A Data-driven, Falsification-based Model of Human Driver Behavior
This work addresses the need for more accurate modeling of human driver behavior in autonomous vehicle systems, though it appears incremental as it builds on existing methods like STL and falsification.
The paper tackled the problem of distinguishing human from non-human driver trajectories by constructing a data-driven boundary using parametric signal temporal logic, resulting in a classifier that separates admissible human examples from inadmissible ones generated through falsification.
We propose a novel framework to differentiate between vehicle trajectories originating from human and non-human drivers by constructing a data-driven boundary using parametric signal temporal logic (STL). Such construction allows us to evaluate the trajectories, detect rare-events, and reduce the uncertainty of driver behaviors when it assumes the form of a disturbance in control synthesis and evaluation problems. We train a classifier that separates admissible (i.e. human) examples - which arise from real-world demonstrations - and inadmissible (i.e. non-human) examples that are generated by falsifying specifications synthesized from the same real-world driving data. Proceeding in this fashion allows for finding a reasonable boundary of human behaviors exhibited in real-world driving records. The framework is demonstrated using a case study involving a human-driven vehicle approaching a signalized intersection.