Robots for Domestic help

Personal robots are becoming more popular as people want to do more with their lives, says Director of the Robotics and Intelligent Machines Center in the Georgia Tech College of Computing. Already the global market for personal robots is growing 400% a year.

Robots are not novel technology, but a new generation of intelligent ‘personal robots’ – ones that work with and directly for humans, especially in the home and workplace – have begun to emerge only recently.

To have a personal robot that does things you need, you have to have onboard processing, perception, motion and power, says a professor at the college of Computing. Until two or three years ago, you couldn’t put all of that on one single platform. But now, we have all the technologies that can support a consumer robot that will not be too expensive.

While some personal robots are already available, important research is under way to address the remaining technical and societal challenges. The ultimate home assistant” in our lifetime, a robot will only be possible through interdisciplinary co-operation.

If you just have computer scientists designing robots, you’re not going to build a robot that’s as good as one that could be built by computer scientists and mechanical engineers working together.

Before robots become part of daily life, improvements are needed in robotic software, robustness, and human machine interaction.

Robots need to be able to interpret their world if they go in a new environment they need to be able to recognize, for example, a chair even though it’s a different chair from one they’ve seen before. From a reliability stand point, personal robots must be robust, able to function 24 hours a day, 7 days a week in a variety of environments.

Designing a robot to survive in this environment is difficult because it needs protection of the electronics and sensors, so material selection is important – and we’ll need the right types of motors and drive systems.

Materials used to build robots must not only protect components, but also protect the humans that interact with them. That requires the development of flexible materials. For example, robotic arm needs to be as flexible as the human arm which won’t break easily, yet as strong to lift and push. Another technical challenge is the integration of various products into one robotic system. Microsoft is attempting to address this problem with its new Robotics Studio operating system.

A standard operating system will accelerate robotics development like IBM’s PC did in the early 1980s. Microsoft is now helping define a standard that’s not been there. Combined with the hardware that’s available, this will be the last domino to fall.

If component integration is the final piece of the puzzle, issues of robot power and human interaction must be addressed first. For humans to effectively communicate with personal, robots, the machines need to be able to understand spoken language and gestures.

It should be a technology that can be given to a 70-year-old and not have to worry about helping her with it. To build robots that people will accept, like, and communicate with, researchers draw upon studies in psychology and human computer interaction.

There is a general hypothesis that robots similar in appearance to what is familiar to us will ease acceptance. If you see a robot that looks and acts like a puppy, you’re going to treat it as somewhat of a subordinate, but gently. On the other end of the spectrum, there are human-looking robots.

It is expected that it can do things like a person, and you’re less patient with it. That raises the question of whether robots should look like humans. The more a robot looks like a person – it’s called the uncanny valley theory – the creepier it seems to humans. If you could design a robot that looks like a person, people might accept it, but if it’s off target, it’s creepy.