Prediction of muscle length during walking by neural networks

Prediction of the muscle-tendon behavior during walking is nowadays undergoing extensive analysis. The aim of this article is to introduce a possible method of predictions of muscle behavior which can be used for rehabilitation, and also for controlling the artificial muscles, actuators of prosthesis or rehabilitation facilities of the future. Our work focuses on predicting muscle-tendon lengths during human gait with the use of angle-time diagram. A group of healthy children was measured using a Vicon motion analysis system. The 3D kinematic data were recorded and the OpenSim software system was used to identify the joint angles and muscle length, which are represented by time diagrams. In conjunction with artificial intelligence, time diagrams offer a wide area of medical applications. We tested and verified new way of prediction of muscle-tendon length based on neural networks. Artificial neural networks for predicting the muscle behavior learned by time diagrams predicted the muscle-tendon behavior of healthy children. The new method based on angle-time diagrams and artificial neural networks for predicting the muscle behavior has never been described or used before. This work has attempted to describe potential ways of applying angle-time diagrams and graphs of the muscle length in conjunction with artificial intelligence. We have shown new methods that have subsequently been proved by simulations in MATLAB software.