IIT Gandhinagar students develop AI system to protect critical establishments from cyber attacks

Two undergraduate students of the Indian Institute of Technology (IIT) Gandhinagar have developed an artificial intelligence-based cyber security system to detect cyber attacks on critical establishments, such as water treatment plants, power grids and manufacturing facilities.

IIT Gandhinagar students develop AI system to protect critical establishments from cyber attacksTheir system, called "TrustGate", combines network traffic and physical sensor data to identify attacks, determine the equipment affected and provide real-time information to plant operators.

Paper leak: NTA debars Latur coaching centre owner's son from NEET-UG exam for 2 years The students won the second prize at the Intel Cup, the international Undergraduate Electronic Design Contest 2026, held at Shanghai Jiao Tong University in China, according to a Press Information Bureau (PIB) release issued on Wednesday.

Third-year computer science and engineering student Miloni Thakkar and electrical engineering student Mitwa Goswami developed the system to address a gap in conventional industrial control system security tools. Oxford English Dictionary adds several words long used in the Caribbean and Bermuda Such tools generally monitor network traffic and can detect unauthorised access, unusual activity or known malware.

However, they may not be able to determine whether a command that appears legitimate at the network level could have dangerous physical consequences for a plant. "TrustGate closes that gap by understanding both the cyber and physical sides of an attack together," Thakkar said.

The project was partly inspired by a 2021 cyber attack on a water treatment plant in Oldsmar, Florida, where an unidentified attacker remotely accessed the control system and attempted to raise the level of sodium hydroxide to nearly 100 times its safe concentration.

Protesting students should be allowed to sit in competitive exams after certain period: UP CM Yogi Adityanath The attack was not detected by an automated security system and was noticed only after a plant operator saw the computer mouse moving on its own and reversed the command.

TrustGate uses a dual-stream Bi-directional Long Short-Term Memory (BiLSTM) model combined with "cross-modal attention". It analyses network traffic and physical sensor readings as two separate streams and allows them to inform each other.

This enables the system to detect attacks that may not appear clearly suspicious in either network data or physical sensor readings alone, including gradual manipulations designed to avoid triggering an alarm. When an anomaly is detected, TrustGate identifies the likely nature of the attack and the specific physical component affected and displays the information on a live map of the plant.

The students used the SWaT (Secure Water Treatment) dataset for testing. It is based on a scaled-down six-stage water treatment testbed maintained by the iTrust Centre at the Singapore University of Technology and Design, and is widely used in industrial control system security research.

Since critical infrastructure may be kept offline for safety reasons, TrustGate has been designed to operate entirely on embedded hardware rather than relying on cloud connectivity. The system uses TPM 2.0 to verify the integrity of its code and hardware at boot and has a self-healing watchdog that automatically restarts a core process if it hangs or crashes.

Goswami led the hardware integration, TPM attestation and embedded Linux deployment, while Thakkar led the AI model architecture, training, testing and OpenVINO deployment.