“People think Robots are more capable than they actually are”

Helen Oleynikova is a NCCR DFAB senior researcher at the Autonomous Systems Lab at ETH Zurich. Her research is focused on autonomous mapping and planning for collision avoidance on drones. Walking through the lab, she gestures at flying platforms, wheeled robots and robotic arms. “These robots are incredibly different, but they all face the same fundamental challenge: interacting with the real world.”

Helen lying on the asphalt fixing a drone. Photo by Helen Oleynikova

The biggest misconception about robots is that people think they are far more capable than they actually are. “Robots are really hard,” Helen says. The real challenge isn’t just recognising objects or understanding images—a task AI has made impressive strides in—but translating that knowledge into real-world actions. “People think, ‘Okay, if an AI can recognise an object in a picture, then a robot should be able to go pick it up.’ But that’s not how it works.” The gap between digital perception and physical interaction is massive.

Helen points to autonomous driving as an example. “For years, people thought self-driving cars were basically solved. Google had mapped all of California. But then what happens when a kid jumps into the street? Or there’s unexpected construction? Or it snows—which it doesn’t do in California?” The unpredictability of the real world, she says, is the hardest problem in robotics. Would a perfect map be the solution? Helen is skeptical. “If we had a perfect, 100% accurate map of the world, if we had perfect sensors and actuators, then sure—robotics would be easy. But that’s not realistic.” Instead, she believes the future of robotics lies in developing systems that can function reliably despite imperfect, messy, and constantly changing environments. “The real challenge isn’t making robots work in perfect conditions—it’s making them work in the real world.”

Mapping the World for Robots

Helen’s journey into robotics began at Olin College of Engineering in Massachussets, where she earned her bachelor's degree before joining Google to work on Street View imagery. “I loved the team and the job, but I missed the physical part of robotics,” she recalls. This realization led her to ETH Zurich for a master’s in robotics, where she focused on computer vision before pursuing a PhD in autonomous mapping and navigation. Her research centered on enabling drones to create real-time 3D maps using only onboard sensors. “The goal was to figure out how drones could use onboard cameras and sensors to create a map of the world and avoid collisions.” Beyond drones, she explored applications in infrastructure maintenance and later worked on integrating mixed reality with robotics at Microsoft’s HoloLens project. “Robots create their own maps, and humans have their own perception. How can we combine these to feel like we’re truly sharing space with robots?”

Three roboticists on the side of the highway near Madrid, Spain, testing their robot for road maintenance and repair for the EU HERON project. Photo by Helen Oleynikova

Robotics: A Field Built on Collaboration

For Helen, one of the defining aspects of robotics is teamwork. “You can’t do robotics alone. There are so many components—mechanical design, actuators, perception, planning, power management. Every piece must work perfectly for the robot to function.” She values the collaborative nature of the field, where researchers rely on each other’s complementary skills to solve complex challenges. Collaboration also extends beyond engineering teams. “If people in a hospital say they need assistance, our job is to figure out how robots can fill those gaps.” Robotics is not just about technology; it’s about solving real-world problems through interdisciplinary cooperation.

Get to know the Autonomous Systems Lab: https://asl.ethz.ch/

The NCCR DFAB Researcher Portraits series introduces you to our researchers and their various fields of expertise. This article is part one of the interview with Helen Oleynikova. To mark International Women's Day on 8 March, the second issue focuses on gender relations in robotics.