MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641048002 A) filed by Vengatesan Krishnasamy; Dayananda Sagar University; R Dhanush; Kruthan B V; Elluri Sheshadri; Manoj MC; N. Bharathiraja; and Jeeva S, Bangalore, Karnataka, on April 15, for 'iot-based smart engine condition monitoring and fault detection system using machine learning.'
Inventor(s) include Vengatesan Krishnasamy; Dayananda Sagar University; R Dhanush; Kruthan B V; Elluri Sheshadri; Manoj MC; N. Bharathiraja; and Jeeva S.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "Conventional approaches to vehicle upkeep typically depend on fixed-interval servicing schedules instead of continuous real-time observation, often resulting in late identification of faults and heightened operational hazards. This paper introduces a smart engine health monitoring platform built on IoT infrastructure and enhanced with machine learning algorithms for classifying engine states in real time. The developed system captures key sensor readings such as engine RPM, lubricant oil pressure, fuel line pressure, coolant system pressure, lubricant oil temperature, and coolant fluid temperature. A dataset comprising 19,535 observations with binary labels indicating healthy and faulty engine states was utilized for experimentation. The data was partitioned into a 70-30 training-testing ratio for model validation. Results from the experiments reveal that the Long Short-Term Memory (LSTM) classifier attained an aggregate accuracy of 64.6%, exhibiting notably higher recall in identifying faulty engine behavior. The system integrates multi-sensor data capture via Arduino-based microcontrollers, cloud-based synchronization through Firebase Firestore, and supervised learning algorithms to identify abnormal engine states. A Flutter-based mobile application provides real-time vehicle health dashboards, trend visualization, predictive maintenance alerts, and automated service appointment scheduling. For anomaly identification, the system utilizes Isolation Forest, a well-established unsupervised learning method. These findings underscore the viability of coupling IoT-driven sensor networks with learning-based models for continuous engine fault identification and remote health assessment. The presented framework offers an extensible architecture for intelligent vehicular health tracking and prospective predictive maintenance upgrades."
Disclaimer: Curated by HT Syndication.
