Towards Ethically Aware Autonomous Systems: Enhancing Moral Reasoning in Robots

Institut
Professur für autonome Fahrzeugsysteme
Typ
Masterarbeit /
Inhalt
experimentell / theoretisch /  
Beschreibung

Background

As robots and autonomous systems become increasingly integrated into human environments, their ability to navigate ethically and respond to unforeseen circumstances becomes critical. This is particularly true for autonomous vehicles (AVs), where ethical decision-making extends beyond technical correctness to include moral reasoning, which allows these systems to navigate complex social interactions. One major challenge is the identification and resolution of "ethical dilemmas" and "anomalous situations" that arise during operation. For instance, Tesla vehicles have faced situations where unexpected behaviors, such as incorrect autopilot deactivation, highlight limitations in the semantic understanding capabilities of current systems. These issues call for a deeper integration of ethical principles into AVs, which are typically resolved intuitively by humans but require sophisticated reasoning from machines.

 

Research Focus

This master’s thesis will focus on advancing ethical decision-making for autonomous systems, specifically exploring how semantic understanding in AVs can be improved using ethical frameworks and large language models (LLMs). The objective is to build on the current limitations of AV systems, which struggle with complex, socially driven situations. By leveraging LLMs for context-based reasoning and decision-making, the project aims to enable AVs to detect and resolve "ethical anomalies" dynamically.

 

Objective

The successful candidate will engage in cutting-edge research aimed at developing a monitoring framework that integrates open-source LLMs with AV systems for ethical reasoning. The project will conduct experiments within a simulated environment (e.g., CARLA or similar), where the AV’s responses to ethically challenging situations will be studied and improved through prompt-tuning methods and scenario-based analysis. Ultimately, this research aims to advance AVs’ ethical reasoning by dynamically incorporating societal values and moral reasoning into real-time decision-making.

 

Your Benefits (What we offer):

  • Engage in a future-oriented, cutting-edge research field.
  • Work with state-of-the-art software development platforms.
  • Possibility of publication as a scientific paper if results meet criteria.
Voraussetzungen
  • A high degree of initiative and a creative mindset.
  • Excellent English skills and strong programming skills (Python or C++).
  • Foundational knowledge in machine learning.
  • Initial experience with autonomous vehicles or robotic systems is an advantage.
  • Experience in software development (Git, Ubuntu) is a plus.
Möglicher Beginn
sofort
Kontakt
Johannes Betz
johannes.betztum.de