MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072838 A) filed by Vellore Institute Of Technology on June 12, 2026, for Hybrid Machine Learning And Fuzzy Logic Framework For Fault-Tolerant Wireless Sensor Networks With Self-Healing Routing.
Inventors include Santhosh Kumar Svn; and Devarla Manoj Kumar Reddy.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: ABSTRACT HYBRID MACHINE LEARNING AND FUZZY LOGIC FRAMEWORK FOR FAULT-TOLERANT WIRELESS SENSOR NETWORKS WITH SELF-HEALING ROUTING A system (100) for fault-tolerant, self-healing, and trust-aware operation of a cognitive wireless sensor network includes a Monte Carlo synthetic telemetry generation module (104) generating node-level telemetry records with overlapping class-conditional distributions and Gaussian noise injection. A LightGBM fault detection module (106) classifies nodes as normal or faulty using gradient boosting with leaf-wise tree growth. A topology construction module (402) builds a two-dimensional network topology with communication links and per-node states. A Mamdani fuzzy inference cluster-head election module (112) evaluates nodes using energy level, distance to base station, trust score, and neighbourhood degree, fuzzifying inputs, applying a fuzzy rule base, and performing centroid defuzzification to elect cluster heads. A Dyna-Q reinforcement learning self-healing routing module (114) learns adaptive routing policies through direct Q-updates and model-based planning. A composite trust scoring and intrusion detection module (116) computes trust scores, classifies attack types, and isolates flagged nodes.
Disclaimer: Curated by HT Syndication.
