Abstract

In this talk I pull apart some of the competencies needed to build “intelligent” self driving vehicles. This talk will discuss the architecture, operation and performance of the “Selenium” Autonomy System which, amongst others, is at the heart of the Oxford Robot Car the LUTZ Pathfinder project, and the Greenwich Gateway project. We will explain the heart of the system and explain how laser and image processing and machine learning techniques have been mixed to create the information engineering software which senses, plans and controls on the behalf of the autonomous vehicles. The talk will consider challenges and solutions arising from time varying environments, weather types, dynamic obstacles touching on calibration, data management, mapping, localisation planning and perception.

Bio: Paul Newman is the BP Professor of Information Engineering at the University of Oxford and an EPSRC Leadership Fellow. He heads the Oxford Robotics Institute (ori.ox.ac.uk) within the Department of Engineering Science which enjoys a world leading reputation in mobile autonomy – developing machines, robots and cars which map, navigate through and understand their environments. Paul’s focus lies on pushing the boundaries of navigation and autonomy techniques in terms of both endurance and scale. The Oxford Robotics Institute has developed a keen focus on intelligent transport for example the RobotCar www.robotcar.org.uk and enjoys collaborations with many industrial partners which provide exploitation opportunities to drive the research. In 2014 he founded Oxbotica – a spinout company focussed on Robotics and Autonomous Systems – and was elected fellow of the Royal Academy of Engineering with a citation for outstanding contributions to robot navigation.