Performance Framework for High-Dynamics Multirotor Systems
An integrated geometry-driven approach to modelling, control, and performance optimisation of racing drones under realistic flight conditions.
An integrated geometry-driven approach to modelling, control, and performance optimisation of racing drones under realistic flight conditions.
This work consolidates a research pipeline developed across multiple studies on racing drone dynamics, co-simulation, data-driven analysis and experimental validation. The objective is to establish a unified framework for analysing and optimising multirotor performance under realistic high-dynamics conditions, where geometry, dynamics, and control are strongly coupled.
The framework builds upon a sequence of complementary contributions:
• Experimental characterisation of racing drone dynamics
• Co-simulation platforms integrating geometry, trajectory, and control
• Data-driven flight analysis using motion capture systems
• Open datasets for validation and benchmarking
These elements provide the basis for a consistent understanding of multirotor behaviour under aggressive flight conditions.
The present work integrates these components within a performance-oriented optimisation framework, linking airframe geometry, nonlinear flight dynamics, and control synthesis. This enables the evaluation of drone trajectories as a function of their geometric configuration, closing the loop between design, dynamics, and control.
The framework explicitly accounts for real-world experimental constraints, including sensor noise, calibration, signal synchronisation and measurement uncertainty. These aspects are integrated into the modelling, control, and optimisation process to ensure consistency between simulation and flight-test data. This enables reliable performance evaluation under realistic operating conditions, supporting implementation in experimental and engineering environments.
The key idea is that airframe geometry should not be treated as a static design parameter but as an active element that shapes flight dynamics and control behaviour. This perspective enables a physically consistent analysis of system performance near both the dynamic and control limits.
This approach is particularly relevant for high-performance flight regimes, such as drone racing and aggressive manoeuvring, where traditional separated design methodologies fail to capture the interactions among structure, dynamics, and control. By integrating these elements, the framework provides a more reliable basis for performance-driven system design.
This work represents the first integrated implementation step towards a broader geometry-driven framework for guidance, navigation, and control. At this stage, geometry is embedded in the modelling, control, and performance optimisation loop, enabling physically consistent analysis under realistic high-dynamics operating conditions. The next stage is to extend this geometric perspective throughout the wider control hierarchy, from flight dynamics and control allocation to geometric guidance, intelligent decision-making, and global optimisation strategies for autonomous systems operating near performance limits.