In a groundbreaking development, a team of engineers at the University of California San Diego (UCSD) has designed a robotic hand that can rotate objects simply by touch, without the need for visual input.
This innovative approach is inspired by the way people can effortlessly manipulate objects without having to see them.
A Touch-Sensitive Approach to Object Manipulation
The team equipped a four-fingered robotic hand with 16 touch sensors spread across the palm and fingers. Each sensor, which costs about $12, performs a simple function: it detects whether an object is touching it.
This approach is unique because it relies on a large number of low-cost, low-resolution touch sensors that use simple binary signals (touch or non-touch) to perform robotic hand rotation.
In contrast, other methods rely on several high-cost, high-resolution touch sensors affixed to a small area of the robotic hand, specifically the fingertips.
Xiaolong Wang, a professor of electrical and computer engineering at UC San Diego who led the study, explained that these approaches have several limitations.
They limit the system’s detection ability by minimizing the possibility of sensors coming into contact with the object.
High-resolution touch sensors that provide information about texture are extremely difficult to simulate and prohibitively expensive, making them difficult to use in real-world experiments.
The Power of Binary Signals
“We show that we don’t need details about the texture of an object to do this task. We just need simple binary signals of whether the sensors are touching the object, and these are much easier to simulate and transfer to the real world,” he said.
The team trained their system using simulations of a virtual robotic hand rotating various objects, including those with irregular shapes.
The system evaluates which sensors in the hand the object touches at any point in time during rotation.
It also evaluates previous movements as well as current positions of the hand joints. Using this information, the system tells the robot hand which joint should go where at the next time point.
The Future of Robotic Manipulation
The researchers tested their system on a real-life robotic hand with objects the system had not yet encountered.
The robotic hand was able to rotate various objects without stopping or losing its grip. The objects included a tomato, a pepper, a can of peanut butter, and a toy rubber ducky, which was the most difficult object due to its shape.
Objects with more complex shapes took longer to rotate. The robotic hand can also rotate objects around different axes.
The team is now working to extend their approach to more complex manipulation tasks.
They are currently developing techniques that will enable robotic hands to catch, throw and juggle, for example. “In-hand manipulation is a very common skill that we humans have, but it is very complex for robots to master,” Wang said. “If we can give robots this ability, it will open the door to the types of tasks they can perform.”
This development marks a significant step forward in the field of robotics, potentially paving the way for robots that can manipulate objects in the dark or visually challenging environments.
