SBIR/STTR Award attributes
Any transient high stress events can affect the ongoing health and safety of the system and require a system which can unobtrusively monitor the electronic health of the critical systems in real-time to prevent system failure or unsafe conditions. Nokomisrsquo; approach to identify functionality and health diagnostic data relies on identifying changes in unintended RF emissions and characterizing them using a hybrid of spectral quantification metrics and machine learning algorithms. nbsp;Nokomis proposes to create a system which can autonomously monitor system and subsystem health to identify possible critical failures or unsafe states and monitor electronics with no a priori knowledge to detect meaningful variations from emissions baseline indicative of critical states. This will allow for more efficient and direct reaction to critical states and allow for countermeasures to protect the system prior to system failure. The autonomous nature of the system addresses trends towards unsafe states without relying on operator response.Most electronic devices have precipitous changes in emissions when under a stress state indicative of system degradation which can rapidly lead to system faults. When the frequency spectrum data is processed via algorithms to quantify these changes, the signal extracted by analysis algorithms clearly differentiates between safe and stressed system states.nbsp; In this effort, Nokomis will develop a methodology to allow for the fault monitoring of electronics without a priori knowledge of the emissions of a standard condition. The proposed effort will autonomously discriminate these critical system faults from changes in emissions related to background changes, normal changes in operation, or expected normal wear. nbsp;
