Sports analytics, tracking how swiftly the ball is moving of how players shift across the field is becoming a core component of how coaches make decisions and fans view games. Data for such analytics is presently sourced through cameras in courts and stadiums and is incredibly costly to acquire.
“There is an immense interest in analysing sports data through high-speed cameras, but a system can cost up to $1 million to maintain and implement. “It is not accessible to big clubs,” says Mahanth Gowda, a Ph.D. candidate in computer science and head author of the study. “Bringing IOT to Sports Analytics, we wish to reduce the expenditure significantly by substituting cameras with cost-effective internet-of-things devices, which cost less than $100 in total, to make it feasible for numerous other organizations to use the technology.”
The group, headed by Romit Roy Choudhury, an associate lecturer of computer engineering and electrical and computer science at Illinois, jointly with Sharon Yang from Intel has introduced advanced movement tracking algorithms from the various incomplete and noisy measurements of inertial measurement unit or IMU sensors and wireless radios, fitted inside a ball and player’s shows. If the technology gains traction, real-time analytics must be possible at anytime and anywhere.
The small sensors, which are wrapped in a protective case and distributed evenly in equipment, use inference algorithms that can track motion to within a few centimetres. They can precisely characterize 3D ball motion, like trajectory, revolutions and orientation per second.
“Such level of accessibility and accuracy could aid players in local clubs read their own performance from their smartphones through Bluetooth, or school coaches could offer quantifiable feedback to their students,” says Roy Choudhury, who is also a research lecturer at Illinois Coordinated Science Lab. The feedback could also aid with identifying and analysing player injuries, like concussions. The sensor inside a soccer ball, for instance, can estimate how hard it hits a player’s head, offering coaches an indication about whether to treat the player for head injury.
“We have really scratched the surface for applications with these sensors. The algorithms offer highly fine-grained detail and precision in measurements, but use popular measuring tools that can be found in any smartphone,” says Gowda.
The group, composed of students Ashutosh Dhekne, Sheng Shen, along with other Intel associates, have also been introducing methods to charge the sensors, comprising harvesting energy from the spin of the ball. “We are motivated to introduce such technology to help coaches make better decisions off and on the field and offer enhanced entertainment to viewers,” says Roy Choudhury. “We intend to bring advanced but cost-effective sports analytics to everyone, anytime, anywhere.”