The Problem of Motion
Imagine you are walking downtown in the middle of rush-hour. You make your way down the crowded sidewalk, easily moving side to side to avoid a collision with another person or object. It comes naturally to us, this ease of motion, and even as toddlers we can do it. It may shock you, then, to discover that the achievement of such autonomous, intelligent motion in robots has been one of the most challenging problems in the history of robotics. But that nut may have just been cracked!
Researchers at the European-funded research consortium Decisions in Motion decided to approach this problem three years ago, and as of this year they are now announcing stunning breakthroughs. The success of their team is a cooperative effort of both neuro- and cognitive scientists who study biological visual systems in humans and high-order mammals such as primates, as well as computer scientists and roboticists.
This rare amalgamation of many disciplines has proved very fruitful to the consortium. The biological researchers are engaged in a dialog with the roboticists and programmers, who then incorporate their findings into neural networks and mobile robots.
“The approach paid off. Decisions in Motion has already built and demonstrated a robot that can zip across a crowded room guided only by what it ’sees’ through its twin video cameras… The computerised brain controls the behaviour of a wheeled robotic platform supporting a movable head and eyes, in real time. It directs the head and eyes where to look, tracks its own movement, identifies objects, determines if they are moving independently, and directs the platform to speed up, slow down and turn left or right.” source
The structure of the robotic “brain” that is being developed by Decisions in Motion is very much inspired by the architecture of a high-order mammal’s biological brain, the flow of information following similar pathways through three layers, or subsystems, that then control the behavior of the robot. The secret of the solution lies in the discovery of anticipatory feedback as observed in organisms such as ourselves. When you make your way through that busy street, you are not plotting the present course so much as you are predicting the next dozen steps. I think that anyone who has misstepped can completely relate; when a person who is walking stumbles, one is already mentally taking those next few steps and it almost comes as a surprise that one is suddenly not along the plotted course anymore and is, instead, on one’s way towards the ground.
The work done by Decisions in Motion will not only benefit movementally challenged bots, but will also dictate further development in the field of electronic mobility assistance for the visually or cognitively impaired. Keep up the good work, guys!
Posted in Features, Real Life Robots | No Comments »