OVERVIEW

The course provides advanced skills in the design of software architectures for embedded, Internet of Things (IoT), and intelligent cyber-physical systems.

Particular attention is devoted to the integration of sensing, intelligent processing through Machine Learning and TinyML, and action in the physical world according to the Physical AI paradigm.

The course addresses software architecture design, application development, the integration of Machine Learning models on microcontrollers, testing and validation, technical documentation, security, and software lifecycle management.

The course also introduces Natural Language Programming and the responsible use of Generative Artificial Intelligence as a support tool for the design and development of intelligent systems.

AIMS AND CONTENT

LEARNING OUTCOMES

The course provides skills for the software design of embedded and intelligent cyber-physical systems, with particular attention to the integration between sensory perception, intelligent processing and action in the physical world according to the Physical AI paradigm. The course addresses software architectures, modeling of responsive systems, integration of TinyML models on microcontrollers, testing, lifecycle, technical documentation and aspects of safety, security and AI governance. Particular attention is paid to the conscious use of generative Artificial Intelligence as a support to software design in the field of Augmented Engineering. Teaching activities include practical exercises in software development and application prototyping on embedded platforms.

AIMS AND LEARNING OUTCOMES

The course provides advanced knowledge and skills for the design, development, validation, and lifecycle management of embedded, Internet of Things (IoT), and intelligent cyber-physical systems, with particular emphasis on software and architectural aspects. Students will acquire knowledge of architectural models for reactive and distributed systems, software development techniques for microcontrollers, and the integration of sensing, intelligent processing, and action according to the Physical AI paradigm.

The course addresses the integration of Machine Learning and TinyML models on resource-constrained embedded devices, the design of intelligent systems based on sensors and actuators, and the modelling of components, behaviours, and interactions using standard methods and tools.

The course also covers software engineering methodologies for intelligent systems, testing, validation, safety and security strategies, technical documentation, and the lifecycle management of firmware and Machine Learning models.

Particular attention is devoted to the use of Generative Artificial Intelligence as a support tool for design, development, and technical documentation activities. Students will learn to use such tools critically and responsibly while maintaining control over design decisions and validation processes.

The overall objective is to educate professionals capable of addressing the entire development process of intelligent cyber-physical systems, from architectural design to implementation and validation, while operating effectively in interdisciplinary and technologically complex environments.

PREREQUISITES

To successfully follow the course, students are expected to have:

* basic programming knowledge, preferably in C, C++, or Python;
* familiarity with software development environments and tools;
* basic knowledge of computer architecture and digital systems.

Previous experience with embedded systems, Internet of Things (IoT), digital electronics, Machine Learning, or software modelling may be beneficial but is not mandatory.

TEACHING METHODS

The course is organised into two main categories of activities:

**Lectures (40 hours):** theoretical and applied sessions covering the design of embedded, IoT, and intelligent cyber-physical systems, software engineering methodologies, Machine Learning and TinyML integration, testing and validation techniques, and AI governance principles.

**Hands-on Sessions (20 hours):** guided laboratory and project-based activities in which students design, develop, and validate intelligent cyber-physical systems using embedded platforms and related development tools.

The learning activities are structured in two complementary phases. In the first phase, students design and develop solutions autonomously, acquiring the fundamental engineering competences required by the course. In the second phase, the same design problems are revisited with the support of Generative Artificial Intelligence, with the goal of developing critical supervision, validation, and responsible AI usage skills.

Attendance of lectures and laboratory sessions is strongly recommended. Course materials, laboratory activities, and project work constitute an integral part of the learning experience.

The course is worth 6 ECTS credits, corresponding to a total student workload of 150 hours.

SYLLABUS/CONTENT

1. Foundations of Embedded Systems and Cyber-Physical Systems
   (Architecture and characteristics of embedded, IoT, and intelligent cyber-physical systems; Physical AI paradigm; integration of sensing, intelligent processing, and action; functional and non-functional requirements.)
   Activities: analysis of case studies and modelling of intelligent cyber-physical systems.

2. Software Architecture Design for Embedded Systems
   (Software and hardware architectures; top-down design; block diagrams, WBS, and hierarchical modelling; functional decomposition; perception–decision–action pipelines.)
   Activities: architectural modelling, ProjectLibre usage, and system design exercises.

