Aus Aifbportal
Wechseln zu:Navigation, Suche

Detecting Cooperative Driving in Deep Neural Network Architectures

Informationen zur Arbeit

Abschlussarbeitstyp: Bachelor, Master
Betreuer: J. Marius ZöllnerNikolai Polley
Forschungsgruppe: Angewandte Technisch-Kognitive Systeme

Archivierungsnummer: 4959
Abschlussarbeitsstatus: Offen
Beginn: 20. Oktober 2022
Abgabe: unbekannt

Weitere Informationen

A well functioning autonomous driving system needs to be able to anticipate the dynamic behavior of other traffic participants. With this knowledge, cooperative driving with non-autonomous cars should be achievable.

Recently, deep neural networks (transformers and graph-neural-networks) have been used on large-scale datasets to predict the future of other road users. In this thesis you will analyze these models to find interactive/cooperative behavior. Your task is to define a metric to determine if a given situation will result in interactive behavior between traffic participants.