Invention Grant
- Patent Title: Hybrid feature-driven learning system for abnormality detection and localization
-
Application No.: US16138408Application Date: 2018-09-21
-
Publication No.: US11146579B2Publication Date: 2021-10-12
- Inventor: Masoud Abbaszadeh , Fernando D'Amato
- Applicant: General Electric Company
- Applicant Address: US NY Schenectady
- Assignee: General Electric Company
- Current Assignee: General Electric Company
- Current Assignee Address: US NY Schenectady
- Agency: Buckley, Maschoff & Talwalkar LLC
- Main IPC: H04L29/06
- IPC: H04L29/06 ; G06F17/18 ; H04L12/26 ; G06N20/00

Abstract:
A cyber-physical system may have a plurality of monitoring nodes each generating a series of current monitoring node values over time representing current operation of the system. A data-driven features extraction computer platform may receive the series of current monitoring node values and generate current data-driven feature vectors based on the series of current monitoring node values. A residual features extraction computer platform may receive the series of current monitoring node values, execute a system model and utilize a stochastic filter to determine current residual values, and generate current residual-driven feature vectors based on the current residual values. An abnormal detection platform may then receive the current data-driven and residual-driven feature vectors and compare the current data-driven and residual-driven feature vectors with at least one decision boundary associated with an abnormal detection model. An abnormal alert signal may then be transmitted when appropriate based on a result of said comparison.
Public/Granted literature
- US20200099707A1 HYBRID FEATURE-DRIVEN LEARNING SYSTEM FOR ABNORMALITY DETECTION AND LOCALIZATION Public/Granted day:2020-03-26
Information query