New defence system for transport smart-tech could save lives

A new system for detecting unusual and potentially harmful data sharing between in-vehicle smart-tech and traffic management technology could prevent 9 out of 10 cyber-attacks.   

Special sensors in much of modern transport act like the eyes and ears of the vehicle, helping it to see what's around it, know how fast it's going, and even ‘talk’ to traffic lights. 

But, in connecting vehicles to traffic management technology like satellite navigation systems, and with more data transferring wirelessly over 4G and 5G networks and cloud-based systems, exposure to malicious interference widens and the risk of cyber-attacks increases. 

A team of experts, led by Dr Zahra Pooranian from the University of Reading's Department of Computer Science, developed the new system, named LFD-IDS (Label Flipping against deep learning intrusion detection system). And now, the team’s development research has been published in IEEE Xplore.

Dr Rahim Taheri, a co-author from the University of Portsmouth’s School of Computing, said: “Strong security measures must be implemented to protect connected vehicle systems. Hackers can create anomalies by altering the values of the data being shared between sensors, in other words ‘flipping the labels’ of that data.  

“Our system specifically targets data transmissions to decipher the information being received and identify any abnormalities so that blockers can be put in place before any harm can be done.”

While the use of smart technology in vehicles and transport systems is helping to make road travel safer and traffic to flow more smoothly, Dr Taheri and his colleagues warn that strong security measures must be introduced to protect connected vehicle smart features from cyber-attacks and to preserve the safety of drivers, passengers, other road users, and the environment. 

The new system works by using artificial intelligence (AI) to help cloud-based systems tell the difference between genuine information i.e. data, and the fake information originating from hackers. 

The focus is on monitoring and analysing sensor data from vehicles - data which includes temperature, tyre pressure, and location information - and incorporating machine learning to classify that data based on learned patterns of sensor behaviour. Inaccurate information that could lead to a vehicle or transport system operating dangerously is identified and the threat to safety avoided. 

“There has been significant scientific and commercial interest in finding cloud-based solutions to keep up with advancements in 5G technology,” added Dr Taheri. “Traditional intrusion detection methods face significant challenges in identifying new and previously unseen attacks. However, by integrating machine learning capabilities in intrusion detection systems, we have shown that it is possible to substantially improve the accuracy in detecting threats.”

Z. Pooranian, R. Taheri and F. Martinelli, "LFD-IDS: Bagging-Based Data Poisoning Attacks Against Cyberattack Detection in Connected Vehicle," in IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 10, pp. 16800-16810, Oct. 2025, doi: 10.1109/TITS.2025.3581102.
keywords: {Cloud computing;Training;Temperature sensors;Cyberattack;Computer architecture;Accuracy;Connected vehicles;Training data;Temperature measurement;Monitoring;Connected vehicle;machine learning;adversarial attacks;cyberattack;intrusion detection systems (IDS)},