SBIR-STTR Award

Human-based expertise to diagnose and repair electromechanical hardware systems is gained only through lengthy experience, but this expertise vanishes with the individual. This research explores methods to capture and embed this expertise to provide hardware self-diagnostics through the use of alterable on-component micro-data bases stored on an add-on "data template." This data template contains electronically addressable (read/write) data on the malfunction history implanted directly onto thephysical component-analogous to a patient's medical records. These on-component data are useable by human diagnosticians, but are intended as a knowledge-base for an embedded expert system to provide decision metrics for hardware self diagnostics. Several technologies exist to implement this concept, ranging from bar coding to add-on or embedded microsensor circuits. The data template has significant potential for fault isolation, diagnostic training, documentation, inventory control, reliability assessment, and parts anti-counterfeiting. Phase I research will establish feasibility of the data template concept, on which a base-line model and specifications for an expert self-diagnostic system will be evolved. Phase II will develop a prototype of the data template for links to the expert system defined in Phase I.The potential commercial application as described by the awardee: The proposed add-on data template can be used to help isolate faults in hardware components, either as a passive unit, or as an active device linked to the hardware's built-in-test circuitry. Such application will greatly enhance self-diagnoses and repair of hardware. The memory element of the template will serve as a distributed data base, linkable to an external diagnostic expert system. In this mode, the template would serve as a diagnostic link between hardware and a human repairman. Conversely, human expertise in specific diagnostics data could be retained in the template memory on the hardware component. By also retaining its maintenance history, the template would greatly aid reliability assessment and inventory control by being able to isolate component-unique factors related to equipment failure. Due to the potential for electronic encoding of the template, which could be integrated into the physical structure of hardware components, anti-counterfeiting and subsequent validation of critical parts could be accomplished with relative ease.

The objective of this research is to determine the feasibility of a method to embed diagnostic expertise within hardware systems to achieve a self-diagnostic capability. For this purpose, a self-contained electronic device is proposed which serves to link hardware, diagnostic information, expert diagnostic systems, human diagnosticians. This link is called the Adaptive Memory Module/Template (AMMT), which is an intelligent software-controlled information logging device to be eventually packaged as a microchip. This AMMT device is to be embedded into a host component as a module, or onto the component as a template. The AMMT is externally powered and interrogated through a radio-frequency probe. It has a battery-backed, non-volatile, crash-proof memory which is capable of data retention for ten or more years. As an option, electronic links between the AMMT and the host component may serve to continuously record host component reliability parameters, such as electrical pulsing, temperature, vibration, humidity, etc., through embedded microsensors. Based on the Phase I research, the AMMT concept is highly feasible with existing technology. This feasibility is significantly enhanced with newer, more comprehensive electronic units. Also, the AMMT has potential commercial applications extending beyond hardware maintenance to such areas as biotechnology, manufacturing, and robotics.