Helping Robots Understand Their Limitations And Adapt
Lee Rannals for redOrbit.com — Your Universe Online
Researchers writing in the journal International Journal of Robotics Research talked about building robots that are aware of their own limitations.
According to researchers at MIT´s Computer Science and Artificial Intelligence Laboratory (CSAIL), if robots are ever going to move around a constantly changing environment effectively, they will need to be more aware of their own limitations.
Leslie Pack Kaelbling, the Panasonic Professor of Computer Science and Engineering at MIT, says most successful robots tend to be used either in fixed, carefully controlled environments or for simple tasks like vacuuming a room. However, taking these robots to a more cluttered, dynamic environment will require more research.
“I would like to make a robot that could go into your kitchen for the first time, having been in other kitchens before but not yours, and put the groceries away,” Kaelbling says.
The team claims they have designed a system that is able to constantly calculate the robot’s level of uncertainty about a given task, such as the whereabouts of an object, or its own location within a room.
The system uses a module called the “state estimation component,” which calculates the probability of any given object being what or where the robot thinks it is. With this, the robot is not sufficiently certain that an object is the one it is looking for, so it knows it needs to gather more information before taking action.
“It´s thinking always about its own belief about the world, and how to change its belief, by taking actions that will either gather more information or change the state of the world,” says Kailbling.
The system simplifies the process of developing a strategy for a given task by making up its plan in stages as it goes along. Kaelbling says they will have to be computing for a very long time to ensure they have a complete strategy formulated before being executed.
Instead of one big, complicated strategy that consumes a considerable amount of power and time, the robot can make many smaller plans as it goes along. The drawback from this is that it leads the robot into making dumb mistakes.
“As we try to get robots to do bigger and more complicated things in more variable environments, we will have to settle for some amount of suboptimality,” said Kaelbling.
She says the system could be used to build more flexible industrial devices, or in disaster relief situations.
We are still a long way away from the world where Will Smith helps stop rogue robots from initiating martial law, but advances are being made in this field on a weekly basis. Last week, researchers announced the advancement in a specific robotic field they created called “Terradynamics.” In this field, researchers try and understand how small-legged robots move and interact with complex materials like sandy beaches.
It is only a matter of time now before a robot capable of moving around on a beach will have the wherewithal to also know how to understand and calculate the changing tide.