This work involves implementing a feedback-loop control for industrial servomotors that generate posterior forces. These forces cause balance disturbances on subjects performing standing movement tasks. Tracking tags or trackers which are used to track movement were attached to the subject. Vicon motion capture cameras captured the movement of the trackers in a previously set coordinate system. Using the User Datagram Protocol, we sent movement data to a computer running on a Simulink Real-Time operating system. Depending on the speed of the subject's movements, a force was generated on the motors on which puller cables are installed. These cables were chained to a special leather belt that the subjects fastened around their waist. The primary objective was to calibrate the servomotors to exert a sufficient force to disrupt the subject's balance by displacing them from their center of mass, while ensuring that the force applied would not evoke a fall. The subject's movements and the targets to which they had to move, were displayed in real-time on a monitor in the form of circles. Additionally, after the completion of each trial, the corresponding result was promptly displayed on the monitor. Through repeated experiments with multiple subjects, we managed to achieve the desired functionality of the system. This enabled us to generate a sufficiently large force with the motors, and capture and display the movement of each individual subject with Vicon motion capture cameras. We were able to then immediately graphically display the saved data after each experiment. These results were analyzed and compared with previous reinforcement learning studies.
|
