مدل شبیه سازی فراروی پرتاب برای آموزش سایبرنتیک

Sports training systems are widely used to improve the overall performance of individuals engaged in various activities. Here, we propose the use of optimized muscle activation signals, hereinafter called OPTIMAS, as a tool for improving the overall sport performance of engaged subjects. The system consists of a series of steps that begin with the calculated OPTIMAS waveforms based on musculoskeletal model. The OPTIMAS waveforms are subsequently compared to the measured EMG waveforms. A trial and error based feedback approach is then taken to narrow the difference of the waveforms to finally suggest the best performance of the individual subject. The results of the proposed training system show that the method is of great advantage with a gain that leads to a better performance especially in the early stage of the training term.

New skills in sports performance are generally attained through successive engagement of the inexperienced subjects with the aim of imitating the body motion of a skilled athlete. This imitation corresponds to the visually observable information from skilled athletes' posture, body parts position, joint trajectories and motion timings, etc. It is known that having an image to imitate accelerates motor learning, especially in the early stage of practice. However, this visually observable information is restricted to kinematic information. Since kinematics information does not have dynamic information such as muscle tension which drives the body motion, a learner needs to develop an internal model, i.e. a correlation map between body motion and motor command, by trial and error. On the other hand, electromyography (EMG) is a graphical record of a biomedical electric currents signal which directlyreflects motor commands from the brain to activate muscles contractions. The general aim of EMG is to analyze the function and co-ordination of muscles in different movements and postures of skilled athletes and inexperienced subjects. The method is however, not necessarily effective as an inexperienced athlete would face difficulty in directly imitating skilled athlete’s EMG due to the inherent difference in the physical characteristics. Therefore, the differences in physical characteristics limits the use of skilled athlete’s EMG signal by inexperienced subjects. In this study, we propose a training system that helps inexperienced athletes to acquire skills through a repeated comparison of their EMG signals measured in real time to that of the signals produced by the optimization calculation using a model considering the physical characteristics of each subject. We call this training system, "cybernetic training" and refers to a feedback based signal produced artificially by optimization of the model calculations. In this work we studied the underhand throw of softball player to verify the validity of the proposed cybernetic training. At first, we calculate the optimal muscle signal for an underhand throw of a softball based on the subject's physical characteristics. Then, the subject practices to improve his skills by comparing his integrated electromyogram (IEMG) measured in real time with that of the OPTIMAS waveform produced by the optimization calculation. Here, the subject aims to approach the OPTIMAS waveform. The preliminary results showed that the proposed method is effective especially in the early stages of training.

In this paper, we proposed a musculoskeletal model and simulation analysis of the upper limb for a cybernetic training system. We obtained OPTIMAS waveforms using the proposed simulation model and the subjects’ IEMG waveforms peak, in a particular timing, was similar to that of the OPTIMAS waveforms attained by cybernetic training. In addition, the subjects that used the cybernetic training were able to throw faster than the ones that did not, and the velocity of the pitched balls were markedly increased in these subjects during the early stages of the training term. Therefore, the proposed cybernetic training was an effective method to improve skills. This simulation model is able to correspond to the subject's physical characteristics by changing the coefficients for the parameters of rigid links and the muscle parameters. Thus, various subjects are able to practice the cybernetic training using individualized OPTIMAS waveforms developed by our simulation model. Ohta proposed a cybernetic training system that is able to perceive tangential acceleration of the hammer by using real-time auditory feedback[3]. As a future work, our training system, based on OPTIMAS, will aim at creating an electric stimulus feedback system in real time to improve the ease of use of the system.