This study focuses on improving quadrotor motion control through physics-informed data-driven dynamic modeling. Traditional modeling approaches often fail to capture complex aerodynamic effects and external disturbances, leading to inaccuracies in control performance. To address these challenges, this research explores various data-driven modeling techniques, including analytical, gray-box, black-box, and physics-informed neural networks. These models are integrated into control strategies such as differential flatness-based control and reinforcement learning-based approaches. A ROS2-based framework is developed to facilitate simulation, real-time control, and evaluation. The results demonstrate that incorporating data-driven models significantly enhances quadrotor stability and trajectory

Бархум З. (науч. рук. Колюбин С.А.) Physics-informed data-driven dynamic modeling for quadrotor motion control // Сборник тезисов докладов конгресса молодых ученых. Электронное издание. – СПб: Университет ИТМО, [2025]. URL: https://kmu.itmo.ru/digests/article/15275