Sleep and activity monitoring for Returning Soldier Adjustment Assessment

This paper describes the development of unobtrusive room sensors to discover relationships between sleep quality and the clinical assessments of combat soldiers suffering from post-traumatic stress disorder (PTSD) and mild traumatic brain injury (TBI). We consider the use of a remote room sensor unit composed of a Doppler radar, light, sound and other room environment sensors. We also employ an actigraphy watch. We discuss sensor implementation, radar data analytics and preliminary results using real data from a Warrior Transition Battalion located in Fort Gordon, GA. Two radar analytical approaches are developed and compared against the actigraphy watch estimates - one, emphasizing system knowledge; and the other, clustering on several radar signal features. The radar analytic algorithms are able to estimate sleep periods, signal absence and restlessness in the bed. In our test cases, the radar estimates are shown to agree with the actigraphy watch. PTSD and mild-TBI soldiers do often show signs of sporadic and restless sleep. Ongoing research results are expected to provide further insight.