Through the wall human movement detection with software defined radio

In this thesis, machine learning models for the detection of human movement behind the wall in different scenarios using electromagnetic waves for wireless communication in the environment are investigated. Software defined radio is used for monitoring and collecting communications signals. There are three main situations in the monitoring area which is empty (no living), and the person is inactive asleep or unconscious (only breathing), and the person is moving (walking etc.). Respiratory function is one of the micro-level movements of living things when they are inactive (sleep, unconscious, etc.). In this thesis, the data collected using a setup, which can sense these micro motions behind the wall, using the electromagnetic waves. It is observed that the amplitude of the received radio signal are changing periodically even with human respiration movement. In the used setup, 900 MHz carrier frequency signals with limited output power were recorded for three defined basic situations with software-based radios (USRP B210) for 10 different states at different distances. In order to distinguish 10 different states, different characteristics were defined in the collected signals and trained and tested with 22 different classification techniques and the most accurate and efficient model is presented. Keywords: Through the Wall, Human Movement Detection, Software Defined Radio (SDR), Machine Learning, Classification