-fidelity CFD-informed models. Each cluster is evaluated based on its “Fidelity Level,” computational overhead, and suitability for tasks such as controller synthesis or failure simulation. A significant portion of this review is dedicated to the critical analysis of validation practices. We examine the evolution of experimental setups, from indoor motion capture systems to modern outdoor telemetry. Our findings reveal a concerning lack of standardization in validation metrics, with many studies failing to quantify the residual error between the model and the physical plant. To address this, the paper proposes a new “Model Fidelity Index” (MFI), a multi-dimensional metric accounting for predictive accuracy across the flight envelope, robustness to parameter uncertainty, and real-time execution capability. By identifying current gaps—particularly the need for better modeling of multi-body interactions and flexible structures—this SLR provides a strategic roadmap for future research, serving as an essential reference for both novice and experienced practitioners in aerial robotics.