Paolo Arena Livres

In this monograph, new structures of neural networks in multidimensional domains are introduced. These architectures are a generalization of the Multi-layer Perceptron (MLP) in Complex, Vectorial and Hypercomplex algebra. The approximation capabilities of these networks and their learning algorithms are discussed in a multidimensional context. The work includes the theoretical basis to address the properties of such structures and the advantages introduced in system modelling, function approximation and control. Some applications, referring to attractive themes in system engineering and a MATLAB software tool, are also reported. The appropriate background for this text is a knowledge of neural networks fundamentals. The manuscript is intended as a research report, but a great effort has been performed to make the subject comprehensible to graduate students in computer engineering, control engineering, computer sciences and related disciplines.

This book presents the result of a joint effort from different European Institutions within the framework of the EU funded project called SPARK II, devoted to device an insect brain computational model, useful to be embedded into autonomous robotic agents.Part I reports the biological background on Drosophila melanogaster with particular attention to the main centers which are used as building blocks for the implementation of the insect brain computational model.Part II reports the mathematical approach to model the Central Pattern Generator used for the gait generation in a six-legged robot. Also the Reaction-diffusion principles in non-linear lattices are exploited to develop a compact internal representation of a dynamically changing environment for behavioral planning.In Part III a software/hardware framework, developed to integrate the insect brain computational model in a simulated/real robotic platform, is illustrated. The different robots used for the experiments are also described. Moreover the problems related to the vision system were addressed proposing robust solutions for object identification and feature extraction.Part IV includes the relevant scenarios used in the experiments to test the capabilities of the insect brain-inspired architecture taking as comparison the biological case. Experimental results are finally reported, whose multimedia can be found in the SPARK II web www.spark2.diees.unict.it

This book guides readers along a path that proceeds from neurobiology to nonlinear-dynamical circuits, to nonlinear neuro-controllers and to bio-inspired robots. It provides a concise exploration of the essence of neural processing in simple animal brains and its adaptation and extrapolation to modeling, implementation, and realization of the analogous emergent features in artificial but bio-inspired an emerging research field. The book starts with a short presentation of the main areas of the Drosophila brain. These are modeled as nonlinear dynamical structures, which are then used to showcase key features like locomotion, motor learning, memory formation, and exploitation. It also discusses additional complex behaviors, such as sequence learning and perception, which have recently been discovered to exist in insects. Much of the material presented has been tested in biorobotics classes for the Master’s degree in Automation Engineering and Control of Complex Systems at the University of Catania. Reporting on the work fostered by several national and international research projects, the book offers researchers novel ideas on how neuro-inspired dynamics can be used in developing the autonomous machines of the future.