# Superhuman Safe and Agile Racing through Multi-Agent Reinforcement Learning Source: [Superhuman Safe and Agile Racing through Multi-Agent Reinforcement Learning]() Collection index: [llms.txt]() Data updated: 2026-06-20T05:00:05.764Z Authors: Ismail Geles, Leonard Bauersfeld, Markus Wulfmeier, Davide Scaramuzza Published: 2026-05-21T17:15:54.000Z ## Abstract Autonomous systems have achieved superhuman performance in isolation or simulation, yet they remain brittle in shared, dynamic real-world spaces. This failure stems from the dominant single-agent paradigm for physical applications, where other actors are ignored or treated as environmental noise, preventing effective coordination. Here we show that multi-agent reinforcement learning provides the essential safety scaffolding required for real-world interaction. Using high-speed quadrotor racing as a high-stakes testbed, we train agents to navigate complex aerodynamic interactions and strategic maneuvering with a variable number of racers. Through league-based self-play, agents evolve sophisticated anticipatory behaviors, including proactive collision avoidance, overtaking, and handling multi-agent physical interactions, including aerodynamic downwash. Our agents outperform a champion-level human pilot in multi-player races at speeds exceeding 22 m/s, while simultaneously reducing collision rates by 50 % compared to state-of-the-art single-agent baselines. Crucially, training with diverse artificial agents enables zero-shot generalization to safer human interaction. These results suggest that the path to robust robotic co-existence lies not in isolated safety constraints, but in the rigorous demands of multi-agent interaction. Multimedia materials are available at: https://rpg.ifi.uzh.ch/marl ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Time Trial Drone Racing on Real-world racing track 3-lap 21 gates](): Time Trial Drone Racing; Real-world racing track 3-lap 21 gates; Average Race Completion: 100. - [Multi-agent Drone Racing (AI Only) on Real-world racing track 3-lap 21 gates](): Multi-agent Drone Racing (AI Only); Real-world racing track 3-lap 21 gates; Race Completion Rate (%): 100. ## Original sources and implementation [Original paper]() [PDF]()