University of São Paulo: Researchers from USP Ribeirão Preto create a treadmill capable of assessing runner fatigue
Technology has evolved a lot over the years and, today, it is increasingly present in all areas. At the same time, the number of people concerned about their health and who started to exercise also increased. Joining technology and sport, then, has become an essential job, mainly to avoid injuries to those who are starting in practice.
In this context, researchers from USP Ribeirão Preto developed a treadmill with sensors and an intelligent system capable of assessing a runner’s fatigue. The new national technology is explained by Sérgio Baldo Júnior, a student in the Biomedical Informatics course at the Ribeirão Preto Medical School (FMRP) at USP and responsible for the project. “From signs of strength, it is possible to extract some characteristics of it and, along with the use of artificial neural networks, it is possible to predict the running pattern of that person and whether he is tired or rested.”
Professor Renato Tinós, from the Faculty of Philosophy, Science and Letters of Ribeirão Preto (FFCLRP) at USP and advisor to Baldo Júnior, says that all the analysis is possible through the artificial intelligence network coupled directly to the equipment and which stores the sensor data . “The person runs on the treadmill and this database will be used to train the neural network, which will then be able to classify that person’s running pattern.”
It has 4 load cells used as vertical force sensors
The force signals emitted by each load cell go through a signal amplifier
A load cell attached to the base of the conveyor
Data are sent through different channels to a data acquisition board
MATLAB software for recording signals in real time
Example of signal acquisition
According to the researchers, it is a treadmill that involves a high level of control, with sensors and motors that reach 40 or 50 kilometers per hour (km / h). Currently, similar treadmills in the national market can cost up to R $ 120 thousand, while imported models can reach US $ 230 thousand. The high costs and the importance, especially for national research, of accurate and more accessible equipment motivated researchers to develop the project.
That’s what Baldo Júnior’s other advisor argues, Professor Paulo Santiago, from USP’s School of Physical Education and Sport (EEFERP). “Due to the impossibility of buying equipment like this, the idea of building our own treadmill came up”. And, according to the professor’s evaluation, they obtained a new treadmill even better than an imported one. “The engine is more powerful and the value, much cheaper”, he guarantees, pointing out that the cost of the treadmill they developed does not exceed R $ 20 thousand.
The scientific and academic importance of the USP treadmill is highlighted by Santiago, who claims to have developed a “cheap and excellent quality unit for use in research that requires accurate and refined monitoring for high-level corridors”. In addition, it can also benefit the population since the idea is that the treadmill reaches the market. “It will help to monitor the training loads so that, at least, the person knows when he is tired or rested, which can help prevent pain and injury”, highlights Professor Santiago.
Using artificial intelligence in the practice of physical exercises, with the power to prevent injuries, made USP Ribeirão Preto’s work awarded the 16th National Meeting of Artificial and Computational Intelligence (ENIAC 2020). The event, which discusses innovations, trends, experiences and developments in the fields of Artificial and Computational Intelligence, was held online between 20 and 23 October.
Recognition of the project by the scientific community was received by academic Baldo Júnior in a “mixture of pride and joy”, as he himself describes his feeling. For the student, winning the prize at an event of this level is very difficult and the fact makes him very excited to continue, since his “work was recognized by those who are a reference in the area”.
The project “Use of artificial neural networks to classify running patterns on a treadmill in highly developed sports” was financed by the Dean of Research at USP, by the São Paulo Research Foundation (Fapesp) and by the Council National Scientific and Technological Development (CNPq).