Innovative Motion Monitoring System on a Bike Trainer – A Breakthrough in Biomechanics Analysis
A research team consisting of scientists from the Lublin University of Technology, the Netrix S.A. Research and Development Center, the WSEI Academy in Lublin and the Polish Academy of Sciences have developed an advanced system based on computer vision, enabling tracking of human movement during exercise on a bicycle trainer. This solution, using machine learning algorithms, allows for a detailed analysis of the biomechanics of the body in real time, without the need to attach sensors or markers to the body. This is a significant step towards increasing the availability and comfort of modern diagnostic methods in sports and rehabilitation.
The foundation of the technology is the use of the MediaPipe model, which allows for the identification of key anatomical points and mapping of the user’s movement skeleton. Thanks to this, it is possible to track and analyze parameters such as angles between joints, which is the basis for assessing riding technique and body posture. Based on the designated spatial points, algorithms were developed to calculate the angles between body segments using vectors, which allows for precise mapping of movement kinematics – using points such as the shoulder, elbow and wrist, it is possible to precisely determine the angle of joint flexion.
In order to verify the effectiveness of the system, comparative tests were carried out with the professional HTC Vive Tracker v. 3.0 motion tracking system. 3.0. The results obtained showed that the mean absolute error (MAE) between the measurements of both systems was only 3.92 degrees, which confirms the high compliance and reliability of the developed solution. Another important advantage is the system’s resistance to changing lighting conditions. The use of two cameras and advanced image processing techniques allows for stable operation even in unfavorable environments.
The designed application offers not only real-time movement visualization, but also comprehensive statistical analysis, generation of graphs and boxplots presenting changes in the angle in individual joints while riding the trainer. This approach enables detailed tracking of exercise dynamics and observation of the user’s progress over time.
The solution is widely used in various fields. It can support the rehabilitation process of patients after injuries, enable analysis of driving techniques in sports medicine and serve to optimize movements in competitive training. By combining modern vision methods, artificial intelligence and kinematic analysis, a system was created that is characterized by high measurement accuracy, while eliminating the need for expensive and complex sensor-based systems.
This breakthrough solution sets a new direction in the development of technologies supporting the monitoring of physical activity and biomechanics of movement.
Want to learn more? Read the full publication: link to publication