Recognition of Radio-Based Device Activity

Abstract

Identifying human actions in their everyday lives permits the development and widespread deployment of human-centric applications such as assisted living, health monitoring,  and so on. Outdated activity recognition methods depend on physical sensors (cameras, measuring devices, gyroscopes, and so on) to gather sensing element readings indefinitely and then apply pattern recognition algorithms to recognize user actions at an aggregate. Though ancient activity identification methods have proven to be useful in earlier research, they do raise some concerns about confidentiality, energy consumption, and training time. A completely new activity recognition approach has been planned in recent years, which takes use of body attenuation and/or channel weakening of wireless radio. In contrast to traditional activity recognition methods, radio-based methods make use of wireless transceivers in the environment as infrastructure, leverage radio communication characteristics to accomplish high recognition accurateness, save energy, and protect user confidentiality. We split radio-based activity recognition into four categories in this paper: ZigBee radio-based activity recognition, local area network radio-based activity recognition, RFID radio-based activity recognition, and other radio-based activity recognition. Every class is introduced and thoroughly reviewed, with some current additions. Then we compare a few example methods to point out their benefits and drawbacks. Finally, we will discuss some future research directions for this new research field.