91411 - Intelligent Robotic Systems

Learning outcomes

At the end of the course, students have acquired knowledge and competencies for designing a system composed of one or more robots (here, we call "robot" an autonomous system which exists in the physical world, can sense its environment and can act on it to achieve some goals). In particular, students know the main models, methods, architectures and tools for programming robots equipped with nontrivial computational and cognitive capabilities.

Course contents

Introduction to robotics

• Brief history of robotics
  
• Notions of robot and its behavior in a physical environment
• Main issues in intelligent robotic systems design

Behavior-based robotics

• Overview of main paradigms for coordinating behaviors
  
• The subsumption architecture
• Motor schemas
• Fuzzy logic and fuzzy systems
• Behavior trees
• Experimental evaluation and parameter tuning of control software for robots: practical guidelines
• Swarm robotics
  

Robot adaptation and learning

• Artificial evolution and evolutionary robotics
• Automatic design of robot programs
• Associative learning
• Reinforcement learning
• Value-based learning and intrinsic motivation
  

Deliberative control

• Informed search algorithms. A* and its variants for path planning problems
• Robot planning: definitions, principles, and main approaches
  
• Navigation problems and main solution approaches

Lab activities

Experiments conducted in simulation and with physical robots (Thymio) to explore different control strategies and apply the knowledge acquired during the course. The laboratory activities provide hands-on experience in designing, implementing, and testing robot control algorithms in both virtual and real-world environments.

Readings/Bibliography

Lecture slides prepared by the instructor, together with selected research papers and additional course materials, will be made available through the course website.

 

Course textbooks:

• S. Nolfi, "Behavioral and Cognitive Robotics - An adaptive perspective", CNR-ISTC, 2021. Available for download: https://bacrobotics.com/
• R. Pfeifer and C. Scheier, "Understanding intelligence", The MIT Press, 1999.

Additional books:

• R. Pfeifer and J. Bongard, "How the body shapes the way we think", The MIT Press, 2007.
• M. Mataric, "The Robotic Primer", The MIT Press, 2007.
• U. Nehmzow, Robot Behaviour: Design, Description, Analysis and Modelling, Springer, 2009.

Teaching methods

The course consists of lectures and laboratory activities. The laboratory sessions are designed to provide students with hands-on experience in addressing the main challenges involved in the design and implementation of robot control programs, while reinforcing the concepts and techniques presented during the lectures. The discussions that take place during the laboratory sessions and in the subsequent lectures constitute an integral part of the learning process, as they provide opportunities to analyse the proposed solutions, reflect on the outcomes of the experiments, and consolidate the knowledge acquired.

Assessment methods

The final examination is intended to assess the extent to which students have achieved the learning objectives of the course, namely the acquisition of the theoretical knowledge presented during the lectures and the ability to design, implement, and evaluate software for robotic systems.

The final assessment consists of two components, each contributing 50% to the overall grade:

• Practical activity (50%)

• Oral examination (50%)

 

The practical activity may be fulfilled in one of the following ways:

• The development of an individual or small-group project, subject to the instructor's approval; or

• active participation in the laboratory sessions, together with the successful, careful, and timely completion and submission of the assigned laboratory work, in full compliance with the instructions and requirements provided by the instructor.

The practical component is assessed on the basis of the student's ability to analyse the proposed problem, adopt a sound engineering and scientific approach, and design, implement, and evaluate an effective control solution for a robotic system. Particular attention is paid to the methodology followed, the quality of the proposed solution, and the student's understanding of the design choices made.

The oral examination assesses the student's understanding of the topics covered in the course. The evaluation is based on four main criteria:

• Contextual understanding, namely the ability to frame concepts within the broader context of intelligent robotics and to identify relationships among the topics covered in the course.

• Clarity of exposition, including the ability to present arguments in a logical, coherent, and effective manner using appropriate technical terminology.

• Precision, referring to the correctness and accuracy of definitions, explanations, and technical details.

• Depth of knowledge, reflecting the student's critical understanding of the subjects, ability to motivate design choices, and capacity to discuss limitations, alternatives, and possible extensions.

The oral examination is graded according to the following general criteria:

• 30/30: all four evaluation criteria are fully satisfied, demonstrating complete and accurate knowledge of the course topics, clear and rigorous exposition, and a deep understanding of the subject matter.

• 27–29/30: the overall performance is very good, with only minor shortcomings in one or more of the evaluation criteria.

• 24–26/30: the student demonstrates an adequate understanding of the course topics, although some important elements are missing or insufficiently developed, without compromising the overall achievement of the learning objectives.

• 18–23/30: the student demonstrates only a basic level of achievement of the learning objectives, with significant weaknesses in one or more evaluation criteria, although the minimum requirements for passing the examination are met.

The mark of 30 cum laude may be awarded in exceptional cases to students who demonstrate outstanding mastery of the course contents, exceptional critical thinking, and an oral performance that significantly exceeds the requirements for the highest ordinary grade.

 

In addition to the official examination sessions (published on almaesami), students may arrange to take the examination at other times, subject to prior agreement with the instructor.

Teaching tools

Students will be provided with teaching materials prepared or selected by the instructor, including lecture slides and scientific papers published in peer-reviewed journals and conference proceedings. Laboratory activities will make use of robotic simulation environments (ARGoS) as well as physical robotic platforms (Thymio).

The recommended textbooks, together with the teaching materials made available through the course website, cover all the topics addressed during the course and constitute the reference material for both the lectures and the laboratory activities.

Office hours

See the website of Andrea Roli