Active Robotics Laboratory

The Active Robotics Laboratory (ARL) carries out advanced research in robot systems. Current areas of research includes space robotics, networked robotics, human-robot interaction & telerobotics, and online robot systems. We also have a special interest in robotics education. The ARL contact details are below for further information click on read more.

Gerard McKee

The Multi-Tasking Rover (MTR) Lead: Antonios Bouloubasis: This research involves the exploration of an innovative new concept for rover systems, combining capabilities for science and robotic work through a combination of a mobile base possessing high reconfigurability for coping with unstructured terrain and and mountable science or tool packs that incorporate science instrumentation or robotic tools respectively. The mechanical system had been developed and tested and the electronic systems, including sensors and controls are currently being integrated. We have a paper accepted for publication at the forthcoming ICRA-07.

Magnetic Compliance for Legged Robots Lead: Richard McElligott: This research investigates an innovative new concept for providing compliance in legged robots, with potential for application in a wide range or areas. Current work is focusing on building a model of the compliance in Matlab. A physical complinace module has been developed in order to tune the model. A design for a hexapod legged robot has been develop and is currently nearing the stage for implementation. It will incorporate the multiple miniature versions of the compliance unit. Experimental studies will investigate the operation and operating parameters of the units under different gait patterns and speeds.

Online Robot Arena Lead: Andrew Gatward: This research involves the development of an online robot arena for robotics education. The project includes the development of both a robot system, which we call the Athlete, and the arena, which we call the Stadium, in which it will operate. The key technical challenges include the development of the electronic systems, from board level design through to system integration, the provision of multple stereo visual feeds from the robot, and the provision of an infrastructure to support student access and assignments.

VISOR: Visual Intelligence for Space Operations Rovers Lead: Xiaofei Nie: The purpose of this research is to improve exploration rover autonomy by designing assistive visual acts for navigation through unstructured terrains to fulfil predefined goals. This research aims to integrate work within the Active Robotics Laboratory on Visual Acts and in the Computational Vision Group on stereo vision and image segmentation to enhance the visual interrogation of unstructured terrain in the context of rover autonomous navigation to selected science targets.

DINERO: Distributed Networked Robotics Lead: Duncan Baker: This work is now completed - Duncan successfully defended his thesis in 2005. The purpose of this research was to investigate a network-centric modular approach to the configuration of distributed robot architectures. A framework, DINERO, was developed to test the approach experimentally. The key ingredients of the approch is the distinction between task modules and resource modules. Resource modules provided primitive functionality, such as sensing or control. Task modules are generic, specifying constraints on resource modules that allow the task module to be instantiated as a resouce module. Robot architectures are constructed in a hierarchical manner, primitive resource modules and instantiated task modules providing the basis for instantiation higher-order functionality through higher-order task modules. Modules communicate via network interfaces and hence can be distributed about multiple platforms, mobile or fixed - hence the reference to 'distributed robot architectures'.

Space Robotics is a strengthening theme within the laboratory, having its origins in original work in the development of the Visual Acts model for cooperative intelligent viewing during remote teleoperation, in collaboration with Paul Schenker of NASA/JPL. Work over the last seven years has included the development of a set of rovers to explore cooperative transport tasks, leading to the development of the MTR. Other work has included the development of a GeoBot system to investigate the expertise that scientists bring to bear in exploring remote environments such as the surface of Mars. Recent work has also included the investigation of magnetic complience for legged science robots and we have also now developed stronger links with our colleagues in the Computational Vision Group to look at the role of vision in science target selection and rover navigation.

Networked Robotics is a continuing theme that has developed over many years. We are currently looking at more generalised fameworks for combining distributed sensing, control and computational modules into a design environment for distributed robot architectures. The research has links to such areas as teleoperation, distributed robot systems, field robot systems, reconfigurable robots and to wider areas including intelligent ambient environments, embedded and embodied robotic intelligence, and sensor networks. A recent paper "What is Networked Robotics?" will be available here for download shortly.

Robotics Education has been a concern of the ARL since its inception. The first projects were motivated by the goal of providing a robotics laboratory facility online so as to provide wider access to robot systems for many more students than normally possible. The key observation here is that when you put students in a lab to program a robot practically all of their time is spent debugging their code and only a small amout of time actually using the robots. The solution we came up with in 1992 was to put the robot system online, provide a video feed so the students could see what was happing to the robot, and provide a control server that allowed them to control the robot. Their task was then to provide the intelligence for a basic control strategy that incorporated a stop-look-think-act type of cycle. A number of pojects were developed and techniques have been refined to improve not only the student experience but aslo monitoring and assessment of the student work. The challenge of using online robots in robotics, and indeed artificial intelligence education has raised the wider question of robotics education. This is one of the issues were are addressing now more earnestly. More to follow.... References to recent and forthcoming publications will be provided shortly.

Contact :
Gerard McKee,
PhD Senior Lecturer in Networked Robotics
Computer Science School of Systems Engineering Reading,
Berkshire, RG6 6AY
United Kingdom
Email: g.t.mckee@reading.ac.uk
Tel: +44 (0)118 378 8617