Researchers at the University of Cambridge in the UK have developed a robot sensor that, combined with artificial intelligence technology, can read Braille at approximately twice the speed of most human readers. This robot was not originally developed as an aid for the visually impaired, but its high sensitivity makes it an ideal model for developing robotic hands or prosthetics. Its sensitivity even matches that of human fingertips. The research results have been published in the latest issue of “IEEE Robotics and Automation Letters.”

Human fingertips are highly sensitive and can help people gather relevant information about the world around them. Fingertips can detect subtle changes in material texture or help people understand how much force to use when grasping objects. For example, picking up an egg without breaking it or picking up a bowling ball without dropping it. These simple actions are quite challenging for robots.

Braille is an ideal test subject for the robot’s “fingertips” because reading Braille requires high sensitivity, with each representative letter pattern consisting of very closely spaced dots. Researchers have used existing sensors to develop a robot Braille reader that can more accurately replicate human reading behavior.

Existing robot Braille readers can only read one letter at a time, which is not how humans read. The newly developed robot sensor has a camera on the “fingertip,” and the information from the camera and the sensor can be used for precise and rapid reading.

The team has developed machine learning algorithms so that the robot reader can “de-blur” the image before the sensor attempts to identify the letters. They trained the algorithm on a set of clear Braille images distorted by artificial blur. Once the algorithm learned to de-blur the letters, they used a computer vision model to detect and classify each character.

Once the algorithm was integrated, the researchers tested the robot Braille reader by quickly sliding it across a row of Braille characters. The robot Braille reader can read at a speed of 315 words per minute with an accuracy of 87%, which is twice as fast as human Braille readers and with comparable accuracy.

Editor’s Note

Braille is a special reading and writing system that is entirely based on touch. Beginners in Braille need to carefully touch and identify the characters using their fingertips, without skipping or reading out of order. It tests memory, tactile sensitivity, and even patience. However, the robot described in this article not only has “fingertips” but also “eyes.” By combining the information from the camera and sensors, it can achieve precise and rapid reading of Braille. With the application of algorithms, machines can also read blurred Braille. This application can be used to develop better Braille readers. Certainly, a sensitive robot that excels in small tactile dots can do much more.

(Source: Science and Technology Daily)