Energy-Aware Environment Configuration in Simulation-based Testing of Autonomous Vehicles
Context
Autonomous vehicles (AVs) are complex cyberphysical systems that require extensive validation to ensure safety and reliability. Since real-world testing is expensive and potentially unsafe, simulation-based testing using platforms like CARLA has become a key component of AV software validation. These simulators reproduce realistic traffic scenarios, sensors, and environmental conditions, but they are computationally intensive and consume significant amounts of energy. As large-scale simulation campaigns become common (e.g., thousands of tests in continuous integration pipelines), improving the energy efficiency of simulation-based testing becomes increasingly important for sustainable software engineering.
Motivation
Most research in simulation-based testing focuses on improving failure detection, scenario generation, or test prioritization. However, little attention has been given to the energy cost of the simulation environment itself. Simulation platforms expose many configuration parameters—such as rendering quality, sensor frequency, number of actors, and weather complexity—that directly influence computational cost. Interestingly, changing these parameters does not always change the safety outcome of a test (e.g., collision or no collision). This raises an important question: Can we reduce simulation energy consumption by adjusting environment configurations without affecting the oracle (test verdict) results? Answering this question could enable more energy-efficient and cost-effective AV testing.
Goal
The goal of this project is to analyze how environment configuration parameters in CARLA influence energy consumption while preserving test correctness. The student will:
- Set up a baseline driving scenario in CARLA.
- Define oracle metrics (e.g., collisions, lane deviations, timeouts).
- Measure CPU energy consumption (and optionally GPU usage), execution time, FPS, and resource utilization.
- Systematically vary environment parameters such as:
- Rendering quality
- Sensor frequency
- Number of sensors
- Weather conditions
- Number of traffic participants
- Evaluate how these parameters influence total energy consumption.
- Verify whether oracle outcomes remain unchanged compared to the baseline configuration.
- Identify configurations that reduce energy while preserving test validity.
Deliverables:
- Scripts for automated experiment runs and energy/resource logging.
- A dataset of configuration vs. energy results.
- Plots and statistical analysis.
- Final seminar report.
- Mid and final presentation.
Requirements
- Programming skills in Python.
- Familiarity with Linux and basic system measurement tools.
- Interest in software testing, simulation, or performance analysis.
- Prior experience with CARLA is useful but not mandatory.
Pointers
- CARLA Simulator: https://carla.org/
- Birchler et al., “Cost-effective Simulation-based Test Selection in Self-driving Cars Software with SDC-Scissor”, ACM Transactions on Software Engineering and Methodology.
- Birchler et al., “A Roadmap for Simulation-Based Testing of Autonomous Cyber-Physical Systems: Challenges and Future Direction”, SANER 2022.
- Masoud Jamshidiyan Tehrani, Jinhan Kim, and Paolo Tonella. “PCLA: A Framework for Testing Autonomous Agents in the CARLA Simulator.” FSE 2025.