The Dream controller (short for Dynamically-Resizing, Ergonomic and Multi-touch) is a new program designed for Microsoft's Surface touch-screen that lets one or more people take control of a single robot, or a swarm of them. The software offers new ways for users to manage search-and-rescue robots through a touch screen interface, while integrating with virtual maps. Several virtual robot controllers are automatically integrated with a view of a virtual map.

When disaster strikes, search-and-rescue teams must quickly gather and assimilate the data needed to find survivors. A team of robots can help scout out for persons stuck in rubble or create new maps of the landscape. But first responders need ways to control those robots, and process incoming information quickly.

Robots are usually controlled using a physical device, like a joystick or games console-type controller. Mark Micire, a researcher at the University of Massachusetts Lowell who deployed search-and-rescue robots at the World Trade Center after 9/11 and a member of the Massachusetts FEMA team, built the Dream system to help first responders. Since they don't need to use new device to control each of several robots while referring to a physical map, he says, it is possible to maneuver more quickly.

"Right now, the state of practice is to use paper maps, with everyone gathered around," says Holly Yanco, professor and head of the Robotics Lab at UMass Lowell. "We've designed the multi-touch application to replace these maps with interactive ones." This lets live data--satellite imagery, sensor data, and video or photography from people, vehicles or robots--to be used nearly instantaneously.

"With our design, a person can select the robot, then place his or her hands down to form the Dream controller to directly drive the robot and see the robot's eye video," says Yanco. "Once done, the controller disappears."

Each controller sizes itself to a person's hand space and finger size, based on contact points made with the screen--so very large or very small hands would be able to control it just as easily. "We aren't aware of any other multi-touch controllers that conform to the placement and size of a person's hand," says Yanco. The Dream controller's hand and finger registration algorithm, which has a patent pending, is faster and outperforms other work, according to Yanco.

Here's a video of the controller being used controlling a real robot (ATRV-Jr):

And this video shows it controlling a swarm of virtual robots:

We finally have a hint of what The Droid Works, a stealthy-mode start-up founded by Helen Grenier, one of the cofounders of iRobot, will develop.

The NSF has given The Droid Works a grant worth nearly $100,000 to develop indoor and outdoor unmanned air vehicles. According to the report, the UAVs will be used for emergency response:

Indoor applications would enable the UAVs to respond to emergency situations that involve large steps, closed doors and rough terrain. The NSF grant will be used to develop indoor flight control and safety technology for the UAVs.

Other researchers are developing UAVs and other robots to assist with a rescue or survey an area following a disaster. Robin Murphy, a professor of computer science at Texas A&M University, is leading one effort to use snake-like robots to rescue people trapped beneath rubble. More recently she developed a way to use multiply UAVs to survey an area and locate survivors.

Imagine a building has collapsed. A team of first responders rushes to the scene and rapidly begins surveying the area for survivors. They draw makeshift maps of the area, so that incoming teams know what's what. But newcomers don't always understand the depictions and each minute is crucial to save survivors.

Robin Murphy from Texas A&M University (TAMU) lab and colleagues have a solution: deploy several small unmanned air vehicles (SUAVs), such as AirRobot quadrotors, to take snapshots of the rubble. The pictures are then uploaded to a software program called RubbleViewer, which quickly builds a three-dimensional map of the area that users can intuitively navigate. More efficient than drawing by hand, this system is also cheaper and more portable than the alternative--using helicopter-mounted lasers to map the rubble.

Pictures from the SUAVs are combined using the algorithms behind the panorama-making software PhotoSynth. RubbleViewer extracts information from PhotoSynth's data points to create the three-dimensional map. It's like putting a blanket over a bunch of needle points, says Maarten van Zomeren, a graduate at the Delft University of Technology in the Netherlands who helped developed the technology under supervisor and assistant professor Stijn Oomes.

While PhotoSynth has been coupled to applications like Live Search Maps and Google Maps to create enhanced, location-embedded panoramic views, RubbleViewer is designed to be fast and easy to build, taking about half an hour to create a topographic 3-D map of an area. What's more, viewers can click on a spot to annotate the map (showing the location of possible survivors, for example) or call up the real photos tied to the spot. See the video below for more.

The program is still a prototype, but expert reviews will come out next month.

Murphy intends to combine RubbleViewer with quadrotors and land-based, search-and-rescue robots to create an easy-to-use first-responders system. Murphy is also working with the Sketch Recognition Lab at TAMU to develop electronic tablets for responders to use. "Because it's an emergency scenario it's really important that people don't have to learn anything but can interact with the world in a way that's natural or intuitive to them," says lab director and assistant professor Tracy Hammond. "We have to enable as opposed to constraining them with technologies."

The team plans to carry out the first tests of the combined system by the end of the summer.