Robots on the march to walking like humans

A psychological theory could kickstart improvements in the way robots are able to walk, thanks to an unique collaboration between a clinical psychologist, robotics engineers and a robotics entrepreneur at the University of Manchester.

Photo
The self-balancing robot used to demonstrate how 'Perceptual Control Theory' can help robots to walk in a more human-like way
Source: University of Manchester

It analysed what happens when standard algorithms driving a self-balancing robot - made from simple Lego - were replaced with those based on 'perceptual control theory'. The theory was encoded into the little droid, allowing it to control what it sensed so that it moved around more effectively, just as humans and other animals can.

Though the robot moves on two wheels, it is an 'inverted pendulum', which requires nimble balancing in a similar way to how our bodies are kept upright when we walk. So, the better the robot can balance, the better prepared it will be for walking like a human.

In the study, the more lifelike robot balanced more accurately, more promptly and more efficiently than its rivals by assessing its environment at least 100 times a second. It also moved to a new location, even when disturbed by sideways nudges, more effectively than its competitors. When algorithm was programmed into the robot, it appeared much less stable and wobbled excessively. Though perceptual control theory has been used widely in psychological therapies, education and parenting interventions, this is the first time that using it in robots has been compared in a 'head-to-head' test.

The new study, compared the same inverted pendulum robot programmed and tuned with three different software controllers. Two of them, proportional control and LQR, are widely used by engineers to build the latest robots. The third, perceptual control theory, was originally derived from engineering, but it takes the 'insider's perspective', specifying the 'desired inputs' or 'needs' of the robot.

Dr Warren Mansell, Reader in Clinical Psychology at The University of Manchester said: "Although this is early work, it is tantalising to see how a scientific theory used to help people with mental health problems might actually help engineers to improve their designs of artificially intelligent devices. "Robots are yet to match the abilities of those in science fiction hits like Star Wars and Blade Runner and none have mastered walking on two feet. But the use of the theory could really enable the transformation of robots into more lifelike machines."

Dr Simon Watson, Senior Lecturer in Robotic Systems at The University of Manchester said: "Nature has developed the most complex machines we know, so being able to implement algorithms inspired by them is an important step in our own creative development capability.

Thomas Johnson, the PhD student who built and tested the robot said "This work has demonstrated the success of controlling robots with perceptual control theory. This paper is a demonstration of how engineers in robotics can find inspiration from the living world."

Computer technologist Dr Rupert Young said: "This research is a peek at a radical new way of understanding how to build robotics systems, that are dynamic and adaptive despite the chaotic, unpredictable nature of the real world. Based upon an elegant and natural approach, this paradigm holds the promise of developing far more sophisticated, autonomous robots"

Subscribe to our newsletter

Related articles

Robots uncover connections between mind and walking control

Robots uncover connections between mind and walking control

Researchers used robots to study how our brains adjust to changes in our walking strides, gaining insights that could be used to develop better physical rehabilitation programs.

An ankle exoskeleton aids running

An ankle exoskeleton aids running

Researchers find that a motorized device that attaches around the ankle and foot can drastically reduce the energy cost of running.

Smart algorithm analyses whole brain vasculature

Smart algorithm analyses whole brain vasculature

Scientists have developed a technique for visualising the structures of all the brain's blood vessels including any pathological changes.

Fighting hand tremors with AI and robots

Fighting hand tremors with AI and robots

Researchers have tapped AI techniques to build an algorithmic model that will make the robots more accurate, faster, and safer when battling hand tremors.

AI challenge aims to improve mammography accuracy

AI challenge aims to improve mammography accuracy

AI techniques, used in combination with the evaluation of expert radiologists, improve the accuracy in detecting cancer using mammograms.

Children benefit from learning with social robots

Children benefit from learning with social robots

Research has shown that explaining aloud what you are learning, whether to yourself or to someone else, leads to a better understanding of the subject.

Neural networks support endovascular stroke therapy

Neural networks support endovascular stroke therapy

Researchers have been investigating whether artificial intelligence might be used to steer a catheter automatically and reliably to a blocked blood vessel.

Robot helps schoolkids break bad hygiene habits

Robot helps schoolkids break bad hygiene habits

A robot which encourages kids to wash their hands has helped pupils at a remote Indian primary school take a fresh approach to hygiene.

AI identified decision-relevant patterns in EEG of coma patients

AI identified decision-relevant patterns in EEG of coma patients

A research team has succeeded in identifying specific patterns in Electro-Encephalogram (EEG) analyses that the deep learning network uses for making prognosis decisions.

Popular articles