Difference between revisions of "Cognitive Robotics Resources"

Revision as of 17:15, 30 November 2014

The resources here are intended to support the IEEE Robotics and Automation Technical Committee for Cognitive Robotics.

Research Material

Research Issues

Controversies in Cognitive Systems Research

White Papers

• Dynamic Field Theory (DFT): Applications in Cognitive Science and Robotics
• Observing Human Behaviour in Image Sequences: The Video-Hermeneutic Challenge
• Cognitive Ontologies: Mapping Structure and Function of the Brain from a Systemic View
• Coordinating with the Future: the Anticipatory Nature of Representation
• Communication and Distributed Control in Multi-Agent Systems: Preliminary Model of Micro-unmanned Aerial Vehicle (MAV) Swarms
• Enactive Artificial Intelligence
• CoEvolutionary Approaches in Cognitive Robotic Systems Design
• Action Selection for Intelligent Systems

Surveys

• Comparative Repository of Cognitive Architectures Biologically Inspired Cognitive Architectures Society (BICA)] comparison of 25 cognitive architectures]
• A Survey of Cognitive and Agent Architectures University of Michigan survey of (mostly cognitivist) cognitive architectures
• A Framework for Comparing Agent Architectures Aaron Sloman and Matthias Scheutz, Proceedings UKCI'02, UK Workshop on Computational Intelligence, 2002

Robot Platforms

• The iCub is an open source cognitive humanoid robotic platform.

Research Networks

• http://www.vernon.eu/euCognition/index.htm euCognition] This network was the first in a series of three networks, culminating in EUCog III below.
• European Society forl Cognitive Systems (originally the European Network for the Advancement of Artificial Cognitive Systems, Interaction and Robotics).

Software Resources

AKIRA

AKIRA is an open source C++ framework for designing distributed, modular agent architectures (e.g., schema-based, behaviour-based, etc.) Some features of the framework include:

• support for decentralized, asynchronous and parallel processing;
• flexible implementation of the beahavior each module, with libraries for implementing soft computing algorithms (neural networks, fuzzy logic, fuzzy cognitive maps, etc.);
• flexible implementation of modules interaction (the default is inspired by complex systems research, i.e. local excitation and global inhibition among the modules);
• communication via blackboard and shared global variables;
• works under linux.

See: Giovanni Pezzulo and Gianguglielmo Calvi (2007). Designing Modular Architectures in the Framework AKIRA. Multiagent and Grid Systems, 3, 65--86.

More details.

AmonI - Artificial Models of Natural Intelligence

Behavior Oriented Design (BOD) and POSH Action Selection, and much more ...

Behaviour oriented design is a methodology for developing intelligent systems. It extends object oriented design to the special problems of proactive systems, including real-time systems for dynamic enviornments. To the extent that these systems are agents, they need goals and priorities; in BOD these are specified using POSH action selection.

More details.

CAST: The CoSy Architecture Schema Toolkit

This a software toolkit to support the developments of intelligent systems based on a space of possible architecture designs described by the CoSy Architecture Schema (CAS). The CoSy Architecture Schema Toolkit (CAST) is a software implementation of this schema designed to allow researchers (primarily in the fields of AI and robotics) to develop instantiations of the schema. The toolkit supports C++ and Java, and provides a communication framework for distributing an instantiation across a network. Its primary scientific purpose is to maintain a separation between a system's architecture and the content of its architecture, allowing one to be varied independently of the other.

More details.

YARP: Yet Another Robot Platform

YARP is a thin middleware for humanoid robots (and more). YARP supports building a robot control system as a collection of programs communicating in a peer-to-peer way, with an extensible family of connection types (tcp, udp, multicast, local, MPI, mjpg-over-http, XML/RPC, tcpros, ...) that can be swapped in and out to match your needs. It also supports similarly flexible interfacing with hardware devices. The strategic goal is to increase the longevity of robot software projects.

More details.

Teaching Resources

Tutorials

• The dynamic neural field approach to cognitive robotics
• Neuronal Dynamics Approaches to Cognitive Robotics
• Tutorial on Embodiment
• Control engineering of autonomous cognitive vehicles - a practical tutorial

Model Curricula

• Cognitive Systems Model Curriculum

Course Material

Artificial Cognitive Systems

Teaching material to support Artificial Cognitive Systems - A Primer, MIT Press, 2014.

Cognitive Computer Vision

This course focusses mainly on generative models and deals with such issues as Bayesian Networks, Gaussian Mixtures, and Hidden Markov Models. The 15 lectures and slides for the entire course are available for download from the ECVision website.

Human Vision

The Human Vision System

Cognitive Architectures

This module is part of a course on computation and cognition that gives a particularly useful introduction to production system cognitivist cognitive architectures.

Degrees in Cognitive Systems

The following is a sample of the Master-level degrees that are available in cognitive systems.

• Universitat Pompeu Fabra, Spain Interdisciplinary Master in Cognitive Systems and Interactive Media (60 ECTS)
• Technical University of Munich Robotics, Cognition, and Intelligence
• University of Warwick M.Sc. in Cognitive Systems
• University of Zurich Master of Science in Informatics: Field of study Multimodal and Cognitive Systems
• Jacobs University, Germany Cognitive Systems and Processes