A DATASET AT A GLANCE...

Security threats are very common in this era of internet and the automotive industry that is heading towards the development of autonomous vehicles has no way to exempt from them. It is because the autonomous vehicles have their microprocessors, controllers and other processing units communicating without a host computer via the Controller Area Network (CAN bus). This vehicle communication protocol, being developed in the 1980s by Bosch, is actually known for multi-master and error-free rapid communication between the vehicular units. However, the CAN bus allows any node to communicate with yet another node on becoming free, without any security checks like, access control, encryption or authentication. This can lead to vehicle manipulation by an eavesdropper to commit a crime like, vehicle theft, disable safety systems or cause an accident too. Denial-of-Service (DOS), spoofing and message replay are a few CAN bus attacks, which emphasize the need for an Intrusion Detection System (IDS) or anomaly detection system in real-time. Researchers aiming to identify an IDS for CAN bus attacks can make use of the CAN-MIRGU dataset from the UC Irvine Machine Learning Repository (https://archive.ics.uci.edu/dataset/1035/can-mirgu). This dataset encompasses a collection of CAN bus attacks from an autonomous vehicle, which can be used for developing a highly-secure in-vehicle network IDS. The dataset is actually introduced by Sampath Rajapaksha, Garikayi Madzudzo, Harsha Kalutarage, Andrei Petrovski and M.Omar Al-Kadri in the Symposium on Vehicles Security and Privacy (VehicleSec) 2024. Though CAN bus was initially developed to serve the automotive industry, it has also gained wide popularity in various industrial and healthcare automations, aerospace and defence sectors. So, the researchers availing the dataset have the chances to secure the automotive applications and extend their systems to other applications as well.