be challenging. However, with the use of player data collection, researchers have been able to gather a detailed dataset that allows for personalized training feedback. This dataset, which includes over 7,763 badminton swings, utilizes wearable sensors and machine learning to provide real-time feedback and optimized movement suggestions. This stroke quality assessment not only offers valuable insights but also makes badminton training more accessible and affordable for athletes of all levels. The foundation has been laid for delivering coaching assistance and feedback through the use of this innovative technology.Advancements in personalized sports coaching are revolutionizing player performance. By harnessing the power of AI and utilizing published datasets, athletes can take their skills to the next level. Cameras and sensors strategically placed on the athlete’s body track joint movement patterns, muscle activation levels, and gaze movements, providing a comprehensive analysis of their performance.
This data allows for personalized feedback on technique and improvement recommendations to be easily accessible to athletes at all levels. A recent study published in the journalOn April 5, 2024, a team of researchers led by Associate Professor SeungJun Kim from the Gwangju Institute of Science and Technology (GIST), South Korea, in collaboration with researchers from Massachusetts Institute of Technology (MIT), CSAIL, USA, announced the creation of a MultiSenseBadminton dataset for AI-driven badminton training.
According to Ph.D. candidate Minwoo Seong, the first author of the study, “Badminton could benefit from these various sensors, but there is a scarcity of comprehensive badminton action datasets for analysis and training feedback.”
Supported by the 2024 GIST-MIT project, this study drew inspiration from MIT’s ActionSense project utilized wearable sensors to monitor various daily kitchen activities, such as peeling, slicing vegetables, and opening jars. Seong worked with MIT’s team, which included Joseph DelPreto, a postdoc researcher at MIT CSAIL, as well as Daniela Rus, the director of MIT CSAIL and a professor at MIT EECS, and Wojciech Matusik. Together, they created the MultiSenseBadminton dataset, which recorded the movements and physical responses of badminton players. This dataset, developed with input from professional badminton coaches, aims to improve the quality of forehand clear and backhand drive strokes. To accomplish this, the researchers gathered 23 hours of swing motion data.The study involved 25 players of different training backgrounds. They were given the task of performing forehand clear and backhand drive shots while being monitored by various sensors. These sensors included IMU sensors to track joint movements, EMG sensors to monitor muscle signals, insole sensors for foot pressure, and a camera to record body movements and shuttlecock positions. A total of 7,763 data points were collected and each swing was carefully labeled based on stroke type, player’s skill level, shuttlecock landing position, and impact location.
e to the player, and sound upon impact. The dataset was then validated using a machine learning model, ensuring its suitability for training AI models to evaluate stroke quality and offer feedback.
“The MultiSenseBadminton dataset can be used to develop AI-based education and training systems for racket sports players. By analyzing the disparities in motion and sensor data among different levels of players and creating AI-generated action trajectories, the dataset can be applied to personalized motion guides for each level of players,” says Seong.
The gathered data can enhance training through haptic vibration or electrical muscle stimulation.This dataset, such as the MultiSenseBadminton dataset, can potentially improve sports training by enhancing movement and refining techniques. It could also be used to create virtual reality games or training simulations that make sports training more accessible and cost-effective, potentially changing the way people exercise.
Ultimately, the researchers believe that this dataset could make sports training more accessible and affordable for a wider audience, promote overall well-being, and contribute to a healthier population.
