Design and application of intelligent fitness system based on android system

Abstract

AbstractThe purpose of this study is to make use of the features of mobile smart phones to create a simple step by step fitness application that can exercise anytime and anywhere. A third‐party open‐source library based on the Android platform was adopted in the entire technology selection, which greatly reduces the writing of a lot of pattern code and enables developers to focus more on the development of fitness system functions. The system has basic functions such as “Fitness data” and “Fitness video courses,” and collects fitness movements through sensors. Based on the extracted eigenvalues and support vector machines (SVM) algorithm, the system determines whether the fitness actions are standard or not, and uses back propagation (BP) neural network to classify and count fitness actions, to achieve recognition of fitness movements. The experimental results show that, the response time, running speed, and load of the system meet daily needs, the system can give the correct instruction probability within 3 s is 74.2%, the correct instruction probability within 3 ˜ 4 s is 17.5%, and the recognition probability after 4 s is only 6.67%, which proves that the intelligent fitness system designed in this study based on the Android system meets basic usage requirements in terms of performance and functionality.