The ISIP40 Research Unit (Information and Signal Processing for Cognitive Telecommunications) research lines are related to theories, methods and techniques for information processing and learning aimed at the development of artificial dynamic cognitive systems with application in the ICT sector.
In particular, the research focuses on how to develop Self-Awareness and transfer functionalities for learning operational knowledge between artificial intelligent agents and between artificial agents and natural agents (eg robots, autonomous driving cars, UAVs, cognitive radios), etc.). In this sense we study at a methodological level aspects related to: 1) the representation and incremental learning of dynamic and causal hierarchical probabilistic models for the analysis and fusion of information and multisensory signals of their own and heterogeneous perception (eg video, radio, control, localization); 2) the use of the models learned for the definition of dynamic predictive semantic filters able to detect situations of non-stationarity with respect to the dynamic rules of the model itself (e.g. Dynamic Bayesian Networks, Switching Models); 3) to the definition of physical and logical architectures of bio-inspired dynamic cognitive systems that use the techniques of representation and inference to the previous points (eg autonomous driving systems, safe intelligent buildings, anti-jamming systems for cognitive radio, etc.) ; 4) to the study of methods and techniques to allow the transfer of operative knowledge between artificial and natural agents (e.g. man) through the elaboration and reorganization of the learned models with respect to situations not previously explored.
The aim of the research is to produce not only theory, innovative techniques and systems but also simulators and demonstrators on a small or natural scale that benefit from the techniques developed. An element of measurability of the relevance of the results obtained will be the publication in international journals, books and conferences. This in the framework of ongoing national and international collaborations (e.g. joint doctorate with Queen Mary University London and Carlos III University of Madrid) and to be developed.
More specifically the methodologies studied aim to find application in the definition of systems and in the realization of prototypes in different sectors that include: 1) terrestrial, air and marine cognitive vehicles; 2) software and cognitive radio; 3) cognitive devices for human-environment interactions (e.g. Smart Glasses); 4) security, intelligent environments and smart communities.
With respect to this element, they constitute objectives of the U.O. the presentation, acquisition and performance of research projects funded by national (MIUR, MISE, etc.) and international (e.g. EU, Marie Curie) and national and international private industries.
Contact person: Prof. Carlo Regazzoni