FORENSIC: Fast and Autonomous Platform Anomalies detections in Cyber Physical Systems

Lead Participant: UNIVERSITY OF ESSEX

Abstract

**Market Need**

Critical systems, such as the power grid, autonomous transportation, industrial robots, and process control, are examples of cyber-physical systems (CPS). Recently, in the second half of 2022, there has been an 80% increase in cyberattacks on such devices. Most interestingly, in 54% of cases, these attacks occur due to unauthorised code or commands silently executed in the systems and compromising the sensor measurements used to control them. When such an attack occurs, the CPS sector faces a huge monetary loss, approximately £20K per minute. Nevertheless, industries still rely on traditional software-based intrusive security mechanisms to tackle attack scenarios. A recent survey found that the current cybersecurity strategy would likely be outdated in two years because the mitigation actions generated are usually generic and difficult to put into context for mission-critical CPS. Additionally, CPS devices are often deployed in remote or physically inaccessible locations. Applying security updates and patches to these devices can be challenging.

**Solution**

FORENSIC is a hardware-software co-design-based solution that rapidly and autonomously monitors the health of the systems by collecting low-level hardware features from the target device that are hard to compromise for threat detection. It utilises innovative technology to raise an alert for anomalies if it observes deviations in the system's health. In addition, it is quicker, cheaper, and less power-consuming than its rivals. The system detects any operational changes in the systems iteratively and autonomously when an attacker might reach the critical component of the system by employing novel innovative AI techniques. Compared to existing solutions, the unique selling point of FORENSIC is that runtime threat detection is based on robust hardware features that are difficult to compromise. Another unique feature of FORENSIC is that it does not rely on the modelling of the software applications running on the platforms and can also be easily calibrated and adapted to different execution platforms.

**Target Markets:**

* UK-based smart manufacturers and critical infrastructure providers
* Automotive industry and healthcare
* Security solution providers within Industry 4.0

Lead Participant

Project Cost

Grant Offer

UNIVERSITY OF ESSEX £59,784 £ 59,784

Publications

10 25 50