Knowing the unknown: Researchers work to build robots' awareness of their own limitations

Mar 27, 2013 by Helen Knight
Credit: Allegra Boverman and Christine daniloff/MIT

Robot butlers that tidy your house or cook you a meal have long been the dream of science-fiction writers and artificial intelligence researchers alike.

But if robots are ever going to move effectively around our constantly changing homes or workspaces performing such complex tasks, they will need to be more aware of their own limitations, according to researchers at MIT's and Artificial Intelligence Laboratory (CSAIL).

Most successful robots today tend to be used either in fixed, carefully controlled environments, such as , or for performing fairly simple tasks such as vacuuming a room, says Leslie Pack Kaelbling, the Panasonic Professor of Computer Science and Engineering at MIT.

Carrying out complicated sequences of actions in a cluttered, dynamic environment such as a home will require robots to be more aware of what they do not know, and therefore need to find out, Kaelbling says. That is because a robot cannot simply look around the kitchen and determine where all the containers are stored, for example, or what you would prefer to eat for dinner. To find these things out, it needs to open the cupboards and look inside, or ask a question.

"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.

And in a paper recently accepted for publication in the International Journal of Robotics Research, she and CSAIL colleague Tomas Lozano-Perez describe a system designed to do just that, by constantly calculating the robot's level of uncertainty about a given task, such as the whereabouts of an object, or its own location within the room.

Uncertainty principles

The system is based on a module called the state estimation component, which calculates the probability of any given object being what or where the robot thinks it is. In this way, if the robot is not sufficiently certain that an object is the one it is looking for, because the probability of it being that object is too low, it knows it needs to gather more information before taking any action, Kaelbling says.

So, for example, if the robot were trying to pick up a box of cereal from a shelf, it might decide its uncertainty about the position of the object was too high to attempt grasping it. Instead, it would first take a closer look at the object, in order to get a better idea of its exact location, Kaelbling says. "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."

The system also simplifies the process of developing a strategy for performing a given task by making up its plan in stages as it goes along, using what the team calls hierarchical planning in the now.

"There is this idea in AI that we're very worried about having an optimal plan, so we're going to compute very hard for a long time, to ensure we have a complete strategy formulated before we begin execution," Kaelbling says.

But in many cases, particularly if the environment is new to the robot, it cannot know enough about the area to make such a detailed plan in advance, she says.

Baby steps

So instead the system makes a plan for the first stage of its task and begins executing this before it has come up with a strategy for the rest of the exercise. That means that instead of one big complicated strategy, which consumes a considerable amount of computing power and time, the robot can make many smaller plans as it goes along.

The drawback to this process is that it can lead the robot into making silly mistakes, such as picking up a plate and moving it over to the table without realizing that it first needs to clear some room to put it down, Kaelbling says.

But such small mistakes may be a price worth paying for more capable robots, she says: "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."

In addition to household robots, the system could also be used to build more flexible industrial devices, or in disaster relief, Kaelbling says.

Ronald Parr, an associate professor of computer science at Duke University, says much existing work on planning tends to be fragmented into different groups working on particular, specialized problems. In contrast, the work of Kaelbling and Lozano-Perez breaks down the walls that exist between these subgroups, and uses hierarchical planning to address the computational challenges that arise when attempting to develop a more general-purpose, problem-solving system. "What's more, it is demonstrated on a practical, general-purpose robotic platform that could be used for domestic or factory work," Parr says.

Explore further: Socially-assistive robots help kids with autism learn by providing personalized prompts

Related Stories

Robots could improve everyday life, do chores

Sep 21, 2010

(PhysOrg.com) -- They're mundane, yet daunting tasks: Tidying a messy room. Assembling a bookshelf from a kit of parts. Fetching a hairbrush for someone who can't do it herself. What if a robot could do it ...

Recommended for you

Top South America hackers rattle Peru's Cabinet

53 minutes ago

The Peruvian hackers have broken into military, police, and other sensitive government networks in Argentina, Colombia, Chile, Venezuela and Peru, defacing websites and extracting sensitive data to strut ...

User comments : 12

Adjust slider to filter visible comments by rank

Display comments: newest first

Lurker2358
1 / 5 (3) Mar 27, 2013
Robots smart enough to cook a meal will kill about half of the food prep industry, maybe more. This would cause an unemployment nightmare both in terms of loss of management level jobs and part time and summer jobs for young people and already under-employed people; World wide.

We're talking totally disruptive technology beyond any one invention in the past several decades.
Lurker2358
1 / 5 (1) Mar 27, 2013
"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.


Going to be hard, IMO.

Things in the kitchen aren't always exactly where they "should" be. Making the Robot know when it's appropriate to just put something where it "fits" as opposed to putting it where it "belongs," as humans do sometimes, will be tricky. Too much leeway and it will just end up putting stuff randomly; Too little and you could end up with the robot in some idiotic endless loop.

Every human is different; Let's say you have two cabinets: one has several canned vegetables, and a can of chili already in it. The other has a few packages of pasta.

The robot has opened a grocery bag and found two items:
One is diced tomatoes with peppers.
Two is a jar of pasta sauce.

I would put the diced tomatoes with the canned goods, and the sauce with the pasta...others may not...
Lurker2358
1 / 5 (2) Mar 27, 2013
Additionally, what if the cabinet is already full? Or the shelf where the tomatoes are is full, but there is an extra can of tomatoes.

Will the robot be smart enough to put it in the nearest empty space, which is the right thing to do, even though it's "wrong", instead of going into some idiot loop?

Another scenario:

Two shelfs, one has random canned fruits, the other has random canned vegetables.

