AI based Predictive Maintenance for Vehicle [Engine Fault Detection]

The AI-based predictive maintenance system uses sound analysis to monitor vehicle engine health and detect anomalies before a major failure occurs. The process begins by capturing real-time acoustic data from a car engine or working motor using a microphone. This audio stream is fed into a machine learning model created via Google Teachable Machine and exported using TensorFlow.js, which runs inside a laptop-hosted HTML browser interface. The browser interface processes the audio features, classifying the sound patterns into normal operation or specific fault states. When the model detects an abnormal engine signature, it communicates alert signals to the IndusBoard Coin development board. The IndusBoard Coin processes these incoming signals to drive interactive indicators, allowing operators to visualize the system status through LED matrices or display components. By bridging web-based machine learning with microcontroller hardware, the system creates a continuous, automated feedback loop that translates complex acoustic data into actionable, visual maintenance warnings.