April 20, 2024

AI-powered robot reads Braille twice as fast as humans

Summary: Researchers developed a robotic sensor that uses artificial intelligence to read braille at a remarkable speed of 315 words per minute with 87% accuracy, surpassing the average human reading speed. This sensor uses machine learning algorithms to interpret braille with high sensitivity, mirroring human reading behavior.

Although not designed as assistive technology, this advance has implications for the development of prosthetics and responsive robotic hands. The sensor’s success in reading braille demonstrates the potential of robotics to replicate complex human touch skills.

Key facts:

  1. The robotic braille reader uses artificial intelligence and a camera-equipped “fingertip” sensor to read at twice the speed of most human readers.
  2. The robot’s high sensitivity makes it an ideal model for developing prosthetics or advanced robotic hands.
  3. This advance challenges the engineering task of replicating the sensitivity of human fingertips in robotics, offering broader applications beyond braille reading.

Fountain: University of Cambridge

Researchers have developed a robotic sensor that incorporates artificial intelligence techniques to read braille at speeds about twice that of most human readers.

The research team, from the University of Cambridge, used machine learning algorithms to teach a robotic sensor to quickly glide over lines of braille text. The robot was able to read braille at 315 words per minute with an accuracy close to 90%.

Although the braille-reading robot was not developed as assistive technology, researchers say the high sensitivity required to read braille makes it an ideal test in the development of robotic hands or prosthetics with sensitivity comparable to that of fingertips. humans.

This shows the robot.
Researchers have developed a robotic sensor that incorporates artificial intelligence techniques to read braille at speeds about twice that of most human readers. Credit: University of Cambridge

The results are published in the magazine. IEEE Robotics and Automation Letters.

Human fingertips are remarkably sensitive and help us gather information about the world around us. Our fingertips can detect small changes in the texture of a material or help us know how much force to use when gripping an object: for example, picking up an egg without breaking it or a bowling ball without dropping it.

Reproducing that level of sensitivity in a robotic hand, in an energy-efficient way, is a major engineering challenge. In Professor Fumiya Iida’s lab in Cambridge’s Department of Engineering, researchers are developing solutions for this and other skills that humans find easy, but robots find difficult.

“The softness of human fingertips is one of the reasons we can grasp things with the right amount of pressure,” said Parth Potdar of Cambridge’s Department of Engineering and a student at Pembroke College, the paper’s first author. . “For robotics, smoothness is a useful characteristic, but you also need a lot of information from the sensors, and it is difficult to have both at the same time, especially when dealing with flexible or deformable surfaces.”

Braille is an ideal test for a robot’s “fingertip”, as its reading requires high sensitivity, as the dots in each representative letter pattern are close together. The researchers used a commercially available sensor to develop a robotic braille reader that more accurately replicates human reading behavior.

“Robotic braille readers exist, but they only read one letter at a time, which is not how humans read,” said co-author David Hardman, also of the Department of Engineering.

“Existing robotic braille readers work statically: they touch a pattern of letters, read it, rise from the surface, move, move down to the next pattern of letters, and so on. “We want something that is more realistic and much more efficient.”

The robotic sensor the researchers used has a camera on the “fingertip” and reads using a combination of information from the camera and sensors. “This is a difficult problem for robotics, as a lot of image processing needs to be done to remove motion blur, which consumes time and energy,” Potdar said.

The team developed machine learning algorithms so that the robotic reader could “wipe” the images before the sensor attempted to recognize the letters. They trained the algorithm on a set of sharp braille images with a false blur applied. After the algorithm learned to blur the letters, they used a computer vision model to detect and classify each character.

Once the algorithms were incorporated, the researchers tested their reader by sliding it rapidly along rows of braille characters. The robotic braille reader could read at 315 words per minute with 87% accuracy, which is twice as fast and about as accurate as a human braille reader.

“Considering we used a fake blur to train the algorithm, it was surprising how accurate it was at reading braille,” Hardman said. “We found a good balance between speed and accuracy, which is also true for human readers.”

“Braille reading speed is an excellent way to measure the dynamic performance of tactile sensing systems, so our findings could be applied beyond braille, for applications such as detecting surface textures or slides in robotic manipulation,” Potdar said.

In the future, the researchers hope to expand the technology to the size of a humanoid hand or skin.

Money: The research was supported in part by Samsung’s Global Research Extension Program.

About this research news in AI and robotics

Author: sara collins
Fountain: University of Cambridge
Contact: Sarah Collins – University of Cambridge
Image: Image is credited to Neuroscience News.

Original research: Closed access.
“High-Speed ​​Tactile Braille Reading Using Biomimetic Sliding Interactions” by Parth Potdar et al. IEEE Robotics and Automation Letters


High-speed tactile Braille reading using biomimetic sliding interactions

Most robotic braille reading sensors employ a discrete letter-by-letter reading strategy, despite the higher potential speeds of a biomimetic swipe approach.

We propose a complete system for continuous braille reading: frames are dynamically collected with a vision-based touch sensor; an autoencoder eliminates motion blurring artifacts; a lightweight YOLO v8 model classifies braille characters; and a data-driven consolidation stage minimizes errors in the planned chain.

We demonstrate a state-of-the-art speed of 315 words per minute with 87.5% accuracy, more than twice the speed of human braille reading.

While demonstrated in braille, this biomimetic sliding approach can be further employed for richer dynamic spatial and temporal detection of surface textures, and we consider the challenges that need to be addressed in its development.

Leave a Reply

Your email address will not be published. Required fields are marked *