Two Master's Theses on AI-Driven and Automated Testing for Automotive Software

About this opportunity - Background

Innovation starts with being curious. At Volvo Cars, we believe that being curious and truly committed to understanding people is the key to future success. We are looking for driven and motivated students interested in pursuing their master thesis within Vehicle Engineering during the spring term 2026.

At Vehicle Engineering, our responsibility spans the entire chassis, interior, climate, body & exterior functions, mechanical Integration and a range of vehicle attributes with impact on design and customer experience of our complete product. Across these domains and all development cycles we will lead an effective delivery operation on complete level in close sync with other Engineering units, functions and stakeholders across the company.

Are you interested in shaping the future of automotive software quality?

The thesis works described below are within the area of Exterior Functions Software Development.

We are offering two master’s thesis projects that explore cutting-edge methods for automated testing in virtualized automotive environments. The first thesis focuses on automatic generation of vTESTstudio test cases from natural language requirements using large language models (LLMs), enabling AI-assisted test development and requirement interpretation. The second thesis investigates fuzzing-based automated testing and continuous integration to improve reliability and robustness in virtualized automotive systems. Both topics offer hands-on experience with state-of-the-art tools and workflows used in the industry and provide a strong foundation for a career in automotive software engineering and testing. In your application, please specify which of the projects you are applying for (you can express interest in one or both projects)

Thesis #1

Automatic Generation of vTESTstudio Test Cases from Natural Language Requirements Using Large Language Models

In the automotive industry, the complexity of Electronic Control Units (ECUs) and distributed systems has grown dramatically. To ensure software reliability and compliance with safety standards such as ISO 26262, companies use rigorous testing workflows supported by Vector’s vTESTstudio and CANoe environments. With recent advances in Natural Language Processing (NLP) and Large Language Models (LLMs) such as GPT-5 and Llama 3, it has become feasible to automatically interpret technical requirements and generate structured artifacts, such as test templates or scripts. This opens an opportunity to accelerate the test design process while improving consistency and traceability.

1.1. Problem Statement:
Manual translation of natural-language requirements into test case designs for vTESTstudio:
•    consumes a large share of development time,
•    introduces inconsistencies between requirement intent and implementation, and
•    limits reuse and automation potential.

There is currently no AI-based system capable of reading automotive requirements and generating vTESTstudio-compatible test case skeletons automatically.

1.2. Research Objectives:
The goal of this thesis is to design, implement, and evaluate an AI-driven pipeline that automatically generates executable test case templates from natural-language requirements.

1.3 Main Objectives:
1.    Develop an NLP model that interprets requirement text and identifies key test elements (signals, preconditions, triggers, expected outcomes).
2.    Translate these extracted elements into structured test case templates compatible with vTESTstudio (e.g., .vtst, .vtestunit, or CAPL).
3.    Evaluate the generated test cases in terms of accuracy, completeness, and engineering effort saved.
4.    Demonstrate integration feasibility within the existing Vector toolchain (requirements tool → vTESTstudio)

1.4 Qualifications:
•    Master’s in Computer Science, Software Engineering, Artificial Intelligence, or Automotive Engineering
•    Specialization or coursework in Machine Learning, Natural Language Processing, or Model-Based Testing
•    Python Programming, NLP/AI Fundamentals, Data Handling & Pre-processing
•    Knowledge of software testing concepts and automotive embedded systems

Thesis #2

Fuzzing-Based Automated Testing and Continuous Integration for Virtualized Automotive Software

Software-in-the-Loop (SIL) testing validates the behavior of individual software components (SWCs) early in the AUTOSAR development process. While traditional tests ensure correctness against expected outputs, they often fail to uncover robustness issues like boundary violations, unexpected input combinations, or unhandled states. Fuzz testing (fuzzing), widely used in security and reliability testing, can automatically generate abnormal or random input sequences to stress-test software logic. Integrating fuzzing into SIL testing, alongside automated CI/CD pipelines, can greatly improve early fault discovery, regression detection, and overall software resilience.

2.1 Research Questions
•    How can fuzzing techniques be adapted for SIL environments to complement standard model- or requirement-based tests?
•    What is the effect of combining fuzz-based and automatically generated functional tests on coverage and fault detection?
•    How can such fuzzing campaigns be automated and integrated into a CI/CD pipeline using SIL test environment?
•    What metrics best represent robustness improvements in early software validation?

2.2 Objectives
•    Implement a hybrid automated testing framework combining model-based and fuzz-based test generation for SWC-level SIL testing.
•    Integrate the framework into a CI/CD pipeline (e.g., GitLab CI or Jenkins) for automated build, test, and reporting cycles.
•    Quantify improvements in fault detection, coverage, and robustness over manual and non-fuzzed tests.
•    Develop best-practice guidelines for applying fuzzing to early-stage automotive software validation.

2.3    Qualifications:
•    Master’s in either Computer Engineering, Electrical Engineering, Embedded Systems, Computer Science (with embedded focus)
•    Knowledge of 
o    Embedded Software Development — familiarity with microcontroller or ECU software architectures
o    Programming in C/C++
o    Software Testing & Verification
o    Version Control & Automation

Duration

•    Tentative proposed thesis work period: 19th January 2026 to 26th of June 2026 (dates can be flexible with +/- 7 days) 
•    The duration for this thesis work is 20 weeks.
•    30 ECTS (Master academic credits) in agreement with your Thesis Advisor in University
•    Each thesis is to be conducted by 2 Students working in pair.