Device-Free, Radio-based Activity Recognition using Smart Home Wireless Communication Technologies

This dissertation demonstrates the use of received signal strength to infer human activities. It has the following contributions: 1) Reference design of a device-free, 2.4GHz IEEE 802.15.4-based sensor system for Activity Recognition; 2) Fundamental description of influences affecting Activity Recognition performance; 3) Software design pattern for device-free, radio-based Inference Systems; 4) Development and characterization of three specialized device-free, radio-based Inference Systems