Skip to content
Technical University of Munich

Practicum: Scenario-Based Testing of Cyber-Physical Vehicles

Description

Scenario-based testing is a methodology that is used to verify the behavior of autonomous cyber-physical systems by searching for unwanted corner cases. The methodology has been used to test autonomous cars and, more recently, drones and drone swarms for unsafe behavior. We can also apply the process to decentralized UAV swarms, which are designed to be robust to failures and scalable (i.e., capable of operating for a wide range of swarm sizes).

This practical course has students working in groups and teaches students how to use scenario-based testing, specifically by using discrete-event simulation, models of a cyber-physical system, and metaheuristic search methods. Then, the students will need to create their own strategy to generate test cases, with the goal of generating challenging test cases in diverse scenarios while also ensuring optimal efficiency in the process.

Students will test autonomous cars or decentralized UAV swarms in their respective simulators. 

Key Topics:

  • Discrete-event simulations.
  • Metaheuristic search methods.
  • Scenario-based testing: combining the above into a process for finding and executing challenging test cases.

Objective:

  • Learn about the scenario-based testing framework and the methodologies that make scenario-based testing possible.
  • Design and implement an automated test case generation tool.
  • Show the effectiveness and efficiency of the implemented tool by creating challenging test cases with as few simulation executions as possible.

Previous Knowledge Expected:

  • Prerequisite courses:
    • Praktikum Grundlagen der Programmierung (IN0002)
    • Programming experience (intermediate); mostly in Python.

Beneficial:

  • Advanced Topics of Software Testing (IN2084)
  • Some familiarity with discrete-event simulations and MATLAB

Teaching and learning method:

  • The course will start with a set of lectures; afterwards, the students will work in groups and meet with their supervisor on a recurring basis.
  • Students will integrate existing tools, simulation environments, and models of systems.
  • Each group will produce a final report and a final presentation.

Preliminary Meeting

For the 2025-2026 winter semester, the preliminary meeting will take place on 17. July at 15:00 via Zoom. A Zoom link will be added here shortly before the meeting time.

Registration

Registration for this course is handled by the TUM Matching Platform (see the deadlines here)

To top