Monday, November 24, 2008

Affordable computer-controlled Braille board under development

From Drobe.co.uk in the UK:

A low-cost computer-controlled Braille board has been prototyped by a RISC OS-using university student.

Undergraduate Edward Rogers hopes to sell his completed units for as little as 200 quid each to schools and families to allow more blind children to continue learning Braille. And he said he wanted to launch his venture using RISC OS-powered kit before offering a package for other platforms.

Edward's mechanical Braille board, also known as a cell display, uses little motors to push Braille alphabet patterns onto a flexible surface that the blind can then read using their fingertips. A computer is connected to the display and software feeds text from web browsers and other applications to the motors.

According to Edward, displays already on the market use complex mechanics and can cost between £3,000 and £10,000 - a price tag that is too much for most parents to swallow, leaving the blind cut off from the internet and other technologies. As part of his degree with the Bristol School of Animation, final-year Edward produced a RISC OS and Windows driver for the Arduino interface kit, (pictured above), devised and built his motor-driven cell display and created the software to link
it all together.

A second prototype is due to be construction by June next year.

Revealing why he chose this particular project, he said: "My enthusiasm for somewhat Heath Robinson-style solutions to problems outside of the general public eye kick started my imagination for the cell display. It was designed solely by me."

Edward, who has had brief contact with a couple of RISC OS companies to gauge their interest in the kit, explained: "Braille cell displays are an electronic mechanical rendering of text from the screen of a computer onto a refreshable Braille board. The primary method used by most blind people to read off a computer is a screen reader with a voice synthesiser.

"Very few blind adults can read or are capable of learning to read Braille because most go blind during their adulthood. Braille reading is largely the preserve of those born blind or blinded during childhood. Therefore most children will learn. This is done with embossed documents rather than computerised ones.

"However when the child finishes the initial learning there is seldom the necessary facilities at affordable prices to allow their skills to develop and so instead it withers. "

Some schools will have a cell display for the student to occasionally practice on but there will usually not be any available for general classroom use. The ideal solution is the portable computers and cell displays known as notetakers, which have essentially the same functionality as a PDA - but these retail for between £3,000 and £10,000.

"The aim of the £200 braille cell display project is, surprisingly, to produce a braille cell display hardware and software package that will retail at or under £200."

Mechanical Braille boards are not a new technology and a lot of time and research has been spent on the subject. Boffins at IBM and the Texas Institute and Stanford Research Institute have explored using electrical shocks, thermal sensing, fluids and other materials that stiffen in the presence of electrical current and similar magic - all of which add up to a rather expensive bill for end-users.

Edward added: "These huge price tags are down to the mechanical complexity of current models. The core element of these cell displays, the piezo-electric pins, are what keep the prices so high as there must be at least 240 of them, costing over £4 each. While there has been various attempts in the past to upgrade the hardware so it can be produced more cheaply, thus far all attempts have failed to make it past either the patent or prototype stage.

"However, where they failed, my project expects to succeed. The simple reason being that the new technological solutions were more complex than the problem. By contrast, this project operates primarily off four electric motors.

"Cell displays require information to be fed to them via a screen reader, which takes the text off the desktop and open windows. There are several large commercial screen readers available for Microsoft Windows, most notably Jaws, but they are all extremely expensive.

"Therefore, while it is essential the reader be made compatible with Jaws, it will be released with a free or very cheap screen reader package. Currently the only screen reader is a simple text file reader and navigator programmed by myself on BBC Basic for Windows, purely for the purposes of testing."

In Edward's system, the blind person's computer talks to the Arduino electronics and sends it the text that needs to be displayed. The embedded microprocessor on the Arduino, programmed by Edward, turns the signals into motor movements. Edward has also started work on basic scanning and OCR support for his system.

He said: "I aim to have a package which runs off RISC OS, possibly through the ROOL ROM and emulator option if it becomes available or on second-hand RiscPCs, and can access the Internet, browse directories and read various types of text files. Running on RISC OS initially will allow it time to develop and iron out flaws before jumping into the Windows ring with the industry giants - as well, that is, as promoting the platform. Early versions of the software will be aimed at schools willing to test the new system rather than personal usage."

The kit has been demonstrated at his university although a field test was halted when the material onto which the Braille is pushed was found to be unsuitable. It's hoped that his cell display will eventually feel as similar to paper as possible.

Edward said: "The initial prototype developed a fundamental flaw which, though it has now been rectified, meant all testing and feedback had to be cancelled, quite literally within the final hour.

"The machine, in its semi-functional form was demonstrated to my university colleagues. However it was due to be tested by a parents association, a teachers group and all the regional branches of the RNIB for the West Country. Where as the university was interested in how far the project had progressed, these other groups were going to test the actual readability of the Braille produced.

"The material used to emboss onto turned out not to be adequate - so while the motors turned out perfectly readable braille, there was nothing from which a blind person could read it. Therefore the tests were cancelled."

Edward also had to learn C++ in order to develop the program code for the Arduino board, which he had to learn as the project progressed.

He said: "I'm not a natural programmer so learning even basic C++ was very difficult. On the mechanical side, the greatest challenge is keeping the machine as simple as possible while building in full functionality."

Asked if he felt his work was pioneering, Edward said: "I wouldn't feel comfortable assuming the work is pioneering. Many others have tried and failed with other more exciting and revolutionary concepts to reach a solution. Elements of these previous projects have been inspirational for me, but will be applied to far simpler technologies in this case. I am wary, when the machine is so far from completion, of putting too much expectation into those four motors."

Edward said he wanted to concentrate on completing his last year at university before completing the Braille board project.