Formal Method and Autonomous Driving

Our research group focuses on the practical application of formal methods to next-generation automotive systems, particularly autonomous driving systems. Our goal is to contribute to a safer, more reliable mobility society by setting new standards in the safety verification of these systems. Unlike existing research, which often addresses limited aspects or simplifies target systems, our comprehensive solutions target real-world autonomous systems, aiming to deliver not only robust methodologies but also practical frameworks and verification tools to be used in the autonomous driving sector.

A Runtime Verification Framework for Autonomous Driving Systems

We propose a framework for runtime verification of ADSs, comprising a scenario description language, a runtime monitor, and an LTL property checker.

Autoware Safety Analysis in cutin, cutout, and deceleration scenarios

We evaluate the safety of Autoware, a leading ADS platform, in the three traffic disturbance scenarios, leveraging our Runtime Verification Framework.