Architecting Safety Supervisors for High Levels of Automated Driving

The complexity of automated driving poses challenges for providing safety\nassurance. Focusing on the architecting of an Autonomous Driving Intelligence\n(ADI), i.e. the computational intelligence, sensors and communication needed\nfor high levels of automated driving, we investigate so called safety\nsupervisors that complement the nominal functionality. We present a problem\nformulation and a functional architecture of a fault-tolerant ADI that\nencompasses a nominal and a safety supervisor channel. We then discuss the\nsources of hazardous events, the division of responsibilities among the\nchannels, and when the supervisor should take over. We conclude with identified\ndirections for further work.\n