Introduction

The document presents a structured and adaptable curriculum for Bachelor and Master level studies in Safe Autonomous Vehicles (SafeAV), with a strong focus on Verification and Validation (V&V) of autonomous systems. The framework serves as a foundation that higher education institutions can adapt and expand when designing their own study modules or programmes related to the safety, reliability, and governance of autonomous technologies.

The curriculum follows a modular structure combining theoretical foundations, applied engineering knowledge, and hands-on experimentation. It is supported by two complementary educational resources developed within the SafeAV project:

• SafeAV Handbook – provides the theoretical and methodological background, including system architectures, sensing, software, and formal V&V methods.
• SafeAV Hands-on Guide – offers practical laboratory and simulation exercises that allow students to perform verification and validation tasks using real and virtual autonomous platforms.

The SafeAV curriculum architecture defines the overall structure, modular hierarchy, and learning flow that connects theoretical knowledge, simulation-based validation, and experimental practice. It ensures coherence between study levels and provides a clear path from basic understanding to advanced assurance of autonomous vehicle safety. Modules are organised in pairs: Part 1 (Bachelor) introduces the concepts, while Part 2 (Master) deepens the same topic through practical verification and validation (V&V) methods. This two-level structure enables a stepwise learning progression across study cycles and gives universities the flexibility to adopt the curriculum or parts of it into existing educational programs.

Each topic therefore exists in two complementary parts:

• Part 1 (Bachelor level) – introduces the fundamental principles, technologies, and system interactions. Emphasis is on conceptual understanding, component function, and system-level awareness.
• Part 2 (Master level) – deepens the focus toward Verification and Validation (V&V), including analytical, experimental, and regulatory methods used to demonstrate safety, reliability, and trustworthiness.

For example, in Hardware and Sensing Technologies Part 1, students learn sensor types, signal processing basics, and data acquisition. In Part 2, they perform calibration, fault analysis, redundancy testing, and scenario-based validation using V&V tools and simulation environments. This two-stage progression ensures continuity between study cycles and supports lifelong learning paths in autonomous vehicle engineering.

The overall curriculum can be described as three integrated layers:

• Conceptual layer – theoretical foundations and system-level understanding (covered mainly in the SafeAV Handbook)
• Practical layer – hands-on experiments, data analysis, and verification in laboratory environments (based on the Hands-on Guide)
• Digital layer – self-study materials, MOOC courses, and AI-supported assistants that guide learning and track individual progress

These layers are interconnected through shared terminology, datasets, and unified learning outcomes across all modules.

The curriculum consists of six interrelated modules that together form a complete 6 ECTS study block but can also be used independently. Each module represents approximately 25–30 hours of student work, combining lectures, laboratory tasks, and self-study. The modular design allows multiple implementation strategies:

• full six-module SafeAV course (6 ECTS)
• selected modules as independent 1 ECTS units
• integration into existing robotics, AI, or control courses
• use for lifelong learning or professional training

Each module includes theoretical reading, guided experiments, simulation exercises, and assessment through a report, presentation, or quiz. The same structure is followed in all modules to maintain coherence across institutions.

The undergraduate programme introduces the building blocks of autonomous systems and their relation to safety assurance. The emphasis is on understanding system components and basic verification of function. Six modules (1 ECTS each) provide foundational knowledge of vehicle architecture, autonomy levels, sensing, computing, software systems, and human–machine interaction.

Modules – Part 1:

• Autonomous Vehicles
• Hardware and Sensing Technologies
• Software Systems and Middleware
• Perception, Mapping, and Localization
• Control, Planning, and Decision-Making
• Human–Machine Communication

Each module combines reading assignments from the SafeAV Handbook with laboratory or simulation tasks from the Hands-on Guide, such as sensor calibration, perception benchmarking, or control-loop validation. The recommended full scope equals 6 ECTS, yet the modular design allows partial adoption depending on local curricula and student pathways.

