“We’re trying to give robots superhuman perception,” said MIT Associate Professor Fadel Adib. Indeed, the world of robotics keeps growing, and a team at MIT isn’t standing by idly.
Attempting to push the technology one step further, Adib and his team have developed a robot that is able to use its senses to find hidden objects. RF-Grasp, as the robot is called, manages to do so by using radio waves, which are able to go through walls.
Combining this radio frequency with traditional computer vision assists RF-Grasp to find and pick up objects that are hidden from plain sight.
The team says its invention could one day revolutionize the e-commerce world, where robots could help in warehouses to find and pick up packages and items that aren’t easily noticeable.
The team plans on sharing its creation in May at the IEEE International Conference on Robotics and Automation.
How the robot uses radio waves to find objects
Radio frequency (RF) identification has been used for many years and for many uses. From locating a lost dog who has an implanted chip to finding that book in the depths of a library, the system uses a combination of a reader and a tag to work.
“RF is such a different sensing modality than vision,” said Alberto Rodriguez, the Class of 1957 Associate Professor in the Department of Mechanical Engineering at MIT. “It would be a mistake not to explore what RF can do.”
Which is precisely what the team decided to look into.
In RF-Grasp’s case, it uses both a camera and an RF reader to locate and then grab an object. It uses a robotic arm and an attached grasping hand, and a camera sits atop the robot’s “wrist.” The RF reader is separate and relays tracking information back to the robot’s algorithm.
Thanks to these two complementary systems, RF-Grasp is able to locate and grasp hidden objects on its own.
The team is confident its new robot would perform well in the aforementioned e-commerce situations, as well as assisting people in their homes to find and collect hidden objects.