Tutorial at the 2025 Eleventh Indian Control Conference (ICC-11)
📍 IISc Bengaluru, India | 🗓 December 18, 2025
Dr. Kushal Chakrabarti
TCS Research, Mumbai
Dr. Dipankar Maity
UNC Charlotte, USA
Cyber-Physical Systems (CPS) are the backbone of modern technology — enabling autonomous vehicles, smart grids, and precision healthcare. However, as these systems become deeply networked, ensuring their security and privacy is critical.
This tutorial bridges control theory, cybersecurity, and privacy-preserving computation. It introduces both theoretical frameworks and emerging applications, preparing attendees to engage with the next generation of secure and intelligent control systems.
Understand the core trade-off between security and privacy in Cyber-Physical Systems. This session will teach you how to analyze this link in stochastic settings using noise, and in deterministic systems through the concept of opacity.
Learn how to protect distributed systems from eavesdropping. You will discover how to use "innovation signals" to secure control systems and analyze the key factors that safeguard client models in Federated Learning.
Learn how to address the critical privacy challenges of delivery drones. This talk presents a series of practical projects for ROS2-based systems, teaching you how to protect citizen privacy from video captured during autonomous navigation.
*Please note: The order and exact time-slots of the talks are preliminary. The final schedule will be confirmed with the official ICC 2025 program.
| Time | Session | Speaker |
|---|---|---|
| 14:00 – 14:05 | Welcome & Introduction | Dr. Kushal Chakrabarti |
| 14:05 – 14:50 | Security vs Privacy Tradeoffs | Dr. Vaibhav Katewa |
| 14:50 – 15:30 | Protecting Distributed Control Systems | Dr. Dipankar Maity |
| 15:30 – 16:00 | Tea Break ☕ | — |
| 16:00 – 16:35 | Protecting Distributed Learning Systems | Dr. Kushal Chakrabarti |
| 16:35 – 17:20 | Delivery Drones and Citizen Privacy | Dr. Vinod Ganapathy |
| 17:20 – 17:30 | Closing Remarks | Dr. Dipankar Maity |
Federated Learning secures CPS by enabling collaborative AI models to optimize physical processes without exposing critical operational data to attacks.
Designing resilient control algorithms for power grids is crucial to counteract false data injection attacks and ensure stable, uninterrupted operation.
Modern autonomous platforms employ privacy-preserving SLAM, encrypted mapping, and decentralized perception for responsible intelligence.