Obiettivo
The project will implement an autonomous system that combines visual and tactile information processing to achieve object recognition, navigation, and memory formation. The tactile system of the robot will adopt the form of artificial vibrissae (hence "artificial mouse"), allowing a direct link to the neurobiology of the respective sensory modules in rodents. The project will combine a synthetic biorobotics approach with parallel neurophysiological experiments in awake behaving rats, and computational modelling that allows to identify relevant information processing principles. The project aims at exploiting biological principles for the construction of adaptive, flexible and robust autonomous architectures; at testing neurobiological hypotheses by implementation; and at creating predictions for physiological experiments from the synthetic and modelling approach. The project will implement an autonomous system that combines visual and tactile information processing to achieve object recognition, navigation, and memory formation. The tactile system of the robot will adopt the form of artificial vibrissae (hence "artificial mouse"), allowing a direct link to the neurobiology of the respective sensory modules in rodents. The project will combine a synthetic biorobotics approach with parallel neurophysiological experiments in awake behaving rats, and computational modelling that allows to identify relevant information processing principles. The project aims at exploiting biological principles for the construction of adaptive, flexible and robust autonomous architectures; at testing neurobiological hypotheses by implementation; and at creating predictions for physiological experiments from the synthetic and modelling approach.
OBJECTIVES
Using a parallel investigation of an artificial and a natural system, the project will address a number of issues that are of crucial relevance for both the explanation of natural behaviour and the creation of ""living artefacts"": How do sensory systems interact that differ profoundly in their morphology, computational architecture and dynamics? What is the nature of cross-modal integration, cross-modal learning and generalisation? How can multimodal input be efficiently used for behavioural control, navigation and spatial/object memory? What is the role of temporal dynamics and neuronal synchronisation in sensori motor architectures? The approach is holistic in the sense that behaviour will be studied at the integral systems level, in both the artefact and the laboratory animals.
DESCRIPTION OF WORK
In contrast to most neuroscience research, this project pursues an integrative approach to an understanding of sensory processing and generation of adaptive behaviour in animals, in the sense that it explicitly studies learning and adaptation in the context of multimodal sensory processing, sensori motor integration, and embodiment of information processing. This integrative approach becomes possible by employing a novel, synthetic methodology, where neural models are derived from the results of neurophysiological and behavioural experiments and validated in real-world experiments using autonomous mobile robots. The project aims at both an explanation of natural behaviour and at the creation of ""living artefacts"". University of Zurich (CH) and Max Planck Institute for Psychological Research (DE) will develop the multisensory robot equipped with visual sensors and an artificial whisker system in order to perform robot experiments on navigation and learning based on multisensory cues. The co-ordinating laboratory, Research Centre Julich (DE) will record neuronal signals from behaving rats that have to combine visual and tactile cues in a behavioural paradigm. International School for Advanced Studies SISSA (IT) aims at unravelling the functional details of the processing stages in the somatosensory pathway to provide the basis for the neural models. ETH/University of Zurich will define a simulated neuronal network according to anatomical and physiological data, use experimental results as input for the simulation studies and derive relevant functional parameters that will be incorporated into the robot architecture.
Campo scientifico
Invito a presentare proposte
Data not availableMeccanismo di finanziamento
CSC - Cost-sharing contractsCoordinatore
20246 HAMBURG
Germania