Skip to main content
European Commission logo
English English
CORDIS - EU research results
CORDIS
CORDIS Web 30th anniversary CORDIS Web 30th anniversary
Content archived on 2024-05-27

Systemic Intelligence for GrowiNg up Artefacts that Live

Article Category

Article available in the following languages:

Growing up infant robots

An innovative approach for structuring a robot's control system through learning is expected to revolutionise future robotics.

Until recently, ontogenesis or simply enabling artefacts to 'grow up' a key feature for 'living artefacts', was considered pure fiction. The SIGNAL project focused on systemic intelligence for developing artefacts that mature by exploiting multidisciplinary research. Thereby, evolutionary algorithms, neuroinformatics, psychology of learning, memory organisation and reasoning, robotics and linguistics were combined. On the basis of a pre-defined genotype that specifies a set of basic abilities, the system's knowledge increases through learning and experience. The artefact undergoes a scheduled sequence of phases ending in a phenotype, that is, an individually matured entity. Part of the project work involved the development of the underlying systemic architectural approach. With the aid of a set of modules with unique functionality, the controller was implemented. Using a multilayered modules' organisation strategy the higher layers serve higher functions. Moreover, two sophisticated control schemes, an upstream sensor and a downstream actuatory system imitate the direction of the neurons in the human peripheral nervous system. This architecture was selected because of the numerous potentiallities it offers when realising modern controller designs. Provision was made for key functions concerning the robot in its environment, the sensory upstream, the actuatory downstream and the action selection mechanism. An important aspect of the functional sections also involves the internal value system (IVS), which is connected with 'drives' and 'emotions'. This system encompasses sensory values that define the action selection mechanism. The IVS values feed both the action selection and the action selection learning modules. The developed system provides a widely applicable approach for intelligent control of autonomous robots and agents. This promising concept of designing autonomous robots that are based on a controller is an area that is still under exploration to discover its full potentialities. For more information click at: http://www.ist-signal.org/

Discover other articles in the same domain of application