The Master’s programme deepens the same thematic areas into Part 2 modules that focus on validation, verification, and system governance. Students explore how safety and reliability are demonstrated through structured testing, scenario generation, formal methods, and compliance with standards such as ISO 26262, SOTIF, and UL 4600. Modules are directly linked to the advanced chapters of the SafeAV Handbook (e.g., 2.7–2.9 and 8 “Autonomy Validation Tools”) and the experimental work described in the Hands-on Guide.

Modules – Part 2:

• Hardware and Sensing Technologies (Validation and Reliability)
• Software Systems and Middleware (Safety and Verification)
• Perception, Mapping, and Localization (Scenario-based Testing)
• Control, Planning, and Decision-Making (Formal and Simulation-backed Validation)
• Human–Machine Communication (HMI Safety and V&V)
• Autonomy Verification and Validation Tools (Integrated Frameworks and Methods)

Students build validation pipelines from model design to field testing, using digital twins and simulation environments. The progression mirrors the V-model lifecycle introduced in the handbook — from design to verification, validation, and governance.

Each module supports flexible learning environments that allow both classroom and remote participation:

• classroom teaching for theoretical foundations
• local laboratories for experiments with sensors, embedded controllers, or robotic platforms
• virtual labs with remote access to simulated autonomous systems
• MOOC-based self-study materials with AI-assisted feedback
• hybrid sessions combining on-site instruction with online validation tasks

The SafeAV Hands-on Guide defines equipment lists, virtual lab configurations, and step-by-step procedures. Remote laboratory setups ensure that students can conduct verification and validation exercises even without physical access to hardware.

Digital tools, Dokuwiki materials, and the MOOC environment allow integration with AI-based assistants that support self-learning, answer technical questions, and provide feedback on simulation or validation tasks.

The integration of artificial intelligence (AI) tools into the SafeAV curriculum is a central element for enabling modern, personalized learning experiences. In addition to supporting individualized study paths for typical learners, it also enhances accessibility and provides improved educational opportunities for students with special needs.

AI technologies are implemented at two levels:

• integration within the learning content to illustrate how AI supports autonomous vehicle V&V (e.g., AI in perception, planning, or safety analysis)
• integration as pedagogical tools to assist students and lecturers throughout the learning process

The following AI-based methods are used within the SafeAV ecosystem:

• AI-powered virtual assistants – LLM-based agents embedded in the MOOC and Dokuwiki environment answer course-related questions, explain theoretical concepts, and provide V&V-related guidance.
• AI-driven interactive simulations and virtual labs – intelligent digital twins and scenario generators support sensor fusion validation, control-loop testing, and human–machine communication studies.
• Personalized AI tutors – adaptive learning systems analyse student progress and recommend additional materials, exercises, or simulations based on performance.
• AI-supported content summarization – automatic generation of concise summaries of lectures, reports, and laboratory documentation helps students prepare for assessment and supports accessibility.
• Automated peer review and feedback – integrated AI tools assist in assessing reports and coding exercises, providing constructive feedback and reducing lecturer workload.

AI-based tools play a significant role in SafeAV by reducing repetitive communication tasks, offering continuous learning support, and improving the overall organization of study activities. These systems provide students with round-the-clock access to guidance and feedback, allowing instructors to focus on higher-level mentoring and project supervision.

To ensure trustworthy and responsible use of AI in education, all implementations follow privacy-by-design principles and comply with relevant data protection regulations. Student data are processed transparently and securely, with anonymized interaction records and clear options to opt out of AI-assisted learning when preferred.

In the long term, the SafeAV approach aims to develop a shared and open AI learning framework that promotes accessibility, multilingual support, and collaboration between partner universities, ensuring sustainable and equitable use of AI technologies in higher education.

The SafeAV architecture is open and adaptable. Educational institutions may:

• adopt the complete curriculum as a dedicated SafeAV course
• integrate selected modules into existing study programmes
• use materials in non-formal education or industrial training

All materials are licensed under Creative Commons (CC BY-NC), allowing reuse and modification while keeping alignment with European learning standards and ECTS principles. This ensures consistency across partner universities while maintaining flexibility for local adaptation and future extension.