Cyber Security of Physical Systems

Predictive Cyber Protection, based on software developed at the Army Research Lab, assesses risk on six metrics: 1) vulnerability, 2) accessibility, 3) resilience, 4) control, 5) effect on the overall system, and 6) recoverability. The risk assessment scores for each of the 6 parameters are plotted on a graph, forming a visually ingestible polygon, of which the surface area is geometrically determined and used as an overall risk score for each asset in the network. The software is particularly adept at assessing the effects of an attack on one asset on the surrounding assets. The software also offers continuous monitoring of the entire network, thereby providing continued, real-time security. This long-term protection is facilitated by our software’s capacity to translate various types of data, as gathered by the various sensors that transmit from each of the network’s assets, into a common language. This translating feature allows the software to be implemented in a wide range of contexts and can also be used as a bridge between contexts. Finally, we have a yet unintegrated machine learning system that will provide not only increased accuracy within the diagnoses, but will also predict changes in the system’s risk portfolio in real time.