Information, Control & Vision
Introduction to Information, Control & Vision Engineering.
This research area spans the domain sensing - control - planning - action. We use common theoretical tools and ideas, but work in an extraordinarily wide range of applications. Within this research cluster we find research on driverless cars, seeing machines, exo-planet detection, video and image understanding, biological networks, vehicle dynamics, wind energy, economic modelling and multi-sensor networks.
The group develops fundamental control methodologies and novel algorithms for important applications, in the areas of energy and automotive, aerospace, medical and process engineering, with sponsors such as Ferrari, ESA, Ricardo, Diamond Light Source and Vestas. The group also has a strong interest on systems and synthetic biology and large scale networked systems.
The Machine Learning Research Group focuses on probabilistic reasoning applied to problems in science, engineering and computing. We use the tools of statistical, and in particular Bayesian, inference to deal rationally with uncertainty and information in a number of domains including astronomy, biology, finance, image & signal processing and multi-agent systems, as well as researching the theory of Bayesian modelling and inference. Our research is diverse in terms of models and application domains, but united by the common theme of probabilistic inference and decision theory.
The Mobile Robotics Group (MRG) researches many aspects of mobile autonomy with a particular emphasis on perception and understanding of large workspaces. Lifelong infrastructure-free navigation lies at the heart of the group's research agenda. This is an information engineering problem encompassing lifelong learning, multi sensor fusion / perception, control, planning, state estimation and software engineering.
Research on sensor systems, visual geometry and active vision is aimed at sensing and interpreting 2-dimensional images in a 3-dimensional world, seeking patterns that can be traced and tracked. Google-like search methods are being devised to enable objects to be identified in that reconstructed world. Earlier research on multiple view geometry generated the spin-off company 2d3 whose software is widely used to "glue" special effects to a live-action background: this work gained the inventors an Emmy award in 2002.