New Robot Technology Allows Robots to Move Freely in Changing Environments

Up till now, one of the biggest hurdles standing in the way of having robots in the home is the constantly changing environment. A book dropped on a table or someone pushing back a chair to stretch their legs makes it impossible for robots to move around in such an environment without either getting lost or avoiding running into, and perhaps wrecking things. Now, researchers at MIT appear to have a found a way around this problem. In a paper they’ve published on the MITNews site, the team describes how they’ve built a system that allows a robot to create a virtual map of the environment in which they are “living” that can be updated on the fly, freeing them to wander about without fear of getting lost or creating other problems due to encountering unfamiliar surroundings. The researchers say the system could also be used by blind people.

In traditional robotic systems, programmers create a map of the environment in which a robot is to exist, the robot then uses that map to find its way around. The problem comes of course when something in the environment changes. Up till now, only multi-million dollar robots being developed by companies with grant money from the government have been built that can adjust to virtually any terrain. This is why a new, inexpensive way to accomplish the same thing for a fraction of the cost is so exciting.

The new technology is based on something called Simultaneous Localization and Mapping (SLAM), which is based on artificial intelligence technology. In such a system, new elements are added to a tree of knowledge as they are encountered in the real world. That way, the robot is able to make sense of the environment each time it moves around.

What’s perhaps just as exciting is that rather than use an expensive laser system to allow the robot to notice changes in the environment, the engineers used a Microsoft Kinect instead. Because of that the new technology behind it could actually wind up in real robots sold to consumers within the next year or so. The Kinect actually allows the robot to build a continuously changing 3D map of its own little view of the world.

In practice, the robot simply uses the Kinect to view a room it has entered, then compares what it’s found with what it saw the last time it was there. Any changes are duly noted and added to the 3D map. Then, the robot can proceed with whatever it’s been asked to do.


People also view

Leave a Reply

Your email address will not be published. Required fields are marked *