Protecting dynamic wireless charging systems from cyber attack with AI
A machine learning technique developed by a European collaboration can protect dynamic wireless charging systems from cyber attack.

Researchers in the UK have developed a technique to protect electric vehicles using dynamic wireless charging (DWC) from cyber attacks.

The team at De Montfort University in Leicester (DMU) propose using city buses to act as mobile charging hubs, using the energy from a focused magnetic field combined with stationary charging posts to power electric vehicles. It would also save money by removing the need for expensive new roads fitted with charging pads and infrastructure. However, to keep charging, an electric vehicle must transmit data regularly on its location and energy to the wireless network so they know when they are close to an available transmitter.

This real-time booking procedure uses a periodic beacon message, known as a Cooperative Awareness Message (CAM), to notify the system of the position of the vehicle. Every beacon message contains a Node Identifier, GPS coordinates, GPS speed, current Timestamp and MAC address of the vehicle. These messages are transmitted several times per second using Dedicated Short Range Communication (DSRC) and Wireless Access in Vehicular Environments (WAVE) technology, based on the IEEE 802.11p standard. However, these messages are vulnerable to a wide range of cyber threats, such as eavesdropping, spoofing and modification attacks

Working with University of Surrey, UK, and the University of Thessaly, Greece the researchers used machine learning to detect cyber attacks on the dynamic wireless charging system that can be applied to other such systems.

“Dynamic wireless charging is a technology with great potential that it combines high-tech communication between vehicles and state-of-the-art technologies for energy transfer, enabling vehicles to extend their travel time without the need for large batteries or extremely costly infrastructure,” said Dr  Leandros Maglaras at De Montfort.

The intrusion detection system (IDS) spots when fake location data is being received. Computer modelling of their system has shown a statistically significant improvement in detecting threats and they plan to test it on electric vehicles soon.

Protecting dynamic wireless charging systems from cyber attack with AI
A machine learning technique developed by a European collaboration can protect dynamic wireless charging systems from cyber attack.
falsification, and can detect spoofing attacks with more than 90% accuracy. The IDS uses a new metric, position verification using relative speed (PVRS). This compares the distance between two communicating nodes that is observed by On-Board Units (OBU) and their estimated distance using the relative speed value that is calculated using interchanged signals in the Physical (PHY) layer. The technique shows an increase in the detection accuracy by 6 percent, a change that Maglaras describes as a “very significant improvement”.

“A spoofing attack is one of the most dangerous attacks for route optimisation systems. This type of attack allows an attacker to spoof its real geographical position in the information sent within interchanged messages, making it appear that the vehicle is in another position,” said Maglaras. 

SMARTROAD STARTUP RAISES $50m

  • WORLD’S FIRST INDUCTIVE CHARGING SMART ROAD
  • MAGNETIC CONCRETE COULD MAKE WIRELESS EV CHARGING AFFORDABLE
  • Other articles on eeNews Power