CHIMP Disaster Robot Aims To Win Robotics Challenge
SONY DSC
The DARPA Robotics Challenge final features 25 robots performing tasks like sawing, driving vehicles and maneuvering in a simulated disaster zone. The challenge was inspired by the Fukushima-Daiichi nuclear disaster. It is pushing forward development of robots and technology to be used in disaster areas and hazardous environments. For the winning team, a prize fund of $2 million is up for grabs. The DARPA challenge in 2004 created self-driving vehicles and this years final could lead to advancements for robotics in emergency response, warehouse automation and also manufacturing areas.
The CHIMP, or CMU Highly Intelligent Mobile Platform, was created by a team at Carnegie Mellon University in order to engineer a robot that could traverse uneven and unpredictable environments. The head of software Clark Haynes said that humanoid robots’ greatest challenge is being able to balance. However, the CHIMP has static stability so it never has to balance and doesn’t risk falling over.
The CHIMP relies on five main areas to work effectively and each area has been carefully developed by the Carnegie Mellon University team. First on the list is situational awareness, which the robot gains from having six cameras and light radar to allow an operator to see the full extent of the area which surrounds the robot. CHIMP can work autonomously but works the best when it is receiving remote instructions from an operator.
CHIMP also has strategic slouch to help the robot with stability but also provide its arms with more reach. The shoulders angle forward which allows the 4.3ft long arms to reach things that would otherwise be too far away. It also helps the robot fit through doorways as it is tall and wide.
CHIMP can work alone and doesn’t have to rely on human instructions, which may have trouble getting through to the robot because of trouble with signals. The CHIMP can control its own hands and find things like power tools using computer vision, it then uses algorithms to hold and work the tool or object.
The robot has legs that look reasonably similar to a humans but it actually doesn’t walk. It instead uses rubber tracks to drive itself over uneven surfaces with stability like a tank. This means it has less risk of falling over, but it can also crawl on all fours if the surfaces are particularly rough.
One of the main things CHIMP relies on is its balance which makes it stable and prevents the robot from falling over in areas where other robots would fear to tread. The CHIMP is statically stable so it doesn’t need to use software, energy or tools to keep its balance. In addition, the robot locks any joints that aren’t being used so it can preserve energy for times when it needs it most. CHIMP also has a brake which locks it in place if the power it relies on is cut off.
This year, the DARPA robotics challenge is being held in California where 25 of the best robotics organisations in the world will come together to compete on the disaster response course. The course is designed to be incredibly difficult, so the teams competing with robotic systems and software will have to work hard to ensure they have a chance of winning the huge cash prize. The teams have a short timeline but are able to create innovative hardware, software, control interfaces and sensors that lead to new developments in robotics for the future, even outside the area of disaster response.
The finals of the competition take place on 5th-6th June and the robots are required to perform consecutive tasks with sketchy communications between robots and operators. The team who wins will receive the biggest prize of $2 million, the runner up will be given $1 million, while the team who comes third will gain $500,000.
CHIMP first took part in the DARPA Robotics Challenge in 2013 but for this year’s competition it benefits from improved mobility, autonomy, speed and ease of use. The Carnegie Mellon University team who controls the robot in the competition is called Tartan Rescue, and they will be hoping that they’ve improved on previous designs and software enough to win the grand prize this year.
