AI Boat Health Monitor[AI Health Monitoring System]

• A real-time marine vessel health monitor, processing GPS and IMU telemetry to detect early signs of mechanical, environmental, and navigational failures before they become critical
• Combines three complementary AI/ML approaches, Isolation Forest for impact detection, Linear Regression for engine health forecasting, and Rolling Z-Score for statistical monitoring covering six distinct sensor categories across the full vessel system
• A synthetic data generator simulates realistic trip conditions including injected failure scenarios such as GPS jamming, hull strikes, vessel instability, and engine cooling failure, providing both training data and a way to validate that every detector fires correctly
• The Streamlit dashboard presents all findings unified into a single alert table with severity labels and plain-English decision support, escalating to a full-screen flashing alarm with audio when anomaly count crosses a critical threshold
• The baseline model is trained offline once on clean data and saved to a ".joblib" bundle, keeping the live dashboard lightweight and focused entirely on inference

Built With

• machine-learning
• python
• scikit-learn
• streamlit