Open Autonomy Kernel (OAK), an Architecture for Mode-Identification and Reconfiguration of Control systems Model-Based
Model-based reasoning systems have been relying upon a single model on which a single reasoning engine operates. A single inference engine physically removes the intelligence from the control system when used to control geographically large systems such as ships, computer networks, power grids and transportation networks. This has the drawback of requiring the transmission of raw data to the inference engine which increases the bandwidth requirements, reduces survivability and limits scalability.
The Johns Hopkins University Applied Physics Lab has developed architecture that is unique as it combines model-based reasoning and autonomous agents to provide for distributed, intelligent estimation and reconfiguration. The OAK architecture provides a goal-directed coordinated intelligence for overall engineering plant management. OAK provides intelligent support for the three features of system control: estimation, communication, and control. A set of local intelligent controllers manages individual system components. Although existing individual subsystem controllers may exhibit varying levels of sophistication and capability for autonomous operation, a common level of functionality is required for coordination of distributed intelligence in the system.
Patent Status: U.S. patent(s) 7574497 issued.
*JHU/APL is seeking an exclusive licensee and development partner for this technology. PatentCONTACT:
Mr. M. T. Hickman