Skip to content
Photo: Klas Rockberg/Svenska Skidförbundet
Alpine skier Felix Monsén. Photo: Klas Rockberg/Svenska Skidförbundet View original picture , opens in new tab/window

New collaboration will help the national alpine team to ski faster

Published: 31 May 2021

AI is becoming an important tool in the competition for Swedish Olympic and World Cup medals. Luleå University of Technology's researchers in robotics and machine learning have started a collaboration with the Boden company TNTX with the goal of helping the alpine national team to ski faster and safer.

– Our skiers are in an extremely competitive environment where hundredths of a second differ between Olympic gold and an honorable fourth place. Skiers are challenged not only by their competitors but also by the surrounding environment. By using modern technology and knowledge, we can now create opportunities for our athletes to focus entirely on optimizing their decisions during the ride, like a Formula 1 driver. Our goal is to compete for medals in all disciplines in the Olympics in Cortina 2026 and this collaboration is a big part of that ambition, says alpine manager Tommy Eliasson Winter.

New visualization tool 

The project aims to improve the performance of skiers. AI and a completely new visualization tool will help national team athletes understand more about the optimal racing line. In a virtual environment, skiers and leadership can quickly get feedback and compare in real time how a choice of racing line affects the time. Factors such as individual skiing technique and snow and weather conditions are also taken into account.

– We develop an AI and race line that is faster than the best skiers in the world. It makes it possible to study and compare your own race choices with an optimized line to quickly see how to improve. It is great that we, as a private company, make this a reality together with the alpine national team and Luleå University of Technology, says Thomas Vikström who has had a long career at Tesla in the USA.

Together with the entrepreneur Thomas Lindgren - who has founded several companies in the gaming industry – Vikström has founded the company TNTX in Boden, which develops platforms for augmented and virtual reality with the goal of having a safer working environment in the mining, steel and forest industry. The technology can also be applied in sports.
– It is partly the same technology that is also used in self-driving cars today. We want to give the Swedish national team another tool to help them become the best in the world, says Thomas Vikström.

Years of experience

New in the collaboration are Luleå University of Technology's research groups in robotics and AI and machine learning.
– The robotics and AI team is bringing in know how in this project on AI in perception, localisation and reasoning that has been developed for embedding autonomy in drones in multiple research and development projects for the last 15 years, says Professor George Nikolakopoulos, who leads Luleå University of Technology's research group in robotics and AI.

Professor Marcus Liwicki leads the research group in achine learning.
– The team contributes with expertise on deep learning and reinforcement learning to optimize the racing part in the virtual environment. Technologies that previously won in virtual racing environment and ATARI games will be tuned to work on the much more complex environment of downhill skiing, says Liwicki.

"Challenges current standards"

Tommy Eliasson Winter points out that, for example, downhill skiers at the highest level have relatively few opportunities to train on slopes where World Cup competitions and championships are held.
– One of our best skiers, Felix Monsén, has been to Kitzbühel maybe eight times, the most experienced have practised there on about 40 occasions. Thanks to the fact that the new tool simulates real environments, as in Kitzbühel, there is good potential to improve and shorten the skiers' development over time. This can take our sport to a new level from a technological perspective and challenging current standards in terms of the ability to analyze and make data-driven decisions, says Tommy Eliasson Winter.