Intellectual Property India Publishes Patent Application for A Hybrid Deep-Learning And Knowledge-Graph-Based System For Intelligent Fault Localization And Root-Cause Analysis In Complex Electronic Systems With Explainable Report Generation And Continuous-Learning Feedback

MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641093053 A) filed by Dr. Asha. T; Arunkumar. A S; Dr. S. Mercy; Pavithra. N; Chamundeshwari. S; Ganavi. V S; and B. Padmavathy on July 31, 2026, for A Hybrid Deep-Learning And Knowledge-Graph-Based System For Intelligent Fault Localization And Root-Cause Analysis In Complex Electronic Systems With Explainable Report Generation And Continuous-Learning Feedback.

Inventors include Dr. Asha. T; Arunkumar. A S; Dr. S. Mercy; Pavithra. N; Chamundeshwari. S; Ganavi. V S; and B. Padmavathy.

The application for the patent was published on August 07, 2026, under issue no. 32/2026.

Abstract: ABSTRACT OF THE INVENTION Title: A HYBRID DEEP-LEARNING AND KNOWLEDGE-GRAPH-BASED SYSTEM FOR INTELLIGENT FAULT LOCALIZATION AND ROOT-CAUSE ANALYSIS IN COMPLEX ELECTRONIC SYSTEMS WITH EXPLAINABLE REPORT GENERATION AND CONTINUOUS-LEARNING FEEDBACK The invention relates to a hybrid deep-learning and knowledge-graph-based system for intelligent fault localization and root-cause analysis in complex electronic systems that overcomes the limitations of single-modality machine-learning fault classifiers and hand-curated symbolic troubleshooting systems. The system comprises an electrical-test-measurement collector (101), an event-log and error-code collector (102), a thermal-image and boundary-scan collector (103) and an operational-telemetry collector (104) which jointly collect multi-source diagnostic signals from an electronic system under test; a signal normalization and de-noising layer (105); a system-component knowledge graph (106) automatically populated from schematic, netlist, bill-of-materials and service-history corpora; a hybrid deep-learning subsystem comprising a convolutional-neural-network waveform encoder (107) and a graph-neural-network topology encoder (108); a knowledge-graph reasoning engine (109); a hybrid deep-learning-and-knowledge-graph fusion module (110); a service-history and bill-of-materials knowledge base (111); a root-cause hypothesis ranker (112); an explainable root-cause-analysis report generator (113); a technician diagnostic user interface (114) with guided diagnostic workflow; and a service-action audit and feedback ledger (115) that closes a continuous-learning feedback loop into the deep-learning encoders and into the knowledge graph. The framework is expressly a decision-support system for the human technician and is not a substitute for the technician's professional judgement, does not itself effect any physical repair, and is not to be operated in configurations that would bypass safety interlocks required by the applicable safety standards for the class of electronic system under diagnosis. Number of Claims: 5 Number of Drawings: 3

Disclaimer: Curated by HT Syndication.