Cybernet Systems Corporation, a leading innovator in technology, has announced that is has received a new patent, Prognostic maintenance system and method. This patent represents the company’s advances in artificial Intelligence, natural language parsing, learning, prediction and prognostication.

The patent describes an improvement on prognostic maintenance by taking historical maintenance information, then replacing or adding to that information for better maintenance-execution decisions based on automated machine learning. The preferred embodiment of the invention breaks down into two phases: the learning process and the execution process.

The learning process corrects and predicts likely maintenance activities or operations from previously stored maintenance records. Those records can include free-form and fixed-record format descriptions of the maintenance operations performed, diagnostic information indicating the maintenance type, and any direct fault measurements of that suggest necessary maintenance. The execution process applies the results learned to specific maintenance requests or work orders for a particular device or system. While the system will accept human-generated input, the preferred embodiment uses fully automated techniques to optimize final maintenance recommendations.

“We’re especially proud of this patent because it demonstrates a significant advancement in the function of existing computer hardware and ancillary equipment, and it advances multiple technologies and fields distinct from computer automation,” said Charles Jacobus, Cybernet’s chairman. “Compared to existing approaches, the most direct benefits come from perfecting device and system maintenance in producing work orders with much better accuracy and relevance.”

The patent also improves upon technologies apart from general-purpose computing, which include aircraft, vehicles, ships, as well as entire systems and subsystems, to determine the required maintenance or operation. Users can also apply Cybernet’s invention iteratively or recursively to any level of system partitioning, regardless of the application area.

The invention is not limited to isolated devices or systems, as it is also applicable to entire fleets or classes of vehicles and devices in determining overall fleet maintenance operations. For example, while users can employ the invention to a specific vehicle or device from a fleet to determine its required maintenance, it can also be used on subsystems within the device, such as the engine, transmission, software or electronic control.

Government records recognize the patent date as July 2, 2024, and list Cybernet’s Kevin Tang, Charles Jacobus and Charles Cohen as the inventors.