As far as sports science is concerned, artificial intelligence (AI) can be a useful tool in many important research areas, says Dr Bence Kopper, Associate Professor at the Department of Kinesiology at the Hungarian University of Sports Science, highlighting four of those areas.
The international scientific symposium entitled "The 2nd Biomechanics in Sport and Ageing: Artificial Intelligence (AI)”, which is organized by the Hungarian University of Sports Science (TF) and the Department of Kinesiology will be held between 15 and 16 October in Budapest, Hungary.
The aim of the symposium is to provide a scientific platform for a state-of-the-art update on the progress of AI in sport biomechanics and ageing. A total of 14 experts from nine countries will present data-based examples of the use of AI in sport and ageing research.
Dr Bence Kopper, Associate Professor of the Department of Kinesiology, gave an interview on the use of artificial intelligence in elite and recreational sport for the elderly.
He said that researchers already have plenty of data from various variables available from the measurement of athletes. Artificial intelligence is playing an increasingly important role in analysing the data, for example in identifying relationships, correlations and patterns. Previously, this has been executed with the use of various statistical softwares, but compared to these solutions, AI is a big step forward, as it is a self-learning machine capable of recognising more complex relationships and patterns. Bence Kopper added that as a lot of new input is accessible from the athletes' movements, AI-based software systems will have to be even smarter in the future, which is possible by the use of machine learning technologies. This is also important because today, as the success of a top athlete depends on tenths or even milliseconds, AI can contribute a lot to the final result.
In this context, the role of artificial intelligence in the development of training plans for elite athletes cannot be neglected either, Bence Kopper comments on the next area. It is crucial that coaches and experts in the field of sports performance are able to seek for the optimal plan and tailor the training of athletes accordingly to achieve the maximum performance achievable for the individual. Coaches need to continuously control a number of components, such as the athletes' workload level during a training session or even the timing of form during the season. Artificial intelligence continuously measures and monitors the athlete's performance parameters longitudinally, and researchers can identify individual patterns from these measurements. These can then be used to determine the workload an athlete should receive in a training session and also the recovery intervals that should be built into their training plan to help them achieve the best possible performance.
The third and one of the most exciting areas where AI can be of great help is in the field of monitoring the rehabilitation of athletes and elderly individuals. Artificial intelligence-based systems, which are similar to facial recognition systems, can recognise patterns, learn to identify the movement of a person's limbs without markers through machine learning, and then fit a body model onto these limbs, from which researchers can extract various kinematic and biomechanical data.
The invention made by Bence Kopper and colleagues is based on this technology: an AI motion recognition system on one hand recognises movement of the limbs, creates a body model, and on the other hand determines kinematic data and joint angles, and by using this data creates feedback about the execution of the movement.
"To illustrate with an everyday example, when an elderly or injured person is undergoing rehabilitation, the physiotherapist teaches them the movements they need to execute at home to recover. The physiotherapist will explain how to hold the limbs while doing the exercise, for example, how far to raise the hand. Our software not only recognises the movements via a simple web camera, but also gives numerical feedback, and gives the exerciser a sound or light signal when the movement has reached the correct joint angle. This is very important because in home rehabilitation, movements often become inaccurate and are being executed incorretly, which can hinder recovery and also pose a risk of injury. Our system works online, in real time, which means that the physiotherapist can follow the patient's movements from a remote location and correct them immediately," he explains.
The fourth direction of development is linked to the fact that smart watches and smart devices, which are becoming more and more widely available, are developing at an astonishing speed, which means that they can now estimate parameters such as distance covered, running speed or acceleration, but most importantly work, energy used, calories burned and power output much more accurately as the technology develops. Based on all of this, professionals can, for example, draw up a more precise training plan but also a nutrition plan that tells the athlete how much of what and what kind of food to eat.
And here's an important point: this processed data including distance, running speed, acceleration, energy can be made more accurate by artificial intelligence, which can then be used by the professionals to make decisions. These developments will create increasingly more accurate performance monitoring applications available not only for the elite athletes, but also for lower-level athletes and recreational sportsmen and women.
The detailed programme of the 2nd Biomechanics Symposium is available here and you can register here.
