Study on Simultaneous-Action Discrimination System Using the Neural Network

Autor: Masato Ito, Hatsuo Yamasaki, Takahiko Murayama, Tomoaki Nakano, Muneo Yamada
Rok vydání: 2019
Předmět:
Zdroj: Intelligent Transport Systems for Everyone’s Mobility ISBN: 9789811374333
DOI: 10.1007/978-981-13-7434-0_7
Popis: Along with the rapid spread of a smartphones, accidents caused by “simultaneous-walking” which is walking while operating a smartphone and “simultaneous-cycling” which is cycling while operating a smartphone has become a social problem. Our research group has been studying algorithms to detect simultaneous-walking and simultaneous-cycling using the smartphone itself, and proposed a method to switch each detection algorithm based on GPS speed. However, in an environment that could not obtain the GPS information, there were problems that cannot be used this conventional method and the detection accuracy changes greatly depending on the surrounding environment. In this study, our research group suggests a discrimination method of simultaneous-walking and simultaneous-cycling which is robust against the above problem by learning tri-axial acceleration sensor information of a smartphone with a neural network.
Databáze: OpenAIRE