WEBINAR: Can we prevent industrial robots going rogue?

IMarEST TV recording by the South East England & London Home Counties Branch on the 20th April 2023. This webinar was presented by Professor Jonathon Loo of the University of West London.

Industrial robots are widely integrated within a network ecosystem for remote interaction, leaving robots subject to cyber-physical attacks, due to vulnerabilities in the network, Robotic Operating System (ROS) and human factors.

Trend Micro, a cyber security software company, found 83,000 robots are exposed on the Internet. Robotic security comes in two forms. The first line of defence is front-end prevention, such as security hardening. It helps, but it isn’t a reliable solution on its own as hackers will always manage to find a way in. Existing solutions are built into the system, scanning for vulnerabilities and attempting to prevent front-line security breaches, such as system penetration. But what happens when this fails or robot controllers are compromised undetected, meaning hackers can take control and launch attacks on robotic operations?

That’s where ROS-PCon (InnovateUK funded PoC) comes in. Standing at the second line of defence, it learns the robotic movement through the robot execution feedback based on deep learning technology while the robot is performing its role in a controlled, best-case environment. It can then highlight anomalies when it spots something that deviates from the control, which would be otherwise invisible to the human eye.

About the Speaker

Professor Jonathan Loo is currently a Chair in Computing and Communication Engineering at the University of West London. He has over 20 years of experience in research, graduated over 20 PhD students and co-authored over 400 academic papers. He recently won InnovateUK grant for developing a proof of concept for robot security surveillance called ROS-PCON (robotic operating system – preventive control).

Watch Now (IMarEST Members only)