Development of a Holistic System for Activity Classification Based on Multimodal Sensor Data
Tony Rolletschke, Rico Thomanek, Christian Roschke, Marc Ritter · Electronic workshops in computing · 2018
As the world of portable computers has evolved, mobile activity and mobility monitoring has become one of the major trends of recent years using widely used technologies such as smartphone sensors and wearables. These technologies are the basis for a wide range of applications in the areas of health monitoring, fitness games and telematics systems in vehicles. In the resulting use cases, the focus is on recognizing, differentiating and qualitatively evaluating different types of movements. The key factor in this context is a high degree of recognition accuracy in almost real time. Due to the ongoing development of mobile devices and the associated increase in performance, it is now possible to use the interfaces provided in mobile operating systems for the use of deep learning technologies. Due to the high availability of the end devices, new context-sensitive applications can be created, which can adapt the program logic to the current environment of a user.