As sensor resolution increases and costs decrease, the amount of data available on mobile robotics platforms is exploding. Unsupervised machine learning algorithms, and their ability to produce useful information without large labeled training sets, are an important tool for benefiting from this abundance. In this thesis the application of unsupervised learning to three subfields of mobile robotics is discussed. Tracking multiple moving objects from an unmanned aerial vehicle, road following in loosely-structured environments, and autonomous offroad navigation. The thesis focuses on building dynamic activity-based ground models for multi-object tracking, the combination of optical flow techniques and dynamic programming to estimate the location of a road, and the use of optical flow techniques to improve the quality of an autonomous robot's obstacle classification.

Proceedings from the annual Robotics: Science and Systems conference, presenting state-of-the-art research on the algorithmic and mathematical foundations of robotics, robotics applications, and robotics systems. Robotics: Science and Systems II spans all areas of robotics, bringing together researchers working on the algorithmic and mathematical foundations of robotics, robotics applications, and analysis of robotics systems. This volume presents the proceedings of the second annual Robotics: Science and Systems conference, held in August 2006. Papers report state-of-the-art research on topics as diverse as Legged Robotics, Reconfigurable Robots, Biomimetic Robots, Manipulation, Humanoid Robotics, Telerobotics, Haptics, Motion Planning, Collision Avoidance, Robot Vision and Perception, Bayesian Techniques, Machine Learning, Mobile Robots, and Multi-robot systems.

This book is the only account of what honeybees actually see. Bees detect some visual features such as edges and colours, but there is no sign that they reconstruct patterns or put together features to form objects. Bees detect motion but have no perception of what it is that moves, and certainly they do not recognize "things" by their shapes. Yet they clearly see well enough to fly and find food with a minute brain. Bee vision is therefore relevant to the construction of simple artificial visual systems, for example for mobile robots. The surprising conclusion is that bee vision is adapted to the recognition of places, not things. In this volume, Adrian Horridge also sets out the curious and contentious history of how bee vision came to be understood, with an account of a century of neglect of old experimental results, errors of interpretation, sharp disagreements, and failures of the scientific method. The design of the experiments and the methods of making inferences from observations are also critically examined, with the conclusion that scientists are often hesitant, imperfect and misleading, ignore the work of others, and fail to consider alternative explanations. The erratic path to understanding makes interesting reading for anyone with an interest in the workings of science but particularly those researching insect vision and invertebrate sensory systems.

Neuromorphic engineering has just reached its 25th year as a discipline. In the first two decades neuromorphic engineers focused on building models of sensors, such as silicon cochleas and retinas, and building blocks such as silicon neurons and synapses. These designs have honed our skills in implementing sensors and neural networks in VLSI using analog and mixed mode circuits. Over the last decade the address event representation has been used to interface devices and computers from different designers and even different groups. This facility has been essential for our ability to combine sensors, neural networks, and actuators into neuromorphic systems. More recently, several big projects have emerged to build very large scale neuromorphic systems. The Telluride Neuromorphic Engineering Workshop (since 1994) and the CapoCaccia Cognitive Neuromorphic Engineering Workshop (since 2009) have been instrumental not only in creating a strongly connected research community, but also in introducing different groups to each other’s hardware. Many neuromorphic systems are first created at one of these workshops. With this special research topic, we showcase the state-of-the-art in neuromorphic systems.

5G IoT and Edge Computing for Smart Healthcare addresses the importance of a 5G IoT and Edge-Cognitive-Computing-based system for the successful implementation and realization of a smart-healthcare system. The book provides insights on 5G technologies, along with intelligent processing algorithms/processors that have been adopted for processing the medical data that would assist in addressing the challenges in computer-aided diagnosis and clinical risk analysis on a real-time basis. Each chapter is self-sufficient, solving real-time problems through novel approaches that help the audience acquire the right knowledge. With the progressive development of medical and communication - computer technologies, the healthcare system has seen a tremendous opportunity to support the demand of today's new requirements. - Focuses on the advancement of 5G in terms of its security and privacy aspects, which is very important in health care systems - Address advancements in signal processing and, more specifically, the cognitive computing algorithm to make the system more real-time - Gives insights into various information-processing models and the architecture of layers to realize a 5G based smart health care system