Development of Training System to Improve Walking Effect - Evaluation of Training Prototype Based on Estimation of Muscle Activity by Using Biped Robot Model and Machine Learning Technique

Abstract

Prevention of mobility loss due to aging is critical to avoid or postpone the necessity for continuous care in late stages of life. In particular, knee extensors have a key role in maintaining balance and stabilizing gait. As loss of muscle strength in knee extensors jeopardize mobility abilities, continuous training of these muscles is fundamental.

Walking is the most popular daily exercise. However, the training effects of simple walking are limited to lower leg muscles and are not sufficient to prevent progressive deterioration of knee extension muscle strength if not integrated with specific exercises targeting those muscles. Therefore, we are working on the development of a training device to apply additional mechanical load on the knee during walking and extend the training effect of simple walking to knee extensors. The training device will be designed for maximum versatility using a compact oil damper with no need for electrical power. The specifications of the device will be determined using results from preliminary research on knee extensor muscle activity during walking, evaluated with an IMU （Inertial Measurement Unit） based measuring system. In this study, we designed a prototype of the device and evaluated if the device could induce sufficient muscle activity in the knee extensor muscle. In the future, we are planning to replace the oil damper with an active actuator and integrate the measuring system in the device to allow adaptive training load depending on real-time measured knee extensor muscle activity.