Sensor Data Fusion for Intelligent Systems Monitoring and Decision Making

Award Information
Agency:
Department of Defense
Branch
n/a
Amount:
$737,053.00
Award Year:
2013
Program:
SBIR
Phase:
Phase II
Contract:
FA8571-13-C-0010
Agency Tracking Number:
F103-255-1119
Solicitation Year:
2010
Solicitation Topic Code:
AF103-255
Solicitation Number:
2010.2
Small Business Information
Data Fusion & Neural Networks, LLC
1643 Hemlock Wy, Broomfield, CO, -
Hubzone Owned:
N
Minority Owned:
N
Woman Owned:
N
Duns:
130770055
Principal Investigator:
Christopher Bowman
Owner
(303) 469-9828
cbowman@df-nn.com
Business Contact:
Christopher Bowman
Owner
(303) 469-9828
cbowman@df-nn.com
Research Institution:
n/a
Abstract
ABSTRACT: The current AF aircraft maintenance system has a high operational cost of unplanned, premature, and missed maintenance. The"red"operational anomaly detection thresholds are not adaptive to individual aircraft history nor to changes in operations causing"ringers"and missed detections. Abnormal correlations in aircraft health sensor data that can be precursors to catastrophic failures are not automatically looked for. The proposed Data-Driven Condition-Based Predictive Maintenance (DCPM) system will be developed based upon a proven data-driven abnormality detection system, ANOM, for spacecraft that is operational at Schriever AFB and at TRL 6 at National site. This system learns normal and detects unknown-signature abnormal behavior for thousands of variables in real-time. ANOM provides context for abnormal clusters with characterizations to which recommended responses can be tagged for historically detected abnormal behaviors. ANOM is affordably adaptive via retraining for changing operational states with automated cluster specified new training sets over time. The abnormality detections from each source will be fused for each aircraft and across aircraft using our Bayesian Fusion Node (BFN) proven software. The BFN will also be used to detect and track relationships amongst these events across the fleet. Our Performance Assessment and Process Management software will evaluate DCPM performance. BENEFIT: The key competitive advantages of the technology in this proposal are affordability, effectiveness and rapid deployment all made possible via a data-driven approach. DF & NN"s Data-Driven Condition-Based Predictive Maintenance (DCPM) technology leverages pattern learning software to automatically discover, detect and characterize unknown abnormal signatures. The technology is extendible and reusable because it is based on the generic Dual Node Network (DNN) DF & RM technical architecture. The data-driven core of the DF & NN ANOM software enables it to be easily applied to detect, recognize, and track abnormalities in any commercial or government system. Operational prototypes of this capability are already operating on-line at to 2 satellite operations (SOPS) sites and these ANOM tools have been applied off-line to over 100 varied and massive data sets for over 200 combined years of data. DF & NN has a pending patent on the ANOM technology and plans to commercialize ANOM to support many DOD and commercial systems in collaboration with our Commercialization Pilot Program team member Lockheed Martin Corporation (LMC) for Space Operations Squadrons (SOPS), cyber, Remotely Piloted Aircraft, and other LMC products.

* information listed above is at the time of submission.

Agency Micro-sites


SBA logo

Department of Agriculture logo

Department of Commerce logo

Department of Defense logo

Department of Education logo

Department of Energy logo

Department of Health and Human Services logo

Department of Homeland Security logo

Department of Transportation logo

Enviromental Protection Agency logo

National Aeronautics and Space Administration logo

National Science Foundation logo
US Flag An Official Website of the United States Government