Generalized Anomaly Detection and Predictive Analytics

Project Overview

This research project focused on Generalized Anomaly Detection and Predictive Analytics. It was sponsored by the Research and Innovation division of Schneider Electric, Hyderabad.

Funding Details

• Funding Agency: Research and Innovation, Schneider Electric, Hyderabad
• Duration: 2022-2023
• Amount: Rs. Five Lakhs

Research Team

• Principal Investigator: Dr. Manohara Pai M.M
• Co-Investigators:
  • Dr. Ajitha
  • Dr. Veena Mayya
  • Dr. Radhika M Pai
• Students: Five students (CSE, ICT, ECE)

Technology Implementation

This research focused on developing robust methodologies for:

• Multivariate Anomaly Detection across diverse industrial systems
• Transfer Learning for cross-domain application of detection models
• Time-Series Prediction for forecasting system behavior
• Explainable AI Techniques for interpretable results

Technical Innovations

The project delivered several significant advancements in anomaly detection.

Industry Applications

This research provided valuable implementations for various industrial systems:

• Manufacturing - Early detection of equipment degradation
• Energy Systems - Identification of inefficiencies and faults
• Supply Chain - Detection of operational irregularities
• Quality Control - Automated identification of defective products
• Cybersecurity - Detection of unusual system behavior patterns

The algorithms developed through this project have demonstrated significant improvements in fault detection accuracy, reduction in false positives, and early warning capabilities, representing a valuable advancement for industrial monitoring and maintenance strategies.