CYBERNET SYSTEMS CORPORATION — Department of Defense SBIR Phase II: AF121-225

Description

Roughly 40% of USAF maintenance records have poorly coded work unit codes. This degrades aircraft pre-induction analysis and fleet-wide analysis of top drivers. Manual methods to scrub the data take years to develop for each air platform. Cybernet designed a machine learning algorithm to automatically correct coding errors without expert input. The Smart Data Cleanser (SDC) performed 4 years of work in 1 day for multiple aircraft platforms and outperformed human experts. It is expected to save $3M annually in unscheduled maintenance and free hundreds of expert hours to discover maintenance risks faster. The SDC has been successfully applied to the C-130, C-5M, F-15, and E-8C platforms, and now touches every plane arriving at Robins AFB. This effort will mature and transition the Smart Data Cleanser technology to the Air Force by (1) Completing integration of the SDC into the Aging Fleet Integrity Reliability and Maintainability (AFIRM) web system sustained by the 581st SMXS/MXDPCC at Robins AFB, (2) Developing high priority updates and enhancements to scale the SDC approach and boost scrubbing accuracy and coverage, and (3) Developing administrative tools and extensions to ensure the maintainability of the SDC application under System Program Offices (SPOs) and service branches.