Get Herald RSSA future full of obedient robots following our every command may not be such an unrealistic possibility after all - even if those commands go unspoken.
A Brown robotics team has recently developed a robot that can follow gestural commands in a variety of environments without having to adjust for changes in lighting, a breakthrough in the robotic world.
While the majority of robots are programmed to recognize specific colors and are constrained to specific lighting conditions, this new robot uses an active light-based system that easily adapts to lighting variance. This ability allows it to operate indoors and outdoors without the need for re-calibration in differently lit environments.
Matthew Loper GS, the lead author of the paper about the project, which was presented at the Human-Robot Interaction Conference in San Diego, Calif., last weekend, said the ability to adjust to light allows the robot to function outside a laboratory setting. "There have been lots of works that have done the kinds of things we've done - person following, gesture recognition, speech recognition," Loper said. "But the important thing is in making a system that has environmental tolerance."
Despite its prevalence in the robotics world, the other important aspect of the robot's design is its ability to respond to nonverbal, gestural commands.
"In the shorter term, we're trying to take a step away from remote-controlled, teleoperation," he said. "We would rather have them interact with people more naturally, able to understand nonverbal gestures, understand speech and follow a person around."
The robot can be programmed to track multiple people and can discriminate between whom to respond to and whom to ignore.
But the design is not flawless. While the tracking distinguishes between two people, the robot can sometimes be tricked because it relies on silhouettes.
"We want to foster natural human-robot collaboration in the long term and the kind of interactions that you can get between people," said Loper, who was responsible for creating the gesture-recognition component. "That a person could interact with a robot in the same way that a person can interact with a person."
