From drone racing to real-world impact
At the Zurich AI Festival, Prof. Dr. Davide Scaramuzza discussed how advances in AI are pushing the capabilities of autonomous robots – and why the next major challenge is making them robust enough for the real world.
At “AI Meets Reality: From Algorithms to Autonomous Systems,” held as part of the Zurich AI Festival 2026, Prof. Dr. Davide Scaramuzza gave a talk titled “From Drone Racing to Real-World Impact.”
In his talk, Davide highlighted both the remarkable progress and the remaining challenges in AI-powered robotics. While advances in computer vision have enabled mobile robots to become increasingly autonomous, humans still outperform robots when it comes to agility, versatility, and especially robustness.
Autonomous drone racing provides a striking example of how AI can already surpass human capabilities in a specific task. After seven years of research, Davide and his team developed the autonomous drone Swift, which was able to race against and outperform world-champion drone pilots. Yet, as Davide emphasized, superhuman performance in a well-defined task does not mean that AI is generally better than humans.
The next major challenge is robustness: enabling robots to operate reliably in complex and changing real-world environments. Advances in agility and perception could ultimately translate into practical applications ranging from search and rescue and infrastructure inspection to wildfire prevention and space exploration.
Davide also highlighted event cameras as an example of sensing technologies inspired by biological vision. By perceiving motion with very low latency, these cameras can help robots operate in challenging conditions and are being applied to both drones and automotive vision.
Ultimately, the goal is to turn advances in AI and robotics into real-world impact. Drawing on examples ranging from collaborations with industry and NASA/JPL to technology transfer and UZH spin-offs, Davide showed how research can lead to applications in areas such as agriculture, emergency response, and spatial AI.
Davide leads the Robotics and Perception Group at the Department of Informatics.