Personal tools
You are here: Home Robocup Research

Research

About Our Research

Our research interests are centered on issues of effective autonomous decision making and strategic behaviour in continually changing worlds. An important component of our work is single and multi-agent robot learning. This is supported by solid foundations in robotic locomotion and full-body behaviours, on-board computer vision and communications.

Our 2011 qualification video gives an idea of the kinds of things our robots can do:

Qualification video image

Our Research Expertise

Strategic behaviour

Multi-agent collaboration

• Real-time control

• Autonomous agents

Sensor fusion

Computer vision

What does Informatics do that is at the leading edge of robotics?

World leading research in robotics is carried out at the Institute of Perception, Action and Behaviour (IPAB) within the School of Informatics. We have expertise in robot learning: examples being robots that learn from imitation and through self-refinement from training; rehabilitation robotics and prosthetics, especially our work with Touch Bionics on the iLimb; bioinspired and biomimetic robotics: including robots that simulate insect behaviour and collective intelligence; as well as robot design: examples include development and control of novel, variable impedance actuators. 

What industry demands or societal needs are being satisfied by this research?

Most of our work on adaptive robotics is at the heart of enabling robots to move from industry floors to our everyday living space. Development of variable stiffness (or soft) yet powerful robot actuators are crucial to the ability of human and robots to cooperate in achieving a task. This can take the form of force augmentation with exoskeletons (lifting heavy loads, teleoperation) or devices that can help with post-stroke movement rehabilitation, aiding greatly the work of a physiotherapist. Robot soccer, as a platform, forces us to work on real world issues such as the need for fast reactions, ability to deal with noisy sensing, dynamic obstacles and cooperative behaviour -- all crucial for deployment of robots in the real world. Our work on reconfigurable robots and distributed sensing-acting can be used in various data collection or surveillance applications in remote areas. The spin-off technology from these developments have the promise to impact much wider fields including field and service robotics, interactive input devices, automated transport management systems, automated trading agents, etc.

What opportunities does this research create for Scotland in the future?

Our expertise in the area of machine learning and adaptive control make us a unique partner for many technology companies (such as HONDA, SONY) and several international research centers (e.g., DLR German Aerospace Labs). This work has also attracted interest from some leading international SMEs in rehabilitation robotics looking to expand in the European market, who are considering Edinburgh as their UK base. Most importantly, by exploiting the excellent Knowledge Transfer and entrepreneurial  culture developed at eh School of Informatics through Scottish Enterprise funded ProspecKT program, we have had several active and prospective spin outs based on the technology we are developing. Last but not the least, Scotland have a football team in the world cup finals!

Selected Publications:

  • 2013  Qualification Team Description Document - Standard Platform League [pdf]
  • 2012  Qualification Team Description Document - Standard Platform League [pdf]
  • 2011 Qualification Team Description Document - Standard Platform League [pdf]
  • 2011 Qualification Team Description Document - 2D Simulation League [pdf]
  • I. Havoutis, S.Ramamoorthy, Constrained geodesic trajectory generation on learnt skill manifolds. In Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010. [Finalist, RoboCup Best Paper Award]
  • E. Ho, T. Komura, S. Ramamoorthy, S. Vijayakumar, Controlling humanoid robots in topology coordinates. In Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010.
  • C. Towell, M. Howard and S. Vijayakumar, Learning Nullspace Policies.In Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010.
  • B. Rosman, S. Ramamoorthy, A game theoretic procedure for learning hierarchically structured strategies. In Proc. IEEE International Conference on Robotics and Automation, 2010.
  • I. Havoutis, S. Ramamoorthy, Geodesic trajectory generation on learnt skill manifolds. In Proc. IEEE International Conference on Robotics and Automation, 2010.
  • I. Havoutis, S. Ramamoorthy,Motion synthesis through randomized exploration on submanifolds in configuration space. In J. Baltes et al. (Eds.): RoboCup 2009, Lecture Notes in Artificial Intelligence, Volume 5949, pp. 92-103. Springer Verlag, Heidelberg (2010).

 


Document Actions