Flatness-Preserving Residual Learning for Real-Time Tight Quadrotor Formation Flight

2026-07-13

AI Summary

This work addresses the challenge of instability and potential collisions in tightly spaced quadrotor formations caused by aerodynamic disturbances such as downwash-induced turbulence. To tackle this, the authors propose a physics-informed residual learning framework that compensates for complex aerodynamic perturbations while rigorously preserving the system’s differential flatness. This enables efficient feedback linearization control combined with feedforward disturbance rejection. Requiring only 30 seconds of training data and operating within a 5 ms control loop, the method achieves high-precision formation flight with low computational overhead. Experimental results demonstrate a 31% reduction in average tracking error compared to a nominal baseline, matching the performance of nonlinear model predictive control while reducing computational complexity by an order of magnitude.

Abstract

Quadrotors flying in tight formations are severely affected by turbulent aerodynamic interactions, such as downwash, that can cause catastrophic collisions if left unmodeled. To compensate for these effects, we propose a physics-informed residual dynamics learning framework that captures complex aerodynamic interactions while ensuring the joint multi-quadrotor system remains differentially flat. We leverage this preserved flatness to design a computationally efficient feedback linearization controller that is easily tunable with linear control techniques and cancels aerodynamic disturbances via feedforward compensation. Hardware experiments demonstrate our framework reduces average tracking errors by 31% compared to nominal baselines. Crucially, our lightweight approach matches the tracking performance of state-of-the-art nonlinear model predictive control (NMPC) while requiring an order of magnitude less computation. We are the first to show that stable, tight formation flight can be achieved with under 30 seconds of training data and a 5ms loop rate, unlocking high-fidelity aerodynamic compensation for compute-constrained flight stacks.

Problem

• quadrotor formation flight
• aerodynamic interactions
• downwash
• collision avoidance
• real-time control

Innovation

• differential flatness
• residual learning
• aerodynamic interaction compensation
• feedback linearization
• real-time formation flight