When virtual reality boosts boxers’ performance

π Science and Technology

π Digital

The 2024 Olympics are just around the corner and it’s not just the city of Paris that’s gearing up… science is getting involved too! At Inria Rennes, we conduct research with the Olympic boxing team and its coach: as part of the REVEA project1, we are looking to determine how virtual reality can help the team work on certain “sub-skills” to improve such as speed, motor coordination and strength training, to name a few. Contrary to what one might expect, our VR system is not intended to replicate an actual match or competition, but rather to identify a specific sub-skill and improve it in a target combat situation. In the case of boxing, the sub-skill we seek to train is the ability to anticipate the opponent’s attacks.

Credit: REVEA.

Without virtual reality, two boxers have to fight head-to-head, one of them is invited to reproduce typical attacks in order to analyze the opponent’s behavior. However, it is very difficult to control and faithfully reproduce situations from one trial to another. This approach also inevitably limits training time, because such exercise is physically demanding – real punches are exchanged and received. In virtual reality, this type of training can be performed with our avatar and a virtual opponent without risk of injury. It also offers a range of INCENTIVES that goes beyond reality: we can play with visual and auditory perceptions or even face an opponent with extraordinary skills to push the limits of traditional training.

A map of sports behavior

For this purpose, our team has developed a virtual reality simulator, which uses three-dimensional measurements of the movements of professional boxers obtained using motion capture devices (through the Immermove platform installed in the M2S laboratory). This data, which includes, for example, light predictive stroke movements, such as hip movements, is then transferred to virtual people whose morphologies may be different. The boxer can then visualize his opponent represented by his avatar and displayed in a virtual reality helmet. As researchers, we then observe the boxer’s reactions to different types of punches and whether he is sensitive to these slight anticipatory movements perceived in the opponent. These observations can be quantified, for example, by success rate or reaction time, in order to create a kind of predictive performance map of the boxer.

The training program includes physical preparation, mental preparation, technical preparation… and virtual reality sessions!

In addition to its application in boxing, we have used our approach to support the goalkeeper training of Stade Rennais Football Club. Our virtual reality system has trained them not to focus on the player with the ball, but to better manage the information available in their peripheral vision. Indeed, if the attacker makes a last-second pass to someone on the other side of the goal, the goalkeepers will not discover the situation until too late and may not have time to react.

Applied research

Our research is highly applicable: our objective is to serve the Olympic teams so that they are ready for Paris 2024. Indeed, thinking about how to incorporate these technologies into the daily life of athletes is one of the demands of the sports world. Therefore we asked ourselves the following questions: how long should the training take with this type of technology? What time of day and what time in the training program? How long should a typical session be? Who should we train with – Olympic hopefuls or all athletes – and how often? Our approach is part of a larger training program, set by the trainer, which includes physical preparation, mental preparation, technical preparation… in addition to these sessions in virtual reality.

We have made our system available to the training center of the Olympic teams and INSEP scientists working with the French Boxing Federation. As a result, we continue to develop it and make it more effective thanks to feedback from athletes and coaches. For example, in our first version, the opponent made a lot of repetitive movements, which undermined the realism of the situation. There is also the issue of improving the equipment used for virtual reality training. Should immersive headphones be wide-angle or not? Are sensors on the hands enough or should sensors be provided on the whole body?

Translate the data

We have a software platform that is based on a commercial tool called Unity, traditionally used for video game development. It loads a three-dimensional environment – in our case, a boxing ring – as well as virtual people and their movements. We write a script in which the movements of virtual people are stored in “motion libraries” that we review. Then there are algorithms related to controlling the movement of the avatar. All these elements provide us with all the metrics that allow us to measure the response times and the waiting capacities of boxers.

We also started working on artificial intelligence, neural networks and deep reinforcement learning to improve the behavior of virtual adversaries. This piece of technology is not yet mature, but early results show that it is possible to reproduce complex interaction behaviors between two boxers, with very few observations of real fights. Pushing this approach to its limits, we can consider simulating the gestures and strategy of a particular boxer from several video sequences of his previous fights.

Interview by Isabelle Dumé


1The REVEA project is funded by the Priority Research Program “Very High Sport” (PPR) of the Future Investment Program, set up in view of the Olympic and Paralympic Games Paris 2024. The program is co-supported by the Ministry of Higher Education, Research and Innovation, and Ministry of National Education, Youth and Sports. (https://www nese/n’t ent_m igrat ion /d nuk jam / p ia 3 livret _ 8 5 3 1 2 7 .pdf)

Leave a Comment