Modeling a Heavy-duty Vehicle Data Collection Process for Authenticating Driver Identity and Analyzing Driver Behavior Under Duress

Heavy-duty vehicle security is critical to safeguarding the supply chain of essential goods and materials. Implementing a continuous driver authentication system based on unique driving behaviors enhances security through improved access control without the use of personally identifiable information (PII). This ensures that only authorized drivers operate the vehicle and reduces the risk of unauthorized use. Designing this complex system to reduce this risk can be done effectively through SysML modeling and MagicGrid systems modeling. Through the use of systems modeling, this study was split into two main objectives. The first objective is to develop this driver authentication system by analyzing behavior patterns through three sets of vehicle sensor data: CANbus, GPS, and IMU data. These collect extensive vehicle data like vehicle speed, engine speed, pedal positions, location, and heading degree. This data is then used to train several different machine learning models to create a driver profile report for identity verification. The second objective is to run cyberattacks on the vehicle and analyze the driver’s behavior with physiological data like heart rate variability and with the vehicle sensor data. The cyberattacks cause a malfunction on the vehicle’s instrument cluster by sending various messages to the CANbus. This is used to create a report on how drivers respond in high pressure conditions, which can be used to train drivers transporting hazardous material to react properly in these situations. The purpose of this paper is to show how the application of the MagicGrid method and SysML modeling improved the development of the stakeholder needs, requirements, use cases, and system structure.