logo-polimi
Loading...
Risorse bibliografiche
Risorsa bibliografica obbligatoria
Risorsa bibliografica facoltativa
Scheda Riassuntiva
Anno Accademico 2018/2019
Scuola Scuola di Ingegneria Industriale e dell'Informazione
Insegnamento 091316 - COGNITIVE ROBOTICS
Docente Matteucci Matteo
Cfu 5.00 Tipo insegnamento Monodisciplinare

Corso di Studi Codice Piano di Studio preventivamente approvato Da (compreso) A (escluso) Insegnamento
Ing Ind - Inf (Mag.)(ord. 270) - CO (482) COMPUTER SCIENCE AND ENGINEERING - INGEGNERIA INFORMATICA*AZZZZ091316 - COGNITIVE ROBOTICS

Obiettivi dell'insegnamento

The course addresses the methodological aspects of Cognitive Robotics. Cognitive Robotics is about endowing robots and embodied agents with intelligent behavior by designing and deploying a processing architecture which makes them apt to deliberate, learn, and reason about how to behave in response to complex goals in a complex world. Perception and action, and how to model them in reactive systems or system using symbolic representations are therefore the core issues to address.

 

Inspiring models of Cognitive Robotics arise from different disciplines: the neural architectures from neuroscience, the basic behaviors from ethology, modeling and reasoning from psychology, the emergent behaviors from sociology. Those models could be implemented in terms of formal logic, probabilistic models, and neural networks to be embodied into physical and computational agents. 

 

Implementation issues are presented in the framework of the ROS middleware for robotics, to give students the experience of a quite professional way to develop and experiment robotics algorithms.


Risultati di apprendimento attesi

 

Dublin Descriptors

Expected learning outcomes

Knowledge and understanding

Students will learn:

·       What are the main approaches to develop cognitive systems starting from the symbolic ones and lending to modern neural networks

·       How cognitive systems based on symbolic representations work, their strengths and weaknesses, their computational complexity, and the main reason to favor deliberative architectures in the design of cognitive systems

·       How cognitive systems based on reactive behaviors work, their strengths and weaknesses, their computational complexity, and the main reason to favor reactive architectures in the design of cognitive systems

·       How cognitive systems based on sub-symbolic representations work, their strengths and weaknesses, their computational complexity, and the main reason to favor neural networks in the design of cognitive systems

·       The basic principles of Human Robot Interaction (HRI), in particular, the notions and techniques for non verbal communication, sensors and actuators used in HRI, the basic design interaction models.

·       The Robot Operating System (ROS) framework and how it can be used in the development of autonomous robots and their simulation

Applying knowledge and understanding

Given a specific domain design for a cognitive robot, student will be able to:

·       Identify which cognitive paradigm to apply in the design of the robot autonomy and in the design of a proper Human Robot Interaction

·       Identify which components to use in the design of an interacting robot: selecting sensors, actuators, behaviors, etc.

·       Model planning problems using an high level language for planning problem definition and applying planning algorithms to find a valid/optimal plan

·       Implement the software system controlling a modern autonomous robot based on the ROS middleware

Making judgements

Given an autonomous robot design problem, students will be able to:

·       Identify the limits of a specific cognitive architecture with respect to the specific domain and compare the pros and cons of each to choose the most effective one

·       Identify the most effective hybridization form in the case of use cases where a standard cognitive architecture is not sufficient to fulfill the task

Communication

Student will learn to:

·       Discuss in written form the pros and cons of different cognitive architectures in the realm of a specific design problem

·       Present in a public venue, in the form of a seminar, specific applications of cognitive robotics or techniques not explicitly presented during classes

Lifelong learning skills

Student will learn to:

·       Face a real life autonomous system design problem with a sound background in artificial intelligence and cognitive architectures

·       Understand complex design of cognitive systems beyond the fundamental ones presented during lectures

·       Develop hybrid cognitive architectures for autonomous systems adapting to the specific problem under analysis

·       Improve their knowledge on artificial intelligence and being able to present this new knowledge to others in seminar-like form

 


Argomenti trattati

The course is composed by a set of lectures on specific topics in cognitive robotics, ranging from the main architectural paradigms used in developing cognitive systems to the design of robots capable of interacting naturally with humans. The course outline is:

  • Cognitive Robotics introduction
    • the sense plan act architecture
    • deliberative, reactive, and neural approaches
    • hybrid architectures and systems
  • Deliberative systems for cognitive robots
    • discrete and continuous planners: action representation and map representation
    • non linear planners, partial order planners, and random planners
  • Reactive systems for autonomous agents
    • behavior based systems
    • rule based systems
  • Bioinspired controllers for autonomous robots
    • neural controllers and neural models of space and paths
    • learning mechanisms in robot and embodied agents
  • User/robot and robot/robot interaction
    • multisensorial interfaces: physical principles and telecontrol.
    • the environment as a communication medium: distributed sensing for robot/robot interaction.
    • interface to interact with real and virtual worlds.
  • Robot system development
    • The use of middleware in robot development
    • The Robot Operating System (ROS)
    • Simulating robots with ROS and Gazebo

A detailed schedule of the course will be provided on the course website (http://chrome.ws.dei.polimi.it/index.php/Cognitive_Robotics) with reference to the course slides and the additional material provided by the teachers.


Prerequisiti

No specific requirements are foreseen to attend successfully the course beside basics principles of programming.


Modalità di valutazione

The course comprises theoretical lectures practical sessions, the grade will reflect both. Part of the grade will be based on a classical written exam (mandatory), while the remaining part of the grade will be based on a seminar-like presentation (optional).

The seminar-like presentation is optional, but in case the student will not give it they will only be graded up to 90% of the maximum grade via the written exam. Upon request, student can opt for a practical project instead of the seminar-like presentation.

 

Type of assessment

Description

Dublin descriptor

Written test

Solution of practical and design problems

·       Definition of PDDL model Computation of linear models for regression and classification on small datasets

·       Design of a robot for human interaction in a specific domain, including sensors, actuators and expected behavior

Answer to theoretical questions

·       Describe and discuss pros and cons of different cognitive architectures

·       Describe the rationale behind a specific cognitive model for the development of human robot interaction

·       Describe the architecture and the functioning of ROS and Gazebo

1, 2, 3

 

 

 

 

 

 

1, 2, 3, 4, 5

Assessment of practical homework

Execution of a practical project (in place of oral presentation)

·       Design of a robotic application

2, 3, 4, 5

Oral presentation

Seminar-like presentation

·       Assessment of the presentation of a topic selected by the student which deepens the content of the course on a specific aspect of Cognitive Robotics

·       Assessment of the competences in understanding and judgement when selecting and preparing the material for the presentation

3, 4, 5

 


Bibliografia

Forme didattiche
Tipo Forma Didattica Ore di attività svolte in aula
(hh:mm)
Ore di studio autonome
(hh:mm)
Lezione
30:00
45:00
Esercitazione
20:00
30:00
Laboratorio Informatico
0:00
0:00
Laboratorio Sperimentale
0:00
0:00
Laboratorio Di Progetto
0:00
0:00
Totale 50:00 75:00

Informazioni in lingua inglese a supporto dell'internazionalizzazione
Insegnamento erogato in lingua Inglese
Disponibilità di materiale didattico/slides in lingua inglese
Possibilità di sostenere l'esame in lingua inglese
Disponibilità di supporto didattico in lingua inglese
schedaincarico v. 1.6.1 / 1.6.1
Area Servizi ICT
27/01/2020