Nick Hawes is senior lecturer in intelligent robotics at the University of Birmingham School of Computer Science and an expert on what makes robots and humans tick. ‘Robots and humans: when two societies meet’ was his Lord Kelvin Award Lecture at the British Science Festival and a timely lesson in the difficulties of defining human skill and programming robots to master that skill.
When is a robot a ‘nobot’ he asks? Well, that of course depends. In its simplest terms a robot is any machine or device that operates automatically with ‘humanlike’ skill. Robots play a huge part in modern manufacturing systems but as predictive or intuitive systems they are a dead loss. They do what we tell them to do extremely accurately and quickly, but nothing else.
Pressure on society
Humankind will eventually develop complex robots able to think and analyse problems on their own terms and that, says Nick, cannot come too soon given the pressures facing society. In terms of global population we will see the greatest growth in people aged 55 and over in the next 30 to 40 years. Inevitably that means increased pressure on healthcare and elderly support systems and some of that could in future be delivered by intelligent robots.
Nick demonstrates a room-mapping robot called Dora (the explorer) whose ability to map its environment using an array of sensors – including an X-box Kinect – and a laptop, is an essential part of the CogX project to develop domestic robots. Dora creates its own map of a house with a laser scanner and can even recognise and categorise objects to find out what kind of room it is in.
Another robot nicknamed Curious George can be taught to recognise and store information about objects. The video below shows what a charming beast the curious robot is. George becomes better at recognising objects under tuition and can then ask questions about objects it does not recognise, pointing to each object in turn: it is ‘curiosity driven’.
Of course there are robotics research groups working on more complex and time demanding tasks than identifying whether a mug is blue. The Oxford Mobile Robotics Group is focused on self-driving cars that engage in life-long learning about their environment. Google has demonstrated its own self-driving cars in the United States.
We will look back at engineering feats such as Honda’s Asimo robot (right) in a couple of decades and laugh at how basic and dated the technology looks. Still, a simple video presented by Nick showing a robot that could make a pancake and serve it up – including opening the fridge for ingredients and cupboards and drawers for cutlery and plates – draws a slight gasp from the audience at the smoothness and ease of the robot’s movements.
Building robots that automatically operate with humanlike skill is still not an easy business. For anyone interested in Nick’s research check out the CogX project, CoSy, the Intelligent Robotics Lab, Media Lab Europe and the Common Sense project.
Turning a robot ‘on’ and ‘off’ is known as ‘powercycling’.
Biomimicry in robots
Another interesting trend we see is in systems that copy nature. Biomimicry is a discipline that relies on evolution as a guide to the most efficient ways of doing things. Sridhar Ravi and his colleagues at Harvard University have been filming the way bumblebees react to the wind and turbulence in a wind tunnel and say their research will help them to design ‘micro air vehicles’ that remain stable in bad weather. “We are currently conducting more experiments including flying other insects in similar wind conditions and identifying the influence of pollen and/or honey on flight stability in the bees,” said Dr Ravi.