Improving Drone Racing Performance Through Iterative Learning MPC

2025-08-01

AI Summary

Autonomous drone racing faces challenges including modeling high-speed nonlinear dynamics and balancing real-time decision-making with safety–time optimality. This paper proposes an iterative learning-based model predictive control (LMPC) framework tailored for racing scenarios. It innovatively introduces an adaptive cost function, a displacement-based local safe set, and Cartesian-coordinate modeling—thereby avoiding singularities inherent in Frenet-frame formulations—and explicitly integrates dynamic weight optimization with real-time updates of safety constraints. The method iteratively refines trajectory performance across successive laps. Simulation and real-world flight experiments demonstrate up to a 60.85% reduction in lap time; even against a finely tuned MPCC++ baseline, it achieves a 6.05% improvement. Crucially, all flights maintain zero collisions, significantly enhancing both racing efficiency and robustness.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Search and Optimization: Learning to Search
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Autonomous drone racing presents a challenging control problem, requiring real-time decision-making and robust handling of nonlinear system dynamics. While iterative learning model predictive control~(LMPC) offers a promising framework for iterative performance improvement, its direct application to drone racing faces challenges like real-time compatibility or the trade-off between time-optimal and safe traversal. In this paper, we enhance LMPC with three key innovations:~(1) an adaptive cost function that dynamically weights time-optimal tracking against centerline adherence,~(2)~a shifted local safe set to prevent excessive shortcutting and enable more robust iterative updates, and~(3) a Cartesian-based formulation that accommodates safety constraints without the singularities or integration errors associated with Frenet-frame transformations. Results from extensive simulation and real-world experiments demonstrate that our improved algorithm can optimize initial trajectories generated by a wide range of controllers with varying levels of tuning for a maximum improvement in lap time by 60.85%. Even applied to the most aggressively tuned state-of-the-art model-based controller, MPCC++, on a real drone, a 6.05% improvement is still achieved. Overall, the proposed method pushes the drone toward faster traversal and avoids collisions in simulation and real-world experiments, making it a practical solution to improve the peak performance of drone racing.

Problem

• Enhancing real-time decision-making for autonomous drone racing
• Balancing time-optimal tracking with safety constraints
• Improving trajectory optimization for faster and collision-free racing

Innovation

• Adaptive cost function for dynamic tracking
• Shifted local safe set for robustness
• Cartesian-based formulation avoids singularities