Taking inspiration from honeybees, an international team of researchers have developed a gesture identification that will allow robots to communicate with each other by “dancing” with a coordinated set of gestures to convey information. The research has been published in the journal Frontiers.
In nature, once a honey bee finds a patch of flowers where it can extract nectar, it goes back to the hive to alert other bees. Once there, it performs a “waggle dance” to tell other bees about the location of the patch of flowers. Other bees interpret the dance’s movements and duration to understand where the patch of flowers is, in relation to the beehive. The researchers used a similar technique to ‘teach’ two robots to communicate with a dance.
“A visual communication system is developed for robots with onboard cameras, using algorithms that allow the robots to interpret what they see. The humans and robots communicate using gestures, such as a raised hand with a closed fist,” said Abhra Roy Chowdhury, senior author of the paper, to indianexpress.com, in an email interaction. Chowdhury is the head of the Robotics Innovation Lab at the Centre for Product Design and Manufacturing at the Indian Institue of Science In Bengaluru.
Chowdhury and his co-author Kaustubh Joshi, a doctorate student and research assistant at the University of Maryland, developed a proof-of-concept for this communication system using two robots that served as proxies for package handling robots.
First, a human operator used hand gestures, such as a raised or clenched fist, to convey a coded message which contained the location of a ‘package’. A robot detected these gestures and decoded the message to understand the package location based on a map of the environment that was encoded into it. It was then able to convey the same information to a second robot through a dance. The robots were able to successfully interpret and relay the information 93.33 per cent of the time during experiments.
Robots typically communicate with each other using different digital networks, including wireless communication. But according to Chowdhury, there could be situations where network communications aren’t available but robot labour is required, like in disaster zones or during spacewalks. Since the robots in the study used simple cameras to identify the gestures, the technology has the potential for scalability.
Chowdhury told indianexpress.com that he and Joshi are now working on making the technology more accurate and more robust so that it can be used to convey more complex messages, tasks and instructions.