The field of neurorobotics is still in its infancy; however, its intersecting motivations are not. On the one hand, theories of neuroscience that require immersion in the real-world can be embedded in mobile agents creating complex patterns of activity believed to be a requirement for understanding higher-order neural function. On the other, the cognitive capabilities of humans remain unparalleled by artificial agents. Emulating biology is one strategy for creating more capable artificial intelligence. Despite these strong motivations for creating neurorobotic entities technological hurdles still remain at all levels. This thesis presents two different contributions to the field of neurorobotics. The first is aimed at reducing the complexity of coupling spiking neural models with virtual agents. This is accomplished through a set of tools that act to abstract the neuroscience details from roboticists and the mechanical details away from the neuroscientists. The second contribution provides an example of how higher-level cognitive theories of speech processing can be integrated into the neurorobotics paradigm. Extracting the emotional content of a speaker, independent of what is being spoken, is a daily act for most people. The neural basis for this ability remains illusive, however cognitive models have been realized. This class of models can be integrated with the biologically realistic neural simulations in a complementary way to expand the capabilities of a neurorobotic system.

The contributions in Toward Learning Robots address the question of how a robot can be designed to acquire autonomously whatever it needs to realize adequate behavior in a complex environment. In-depth discussions of issues, techniques, and experiments in machine learning focus on improving ease of programming and enhancing robustness in unpredictable and changing environments, given limitations of time and resources available to researchers. The authors show practical progress toward a useful set of abstractions and techniques to describe and automate various aspects of learning in autonomous systems. The close interaction of such a system with the world reveals opportunities for new architectures and learning scenarios and for grounding symbolic representations, though such thorny problems as noise, choice of language, abstraction level of representation, and operationality have to be faced head-on. Contents Introduction: Toward Learning Robots * Learning Reliable Manipulation Strategies without Initial Physical Models * Learning by an Autonomous Agent in the Pushing Domain * A Cost-Sensitive Machine Learning Method for the Approach and Recognize Task * A Robot Exploration and Mapping Strategy Based on a Semantic Hierarchy of Spatial Representations * Understanding Object Motion: Recognition, Learning and Spatiotemporal Reasoning * Learning How to Plan * Robo-Soar: An Integration of External Interaction, Planning, and Learning Using Soar * Foundations of Learning in Autonomous Agents * Prior Knowledge and Autonomous Learning

Neuro-robotics is one of the most multidisciplinary fields of the last decades, fusing information and knowledge from neuroscience, engineering and computer science. This book focuses on the results from the strategic alliance between Neuroscience and Robotics that help the scientific community to better understand the brain as well as design robotic devices and algorithms for interfacing humans and robots. The first part of the book introduces the idea of neuro-robotics, by presenting state-of-the-art bio-inspired devices. The second part of the book focuses on human-machine interfaces for performance augmentation, which can seen as augmentation of abilities of healthy subjects or assistance in case of the mobility impaired. The third part of the book focuses on the inverse problem, i.e. how we can use robotic devices that physically interact with the human body, in order (a) to understand human motor control and (b) to provide therapy to neurologically impaired people or people with disabilities.

"This book argues that computational models in behavioral neuroscience must be taken with caution, and advocates for the study of mathematical models of existing theories as complementary to neuro-psychological models and computational models"--