Humboldt-Universität zu Berlin - Mathematisch-Naturwissenschaftliche Fakultät - Institut für Informatik

Eine Zoom-Einladung finden Sie hier. (nur mit Informatik-Account)

Zusammenfassung:
With the increasing autonomy of vehicles in road traffic, attention must be paid to safety in automated driving systems. Runtime monitoring of temporal logic is a technique suitable to analyze whether a system acts as specified during operation. However, current techniques for runtime monitoring lack the support for observing the movement of geometric objects under uncertainty. Therefore, this thesis presents how to perform temporal monitoring of probabilistic spatial relations. First, uncertainty is modeled into geometric objects and spatio-topological relations are adapted to accept such probabilistic geometric objects. Then, this thesis presents Temporal Probabilistic Spatial Logic (TPSL), which is based on Metric Interval Temporal Logic (MITL) and supports probabilistic spatial relations nested within temporal operators. Experiments are performed using autonomous driving simulator CARLA to highlight usability of TPSL. In the end, monitoring such TPSL expressions allows to detect possibly harmful movement and to react accordingly during driving.