SBIR-STTR Award

This proposal is fully responsive to DARPA topic SB052-012. It provides research toward an innovative non-intrusive health monitoring system for military personnel recovering from combat injuries. The proposed approach will collect and analyze real time data of vital signs, patient activity, fall acceleration and location parameters to detect deviations from learned norms, from modelled expectations, and to make predictive determinations. The research will be conducted with experts in military medical care and the VCU Medical/NASA biosensor team to evaluate the efficacy of intelligent software approaches to "trip-wire" monitoring of a recovering patient. The commercialized system will work across all echelons of military medical care from noisy evacuation transports to stateside medical facilities and even the soldier's home. It will meet HIPAA and DOD privacy and security guidelines. The technical vision includes small wireless non-intrusive wearable sensors of socially acceptable form factors, secure wireless networks, intelligent analysis software, displays for medical personnel, and interfaces to medical record systems. It will be a highly flexible and scalable standards-based platform to accomodate likely advances in miniaturized epidermal, implantable and swallowable medical sensors, patient activity sensing, and evaluative software.

Keywords:
Artificial Intelligence, Cognitive Systems, Non-Invasive Medical Monitoring, Uncertainty, Automatic Alerts, Wireless Networks, Vital Signs, Sensor Networks

AFrame's proposal to deliver a prototype system for Phase II will improve care for recovering warfighters and lead to easier transitions through the stepdown from hospital and rehabilitation settings to homecare and independence. Using advanced probabilistic reasoning, the system determines a patient’s vital sign baseline and sends intelligent alerts on important variations from that baseline during recovery. The system is being developed and will be field deployed in military treatment facilities and outpatient settings. It comprises a portable wearable monitor that transmits biometric and other relevant patient data wirelessly and securely via a reliable mesh network infrastructure to a central server for intelligent, predictive medical inference, alerting and reporting.

Keywords:
Automated Medical Condition Assessment, Cognitive Learning For Medical Decision Support, Intelligent Alarm Filtering, Bayesian Neural Nets, Predictive