Robotic finger with very high precision sense of touch, Researchers have introduced a new type of robot finger with a sense of touch. Their finger can find touches with very high accuracy <1 mm on large surfaces and many curves, similar to human parts.

“There has long been a gap between a standalone touch sensor and a fully integrated finger touch. The feeling of touch is far from universal in robotic manipulation,” the researcher said.

In this document we have shown a multi-curved robotic finger with precise contact localization and detection of normal force on complex 3D surfaces.

Current methods of building touch sensors have proven difficult to integrate into robotic fingers due to many challenges, including difficulty covering multi-curved surfaces, large numbers of wires, or difficult placement on small fingers, resulting in the use in skilled hands being prevented. Robotic finger with very high precision sense of touch.

The Columbia engineering team has chosen a new approach: the use of overlapping new signals from light transmitters and receivers, which are embedded in a transparent wave layer that covers the functional area of ​​the finger.

By measuring the transport of light between the transmitter and receiver, they have shown that they can receive a very broad set of signal data that changes in response to finger deformation due to touch.

They then demonstrate that purely data-based deep learning methods can extract useful information from data, including contact positions and normal forces, without the need for analytical models.

The end result is a fully integrated, wireless sensory robotic finger that is manufactured using inexpensive manufacturing methods and designed for easy hand-to-hand integration.

The finger also has more than 30 photodiodes that measure how light is reflected. Every time a finger touches something, its skin changes shape so that light enters the transparent layer underneath. Robotic finger with very high precision sense of touch.

By measuring how much light reaches each diode from each LED, the researchers reach nearly 1,000 signals, each containing information about the contact made.

Because light can also be reflected in curved spaces, these signals can mask complex 3D shapes such as fingers.

“The human finger provides unusually wide contact information – more than 400 small sensors per square centimeter of skin,” the researchers said. It is important to ensure that all contacts on all sides of the finger are closed. We basically build tactile robot fingers without blind spots. Robotic finger with very high precision sense of touch.

Second, the team designed this data to be processed by a machine learning algorithm. Because there are so many signals, some of which overlap, the data is too complex for people to interpret.

Fortunately, modern machine learning techniques can learn to extract information that interests researchers: where fingers touch, what fingers touch, how much force is applied, and so on.

The team also built a finger so that the finger and other people could be placed on the robotic arm. Integrating the system into your hands is easy: thanks to this new technology, fingers collect nearly 1000 signals, but only require a 14-core cable to connect it to the hand, and there is no complicated sticky electronics. Robotic finger with very high precision sense of touch.

Researchers already have two skilled hands (able to capture and manipulate objects) in their laboratory to be equipped with these fingers. One hand has three fingers and the other four.

In the coming months, the team will use this hand to try to show skilled manipulation based on touch and proprioceptive data.

Eligible robot manipulation is now needed in various fields such as manufacturing and logistics and is one of the technologies that will be needed in the long run to enable personal robot support in other fields such as healthcare or services.