Intellectual Property India Publishes Patent Application for 'Development Of An Optimized Autoencoder System And Method For Detecting Faults Of Drilling Machines In Underground Coal Mines And Bauxite Mines' Filed by Indian Institute Of Technology (Indian School Of Mines)

MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202631026571 A) filed by Indian Institute Of Technology (Indian School Of Mines), Dhanbad, Jharkhand, on March 6, for 'development of an optimized autoencoder system and method for detecting faults of drilling machines in underground coal mines and bauxite mines.'

Inventor(s) include Bandita Sarkar; Kaushik Mazumdar; and Biswaranjan Das.

The application for the patent was published on March 13, under issue no. 11/2026.

According to the abstract released by the Intellectual Property India: "Development of an Optimized Autoencoder System and Method for Detecting Faults of Drilling Machines in Underground Coal Mines and Bauxite Mines The present disclosure relates to a fault detection system for a drilling machine based on graph-based machine learning and real-time sensor data analysis. The system comprises a sensor signal acquisition unit configured to receive operational signals from one or more sensors mounted on the drilling machine and to condition and convert the received signals into digitized operational signal data representing real-time machine operating conditions. A graph construction module transforms the digitized operational signal data into a graph data structure including nodes corresponding to sensor channels and edges representing inter-sensor relationships. The system further includes a graph attention autoencoder having a graph attention encoder configured to generate low-dimensional latent node embeddings through weighted aggregation of neighbouring node features, and a graph decoder configured to reconstruct the graph data structure from the latent node embeddings. A reconstruction deviation computation unit determines reconstruction error between the original and reconstructed graph structures, and a fault decision unit identifies a fault condition when the reconstruction error exceeds a predefined deviation threshold and generates control output signals indicative of the detected fault."

Disclaimer: Curated by HT Syndication.