The robot opens a bag and finds a can of tomatoes. Where should it put the tomatoes?

Even though the tomatoes are a fruit, they should go with the vegetables, because we tend to use tomatoes in combination with vegetables, i.e. in soups, pastas, Mexican dishes, etc.

Also, when you shop for groceries and you buy a bunch of random stuff in bulk if it's on sale, which doesn't fit anywhere, what will the robot do?

Will it be smart enough to look for space in your garage or a cabinet which isn't normally used for food? We have random stuff mixed in with the pots and such to save space...
muggins
2 / 5 (1) Mar 27, 2013
Adapt and overcome. If a robot takes over a mundane task and puts people out of jobs then so be it. These people can then go on to get an education or move on to more mentally demanding jobs. We should not stifle innovation because of job losses. Natural selection in the work place.
Eikka
not rated yet Mar 27, 2013
Will the robot be smart enough to put it in the nearest empty space, which is the right thing to do, even though it's "wrong", instead of going into some idiot loop?


That's where the robot asks you "Where do you want me to put the tomatoes?" or simply decide to re-order your cabinet.

These people can then go on to get an education or move on to more mentally demanding jobs.


What if they are not capable? Half the population has an IQ of 100 or less.

You're really going to make a lot of people redundant this time, with no proper alternative work because there's no need for 100 million street artists or hairdressers. Plus, it's bad for the people in the more mentally demanding jobs as well, because there will be more competition and the wages will go down.
muggins
1 / 5 (1) Mar 27, 2013
A lot of people are in dead end mundane jobs because they never tried at school or lacked confidence. Basing it off IQ is a flawed method. You can improve your IQ, you are not born with it. University is a great place to get people interested in learning which is what it comes down to.
Noumenon
2.1 / 5 (7) Mar 27, 2013
Robots smart enough to cook a meal will kill about half of the food prep industry, maybe more. This would cause an unemployment nightmare both in terms of loss of management level jobs and part time and summer jobs for young people and already under-employed people; World wide.


If it is economically unsustainable that robots take over certain segments of labour, then it is unsustainable, thus will not occur.

If it is sustainable then that particular job will become obsolete, as standard of living evolves.

There used to be horse carriage makers and horse shoe black smiths,.... when the automobile was developed some complained that it would put them out of work, and wanted such progress to be stopped. Capitalism is self regulating, children.
nowhere
not rated yet Mar 28, 2013
Adapt and overcome. If a robot takes over a mundane task and puts people out of jobs then so be it. These people can then go on to get an education or move on to more mentally demanding jobs. We should not stifle innovation because of job losses. Natural selection in the work place.

A robot smart enough to cook food will be able to perform a massive amout and variety of "mundane" jobs. Fast forward 20-100 years and now you have almost no jobs that the most advanced robot cannot preform. How do you compete with corporations who can send in a robot to do your job better than you and at a fraction of your wage, while still working 23 hours a day 7 days a week? The entire economic system will need to be redesigned, or the human population will need continuous reduction till the system reaches equilibrium.
Noumenon
2.8 / 5 (9) Mar 28, 2013
How do you compete with corporations who can send in a robot to do your job better than you and at a fraction of your wage


With respect, you don't know what you're talking about,... and such a state is the reason liberals become liberals. Nothing you just stated makes any sense.

Corporations will not be able to sustain themselves if they don't have employed people that are able to buy their products. So this imagined nightmare of robots taking all the jobs and causing mass unemployment is nonsensical .

Ford understood this way back when he started making cars, which is why he paid people a good wage, ... so his employees could afford to buy his products, thus folding that back into the corporation.

Robots will make some jobs obsolete, only if the economic conditions allow that job to go obsolete. This is how we evolve technologically, as has been the case since the industrial revolution.

You guys watch too many cartoons, thinking robots can do human thought tasks.
Eikka
not rated yet Mar 30, 2013
Ford understood this way back when he started making cars, which is why he paid people a good wage, ... so his employees could afford to buy his products, thus folding that back into the corporation.


Ford understood that the car must be cheap enough that his own poorly paid employees can purchase one, so he made a cheap car.

It wasn't about fair wages, it was about market research and segmenting. Cars were previously vanity items for the filthy rich, and Ford needed to figure out what kind of a car could sell to everybody, and that's the metric he used: whether his own people could afford it.

No company ever prospered by selling the product to its own workers.
LariAnn
1 / 5 (2) Apr 01, 2013
So long as we are imagining the future, perhaps a lethal virus will "show up" coincidentally with the emergence of the independently capable domestic/industrial robot, reducing the human population drastically at the same time as the robot population explodes. Equilibrium will be reached when the number of capable humans is adequate for the tech jobs available. Many more jobs will open up as humanity begins serious exploration and exploitation of the solar system using advanced autonomous robotic manufacturing facilities and explorer teams. At that time, genetically engineered superior human genomes will be available to be used for procreation of intellectually capable human engineers and technicians, at least until the robots become advanced enough to take over their own engineering and development. Humans will move gradually into the role of pure thinkers and observers rather than doers, as the robots become so advanced that they conduct their own research independently.
nowhere
not rated yet Apr 03, 2013
Corporations will not be able to sustain themselves if they don't have employed people that are able to buy their products. So this imagined nightmare of robots taking all the jobs and causing mass unemployment is nonsensical.

Incorrect. The market will simply shift over time until more products and services are focused on machines and their sector. You see the value in the market hasn't disappeared, it has only moved from being in the hands of people to the hands of the corporation that owns the machines doing the work. Industries that are no longer necessary or lack the financial backing will fade and may even dissappear all together.