on 26-09-2007 19:08

Editor's rating

Views 939

Favoured 52

A timid-looking four-legged robot about the size of a Chihuahua might seem like an unlikely companion for soldiers of the future.

Yet the robot, called LittleDog, could ultimately help researchers create more sophisticated robotic assistants for military personnel, including automated "pack-mules" capable of hauling heavy loads over tough terrain.

This is because LittleDog is remarkably agile for a robot when faced with treacherous, uneven terrain. Researchers are also fine-tuning its movement to be even faster and more animal-like over rough terrain.

LittleDog was created for the US Defence Advanced Research Projects Agency (DARPA) by US robotics company Boston Dynamics.

And now DARPA has selected six university research teams, including ones at MIT and Stanford, to compete to develop the best algorithms for controlling the robot puppy. The agency hopes this will help identify the best adaptive strategy for moving over irregular surfaces.

The robot has three motored joints on each leg, and its movements are controlled precisely by an on-board computer. An internal gyroscope lets the robot sense its orientation, while an external motion-capture system monitors the precise position of each limb and joint as it moves.

Monthly testing

The six teams have each been given a LittleDog and a section of near-identical artificial terrain for the robot to cross.

Every month, following rigorous testing in their own labs, each team sends its control algorithms to DARPA, and the agency then tests these on an identical LittleDog and the same fake terrain to determine which lets the robot move fastest.

But over the course of 2008, new algorithms will be tested by DARPA on more unpredictable, previously unseen, terrain. So the algorithms used will need to quickly identify footholds and the best route across the terrain. At the end of that year, an overall winner will be announced.

Dynamic advances

"It's getting a bit more competitive," admits Jerry Pratt, a member of the LittleDog team at the Institute for Human and Machine Cognition (IHMC) in Florida.

Pratt notes that most LittleDog teams are starting to move from a static approach that involves keeping three of the robot's legs rooted to the ground to a more dynamic one that sees only two legs planted at any time. "It's a lot trickier because you have to take dynamics into consideration," he told New Scientist. "Walking becomes more like controlled falling."

Although the current set-up uses an external motion- capture system, LittleDog's descendants will ultimately need to examine the terrain ahead for themselves. "It would have to identify the best places to step," Pratt says.

Boston Dynamics has previously demonstrated a much larger four-legged robot called BigDog. Internal sensors and motors allow this robot to rapidly regain its balance after slipping or being pushed, but BigDog is unable to tackle the kind of irregular terrain faced by LittleDog.

Trial and error

"Legged locomotion on extreme terrain is a very complex task for robots," says Dennis Hong, an expert on walking robots at Virginia Tech, US, who is not involved with any of the LittleDog projects.

Hong adds that adaptive learning offers the best hope of tackling such a complex problem. "Think of a human baby learning how to crawl," he says. "Initially, the baby does not know how to coordinate the complex motion of its limbs for crawling, but learns by trial and error, and eventually learns how to walk and even run."

Max Lungarella, a robotics researcher at the University of Zurich in Switzerland, says the project provides a great opportunity for comparing different approaches. "What is really interesting about the whole project is the idea of a common research platform," he told New Scientist. "A lot of research in robotics is done on platforms built ad-hoc."

Lungarella also believes robots capable of covering such hazardous terrain could find plenty of uses: "from search and rescue operations to exploration of unknown areas, and maybe an alternative form of transport for humans in terrains inaccessible by vehicles on wheels like mountains or forests."

Robert Richardson, at the University of Manchester, agrees. "Robotics is generally moving away from the three D's (dull, dirty and dangerous), towards the three E's (education, entertainment, and everyday)," Richardson says. "I think there will be no shortage of applications for such a system."

Recommend this article...

Last update: 26-09-2007 20:27

Published in : Bot News, General

Users' Comments (0)

Add your comment