A New Mapping Algorithm For Vehicle CAN BUS Mapping Based on Correlation Method

Abstract

Nowadays, advanced control systems help significantly improve the performance of vehicles from many perspectives such as safety, reliability, comfort, and so on. And those features require sophisticated controllers known as Electronic Control Unit (ECU). ECUs will electrically guarantee critical on-vehicle systems working correctly by communicating through different in-vehicle networks. Control Area Network (CAN) is the most popular in-vehicle network. A CAN mapper, which maps CAN messages with their corresponding source/destination ECUs, is considered as a useful tool for in-vechicle networks to help attack detection and aftermarket vehicle improvement, similar to Nmap for modern IP networks. The major challenge in developing such a tool is the broadcast nature of the CAN network where CAN messages have no transmitter information by design. Existing CAN mapper performs poorly when it is used for message source mapping over complex CANs due to the limitation of their hardware characteristic-based algorithm. Our goal is to develop a source mapper tool to organize CAN network messages by their control area with improved mapping accuracy in complicated network environment. Toward this goal, we propose Covariance, a new mapper algorithm which uses correlation information among CAN message timestamps to map in-vehicle networks. The covariance algorithm maps messages to source ECUs based not only on hardware characteristics but also on network function characteristics. That is why it will work better in a more complicated network environment than the previous Canvas source mapping algorithm which is only based hardware characteristics. We implement Covariance and test it over data collected from the Arduino emulator, dashboard emulator, manufacturing development bench, and testing vehicles by six data logging tools. Our new Covariance mapper tool could reach an average of 77.8% accuracy based on our testing results compared with an average of 51.9% of existing mappers. In addition to mapper algorithm design and development, this thesis also contributes to the setup of ECU information database, including message information and vehicle information affiliated, from some current Stellantis model vehicles.