3. Embedded Application Modelling and Design
   (Finite State Machines, event-driven systems, physical and digital user interfaces, UML modelling of structure and behaviour.)
   Activities: FSM development, UML modelling, and software component design.

4. Intelligent Agents and Autonomous Cyber-Physical Systems
   (Reactive and deliberative models, autonomous behaviour, agent–environment interaction, control architectures, and distributed intelligent systems.)
   Activities: design and simulation of embedded intelligent agents.

5. Machine Learning and TinyML for Embedded Systems
   (Machine Learning foundations for cyber-physical systems; data acquisition; model training and validation; deployment on microcontrollers; inference on resource-constrained devices.)
   Activities: development and integration of TinyML models.

6. Embedded Application Development and Hardware–Software Integration
   (Microcontroller programming, sensor and actuator integration, communication protocols, and management of physical system resources.)
   Activities: implementation of embedded prototypes and hardware–software integration.

7. Testing, Validation, Safety, and Security
   (Testing and verification strategies; functional validation; robustness; reliability; safety and security of intelligent embedded systems.)
   Activities: design and execution of test cases and system validation procedures.

8. Lifecycle Management and Technical Documentation
   (Software development methodologies; requirements management; version control; lifecycle management of firmware and Machine Learning models; technical documentation and Model Cards.)
   Activities: project documentation, release management, and lifecycle analysis.

9. Natural Language Programming and Augmented Engineering
   (Use of Generative Artificial Intelligence to support design, development, testing, and documentation activities; prompting techniques; supervision and validation of AI-generated outputs; comparison between autonomous and AI-assisted development processes.)
   Activities: guided use of Generative AI tools in software engineering tasks.

10. AI Governance and Professional Responsibility
    (Ethical, organisational, and professional aspects of AI adoption in intelligent systems; transparency, traceability, accountability, safety, and sustainability.)
    Activities: analysis of case studies and discussion of engineering decisions.

RECOMMENDED READING/BIBLIOGRAPHY

All materials used during lectures and lab sessions will be made available progressively on the AulaWeb platform, in the section “Materials used in class”, together with links to relevant resources and freely accessible references.

LESSONS

LESSONS START

https://corsi.unige.it/en/corsi/11959/studenti-orario

Class schedule

The timetable for this course is available here: Portale EasyAcademy

EXAMS

EXAM DESCRIPTION

In order to take the exam, students must register online through the Student Portal at:
https://servizionline.unige.it/studenti/

The exam consists of an individual oral examination on the course content, combined with the discussion of the assigned project and the related technical documentation.

ASSESSMENT METHODS

The assessment consists of the discussion of a project developed during the course, an individual oral examination on the course topics, and the evaluation of the technical documentation produced by the student.

**Project Discussion and Oral Examination (approximately 70% of the final grade)**

This part of the assessment is intended to verify the acquisition of the design, methodological, and technical competences addressed during the course. Evaluation criteria include:

* the ability to design and develop embedded, IoT, or intelligent cyber-physical systems;
* the understanding of software architectures and design solutions adopted within the project;
* the ability to integrate sensing, intelligent processing, and actuation into a coherent system;
* the ability to justify and critically discuss design decisions;
* knowledge of the course topics, including Machine Learning, TinyML, testing, validation, safety, security, and lifecycle management;
* the ability to use Generative Artificial Intelligence tools responsibly and critically as support for engineering activities;
* appropriate use of technical terminology and effective communication of technical concepts.

**Evaluation of Technical Documentation (approximately 30% of the final grade)**

Documentation will be evaluated in terms of completeness, clarity, consistency, and technical quality. It may include project specifications, architectural descriptions, models and diagrams, software documentation, testing and validation results, and documentation related to Machine Learning models where applicable.

Particular attention will be devoted to the traceability of design decisions and to the documentation of any use of Generative Artificial Intelligence tools during project development, testing, or documentation activities.

Detailed assessment criteria and examination procedures will be presented during the course.

FURTHER INFORMATION

Students with disabilities or specific learning disorders (SLD) may request compensatory or exemptive measures for the exam. These measures will be defined on a case-by-case basis in consultation with the Faculty Disability Liaison of the University Committee for Student Support. Students who wish to make such a request are invited to contact the course instructor and copy the Liaison Officer:
https://unige.it/commissioni/comitatoperlinclusionedeglistudenticondisabilita.html