©School of Informatics, University of Edinburgh and Edinburgh Inspiring Capital

'cubed' webzine

Evi knows what you want
New Adaptive Robots
Thyme heals
Dance And Parkinson’s Disease
High speed light harvester
Secrets of Regeneration
Previous issues

The University of Edinburgh

School of Informatics

Professor Sethu Vijayakumar

Statistical Learning and Motor Control Group

cubed logo © British Council

New Adaptive Robots

Leonardo Da Vinci's robot
In the 15th century, Leonardo da Vinci designed a mechanical device resembling a knight in armour. The mechanisms inside were designed to make the knight move as if there was a real person inside. In the 21st century scientists are continuing this vision of making robots more human-like. While robots may never have a heart like the Tin Man in the movie, Professor Sethu Vijayakumar and his team at the University of Edinburgh want to make robots that not only move more like humans, but are also able to interpret the context of their actions and improve by continuously learning from their experiences just like humans do.

Adaptable and dynamic
Their innovative work is focussing on two things: firstly making the robot capable of adapting to new situations, learning ‘on-the-fly’ to better perform a goal; secondly, to make the robot’s movement more dynamic.

As Vijayakumar explains, ‘when we do everyday tasks like walking or running, we exploit the natural dynamics of our body. To mimic this capability in robots and generate more efficient movements is a significant challenge.’

Their walking robots are unique in being able to simultaneously adjust the stiffness, 'damping' (the amount of resistance experienced) and position at the hip, knee and ankle joints. Using this ability allows the robots to walk with human-like energy efficient gaits. Adjusting their damping and stiffness enables the robot to deal with differences in terrain and unexpected changes, as well as to undertake energetic tasks like jumping, throwing and even dancing.

©School of Informatics, University of Edinburgh and Edinburgh Inspiring Capital

Wearable robots
For the team the biggest challenge is developing technologies that can scale. As Vijayakumar points out, ‘as we consider more complex, useful and realistic robots, if the technology we develop to control them and adapt takes forever to compute, then that’s no good. Another challenge is robust sensing. Humans are good at combining all their senses to perceive the world, but issues such as noise and partial observability makes it hard to translate this capability to robotic sensors.'

Designing the robots to cope like humans continues. The team has also built prototype prosthetic robotic devices, which can interface directly with the human body and be controlled using human muscle signals. Wearable robots are science and not fiction.

LearnEnglish Science activities
Why not do a language activity based on this cubed story, New Adaptive Robots? You can double-click on any word on this page for a dictionary definition.

DCSIMG