Sistema electrónico para la detección de posición angular y ergonomía de ciclistas empleando visión artificial

Abstract

This research project proposes the implementation of an electronic system for the angular detection and ergonomics of cyclists using artificial vision, so that the adjustments recommended by it help to have a correct ergonomic positioning at the time of cycling routes. Therefore, it is very important to perform a biomechanical analysis of the body gestures generated by the cyclist during the pedaling process. The electronic system for man-machine coupling in cyclists, bikefit, consists of a mini computer Jetson Nano which is connected to different input and output peripherals that allows 2D detection of the skeleton of the athlete to be analyzed. It shows in video and in real time the different points of the joints that allows to calculate the angles generated in each limb, thus obtaining values that can be compared with previous studies and establish recommendations for bicycle adjustments. This in order to have an optimal position to increase performance and avoid injuries in the short or long term. All the data obtained are displayed in an easy to interpret interface, both for the person performing the analysis and for the person being analyzed. Two reports are generated, the initial report that indicates the adjustments and angles with which the cyclist has arrived and the final report that shows the fitting process with the angles and recommendations executed, this report has the function of informing the cyclist of the variations made on his bicycle. The results obtained in the performance tests are consistent, showing that the system has a confidence value of 95.22%, after comparison with a manual measurement method with a goniometer, demonstrating that the prototype is valid for bike fitting analysis.