flight dynamics simulation

This work addresses the challenges commonly encountered in multirotor flight control research—namely, the complexity of simulation-to-hardware workflows, poor code readability, and limited extensibility—by presenting a lightweight, modular, full-stack open-source flight control system built on ROS 2 and ROSflight 2.0. The proposed architecture enables seamless deployment between simulation and real hardware while significantly simplifying code structure without compromising performance. Its high modularity enhances both readability and extensibility, facilitating rapid research iteration. Experimental results demonstrate that the system achieves waypoint tracking performance comparable to state-of-the-art flight controllers, yet with a smaller, cleaner codebase that streamlines development and validation in academic settings.

We present a modular, high-fidelity simulation framework for the development and benchmarking of flight control strategies in fixed-wing aerial robots operating near the ground. Unlike existing simulators that rely on simplified or hover-oriented dynamics, our framework models full 6DoF rigid-body physics, semi-empirical ground-effect aerodynamics, actuator dynamics, sensor noise, and environmental disturbances. This physical realism, combined with modular component design, enables systematic analysis of low-altitude flight behavior under realistic conditions. The simulator supports both CPU and GPU backends via Torch and NVIDIA Warp, enabling high-throughput parallel execution suitable for large-scale reinforcement learning training and optimal control rollouts. A unified interface accommodates a range of controllers (both RL and optical control algorithms) across tasks such as altitude regulation and trajectory tracking. Cross-validation with X-Plane and JSBSim is also supported to facilitate engineering integration and visual fidelity.

本文通过动态反馈线性化方法，使多旋翼无人机模拟固定翼飞行器的动态特性，以解决固定翼飞机训练和仿真中的操作限制问题。

Ensuring simultaneous obstacle avoidance and geofence compliance for fixed-wing UAVs during flight poses significant safety challenges under nonlinear kinematic constraints. Method: This paper proposes a real-time assurance (RTA) framework based on control barrier functions (CBFs), specifically designed for nonlinear kinematic models of fixed-wing UAVs. We systematically formulate and compare multiple CBF variants to jointly enforce collision avoidance and geofence constraints in a unified, formal safety guarantee. The architecture operates at the command layer, dynamically modifying control inputs to ensure closed-loop safety under both constraints. Contribution/Results: We provide rigorous theoretical proofs establishing formal safety guarantees. Extensive validation—across both kinematic and high-fidelity dynamical simulations—demonstrates zero constraint violations and zero collisions. The approach significantly enhances the verifiable safety assurance capability of fixed-wing platforms operating in complex, constrained airspace.

To address the instability of reinforcement learning policies during zero-shot sim-to-real transfer of quadrotor control—caused by simulation-to-reality discrepancies—this paper introduces SimpleFlight, the first integrated PPO training framework specifically designed for zero-shot deployment on real hardware. The framework systematically identifies and jointly optimizes five critical factors: domain randomization, observation normalization, reward shaping, action smoothing, and dynamics-aware state encoding. Evaluated on the Crazyflie nano-quadrotor platform, SimpleFlight achieves stable trajectory tracking without any fine-tuning, reducing tracking error on jagged trajectories by over 50%. To foster reproducibility and community advancement, we open-source the complete implementation, pre-trained models, and Omnidrones—a high-fidelity, GPU-accelerated quadrotor simulator optimized for rapid RL training.

This work addresses the challenge of coordinate-free inverse flight dynamics modeling for fixed-wing aircraft, particularly the difficult mapping from trajectory to control inputs in tethered flight. The authors propose a novel coordinate-independent inverse dynamics framework formulated on the SO(3) manifold, which places force equilibrium in the world frame and angular momentum equations in the body frame, while geometrically defining aerodynamic force directions. Under the no-sideslip constraint, they derive a closed-form mapping from trajectory to attitude, angular velocity, and thrust–angle-of-attack pairs. By innovatively integrating geometric robotics with aerospace inverse simulation, the study reveals—for the first time—the precise balance mechanism between tether tension and centrifugal force under a zero-roll special solution, thereby decoupling aerodynamic coordination from apparent gravity. Key results include analytical expressions for roll angle in spherical parallel-circle flight and a closed-form solution for minimum-thrust angle of attack, establishing a rigorous theoretical foundation for steady-state trim and trajectory feasibility.

This study addresses the challenge of accurately modeling flexible cables in multi-UAV slung-load systems by departing from the conventional massless rigid-link assumption. A high-fidelity discrete-link cable model with optimizable parameters is proposed, formulated using Euler-Newton dynamics and efficiently simulated through the integration of Featherstone’s rigid-body algorithm with parameter sweeping techniques. Experimental validation against real-world flight data demonstrates that payload translational errors remain below 132 mm and attitude errors are less than 11.4°. Furthermore, the empirically collected dataset has been made publicly available as open source. This work provides a highly reliable research benchmark for advancing the modeling and control of cooperative multi-UAV aerial transportation systems.

This study addresses the low efficiency and high risk associated with advanced aerobatic flight training by introducing, for the first time, a reinforcement learning–based approach to advanced jet trainer systems. The authors develop a high-fidelity, interactive AI instructor module that integrates flight simulation, aerodynamic modeling, and intelligent agent reinforcement learning to accurately replicate a variety of complex aerobatic maneuvers. This integrated framework significantly enhances both the safety and precision of pilot training, offering an efficient and intelligent auxiliary tool for trainees. By doing so, the work bridges a critical gap in the application of artificial intelligence to aerobatic flight instruction, demonstrating the potential of data-driven, adaptive learning systems in high-stakes aviation environments.

This study addresses the challenge that conventional fixed-wing control strategies struggle to accommodate dynamic configuration transitions during rotor hover mode for biplane morphing unmanned aerial vehicles (UAVs). Taking the MetaMorpher platform as the research subject and building upon flight dynamics modeling, this work proposes an innovative flap phase-synchronized propulsion strategy as a lightweight and decoupled alternative to direct motor control. By integrating numerical simulations with vertical dynamics experiments, a closed-loop validation pipeline spanning conceptual design, algorithm development, and physical prototyping is established. The results demonstrate that the proposed strategy achieves precise reference tracking and exhibits excellent steady-state consistency across varying configurations, successfully enabling stable hover control for morphing UAVs.

Modeling the aerodynamic forces acting on the tether of a tethered multirotor UAV—particularly under high-speed platform motion or strong wind conditions—presents significant challenges in balancing modeling fidelity and real-time computational efficiency. To address this, this paper proposes a dual-path quasi-static modeling framework: (i) an analytical method based on catenary theory incorporating uniform aerodynamic drag, yielding solutions in under 1 ms; and (ii) a numerical method employing piecewise mass-point discretization coupled with CasADi/IPOPT-based nonlinear optimization, accelerated via warm-starting and analytical initialization to achieve real-time solutions within 5 ms. This framework uniquely unifies physical fidelity and computational efficiency for the first time. Experimental validation using force sensors confirms that the model meets engineering accuracy requirements. The resulting lightweight, scalable model has been successfully deployed in online control, trajectory planning, and offline optimization applications.