Maintainers Automated Troubleshooting Assistant (MATE)

ABSTRACT: Impact Technologies, LLC, proposes to develop and demonstrate an intelligent and systematic methodology to troubleshooting that aims to improve current practice and expedite the diagnosis/repair cycle. An expert decision support system will be developed to assist the maintainer in navigating through complex system interconnections while reducing the variability of coupled components/subsystems into a well understood series of steps. The enabling technologies borrow from Case Based Reasoning and Model Based Reasoning to integrate troubleshooting practice with technologies from Condition Based Maintenance. The vision is the development of a rigorous and automated diagnosis and troubleshooting system that will optimize maintenance, repair, and overhaul of complex assets. BENEFIT: In a typical military or industrial environment, complex systems (machines, aircraft, etc.) experience fault/failure modes that must be diagnosed accurately and rapidly in order to sustain a high level of operational availability. Considerable downtime translates into loss of productivity and increased maintenance costs. Current diagnostic and repair techniques rely on either evidence from a machine"s internal checks (fault indicator light or code) or an alert from the machine"s operator. An"expert"then observes the faulty system, determines the root cause of the problem, and composes a work order to schedule people, tools/equipment, or materials for repair and maintenance. If the problem is not diagnosed correctly and repaired, the troubleshooting and repair task is passed on to other technical personnel until a successful solution to the problem is reached. Unfortunately, these methods often do not offer sufficient information for the accurate and rapid diagnosis of the problem and rely heavily on the experience of the maintenance technician. Additionally, the formal and informal knowledge that the traditional troubleshooting approach relies upon is extremely difficult to maintain due to the transience of the workforce. Information management technologies, combined with proven techniques for knowledge extraction, evidence-based reasoning, and continuous learning offer a potential solution to this diagnostic dilemma. Impact Technologies, along with experts from the Georgia Institute of Technology, proposes to develop an intelligent and teachable system that learns through experience to provide expert decision support for field maintenance. This intelligent system will, over time, become the"learned expert,"or"esteemed colleague,"to the diagnostician. The proposed intelligent diagnostic expert will apply across mechanical and electronic complex system domains.