Sports training relies heavily on practice, but what if athletes could learn from the techniques of professionals? Thanks to AI-based personalized coaching assistants, this dream is becoming a reality. These assistants use advanced technology, like cameras and sensors, to track every move an athlete makes during training sessions.
Researchers at the Gwangju Institute of Science and Technology (GIST), South Korea, along with collaborators from Massachusetts Institute of Technology (MIT), USA, have developed a groundbreaking tool for badminton training: the MultiSenseBadminton dataset. This dataset captures the movements and physiological responses of badminton players, aiming to enhance their techniques.
WHAT IS MULTISENSEBADMINTON?
Inspired by MIT’s ActionSense project, which tracked everyday tasks using wearable sensors, the MultiSenseBadminton dataset focuses on improving forehand clear and backhand drive strokes. The researchers collected extensive data from 25 players with varying skill levels, using sensors to monitor joint movements, muscle signals, foot pressure, and even shuttlecock positions.
Each movement was carefully analyzed and labeled based on stroke type, player skill level, shuttlecock position, impact location, and sound upon impact. This meticulous labeling ensures the dataset’s quality, making it suitable for training AI models to evaluate stroke quality and provide feedback.
WHAT DO EXPERTS THINK?
According to Associate Professor SeungJun Kim, this dataset can revolutionize badminton training by creating personalized motion guides for players at different skill levels. By analyzing differences in motion and sensor data, AI-generated action trajectories can offer tailored feedback to enhance players’ techniques.
The data collected can also be used to enhance training through methods like haptic vibration or electrical muscle stimulation, promoting better motion and refining swing techniques. Additionally, this dataset could fuel virtual reality games or training simulations, making sports training more accessible and affordable for all.
In the long run, the researchers believe that the MultiSenseBadminton dataset could promote overall well-being by making sports training more accessible to a broader audience. With AI-driven coaching assistants and advanced datasets, the future of sports training looks promising, paving the way for healthier populations and improved athletic performance.
The Sportz Planet Desk,
Atharva Shetye.
