Cyber-physical security for HVAC systems

PI: Burcu Akinci

Co-PI(s): Mario Berges

University: Carnegie Mellon University

Industry partner: Carnegie Bosch Institute/Bosch Research and Technology Center North America

Building automation systems are known to be one of the most vulnerable systems to cyber-attacks in the built environment, particularly in safety-critical infrastructure, enterprises, or research facilities. Especially in highly interconnected systems, these attacks can have severe impacts, and thus, should be detected quickly and reliably. In this project, we intend to focus on cyber-physical security for HVAC systems. Our objective is to develop an algorithm to detect cyber-attacks by leveraging the physical properties of HVAC systems, which are inherently given but rarely used for the detection of cyber-attacks. In addition, we plan to perform experiments on a real HVAC system at CMU. We expect to be able to detect cyber-attacks that are not possible to be detected reliably with existing IT-focused approaches. Our research would help raise awareness of the threat of cyber-attacks on HVAC systems, support the detection of attacks on CMU or PA enterprises, and in the long term, avoid unnecessary costs, reduce the risk of damages, and make systems more secure.