Direkt zum Inhalt springen
login.png Login    |
de | en
MyTUM-Portal
Technische Universität München

Technische Universität München

Sitemap > Schwarzes Brett > Abschlussarbeiten, Bachelor- und Masterarbeiten > Masterarbeit "Motion Planning for Autonomous Driving in critical street scenarios"
auf   Zurück zu  Nachrichten-Bereich    vorhergehendes   Browse in News  nächster    

Masterarbeit "Motion Planning for Autonomous Driving in critical street scenarios"

31.07.2017, Abschlussarbeiten, Bachelor- und Masterarbeiten

In cooperation with the Audi AG, the fortiss GmbH (An-Institut of the Technical University of Munich) develops new motion planning approaches for autonomous vehicles.

Autonomous Cars are supposed to react on various situations, which may include obstacles, traffic participants, street conditions and others. Covering all possible scenarios when developping autonomous driving functions is hard to achieve.

As part of this thesis, generic scenarios are supposed to be generated automatically on behalf of a definted criticality measure. A scenario consists of a semantic description of the scene, dynamic elements and both start and end pose of the vehicle.

First of all, street information need to be extracted from public street material. In order to decide for one line, we need a line center, for which an optimization approach should be used to generate it. The next step is to physically model other traffic participants to include them in the scenario to let the motion planner react on them. Various approaches on how to predict their motion into the future can be used. The extracted road information shall be used to generate new stochastic ones. Finally, a method shall be defined to quantify the criticality of a scenario.

Your Tasks:
- Literature search
- Extraction of street information from Open Street Maps (z.B. JOSM)
- Generation of a stret center line using optimization
- Probalistic modelling of other traffic participants
- Definition of a metric to describe criticality of the motion planning
- Generation of Scenarios (e.g. using Markov Decision Processes, Monte Carlo Simulation) based on the criticality measure, planning horizon, uncertainties)

Excited to contribute in the field of autonomous driving? Please send your CV, transcript of records and a short motivation letter to esterle@fortiss.org.

Kontakt: esterle@fortiss.org

Mehr Information

https://www6.in.tum.de/pub/Main/StudentProjects/20170724_esterle_scenarios_motionplanning.pdf

Termine heute

no events today.

Veranstaltungskalender