Predictive Fault Detection for Unmanned Communications Facilities

Communications systems depend on the health of their equipment and components to perform reliably over extended periods of time. Maintaining the health of equipment in remote, unmanned communications terminals is critical, because unscheduled repairs have significant impact on system availability. In Phase I of this SBIR project Instrumental Sciences, Inc. (ISI) researched methods to improve the effectiveness of communications systems/facilities maintenance through development of new predictive fault detection (PFD) methods and architectures, employing multivariate statistical analysis, trending analysis, and Bayes cost decision algorithms to produce an innovative statistical analysis engine capable of processing a variety of input signal types. In the Phase II continuation, ISI will develop an engineering prototype PFD system, building on the successful Phase I results, integrate it with an IFICS Data Terminal (IDT) testbed, and conduct tests in the IDT environment. The ISI approach combines statistical signal processing, trend analysis, identification, and decision-theoretic techniques with model-based representations of measurable equipment properties that can yield clues concerning the condition of the covered equipment.