Design of artificial intelligence-based electronic Malay language learning tool for visually impaired children

• Main Author: Yeoh, Sing Hsia
• Format: Thesis
• Language: English
• Published: 2011
• Online Access: http://psasir.upm.edu.my/id/eprint/41689/1/FK%202011%20138R.pdf; http://psasir.upm.edu.my/id/eprint/41689/

For many years, the application of assistive technologies for the disabled has been given little attention, despite the undoubted need for more. Disable people especially for those who are blind, always face a lot of difficulties in their learning process. Personal teachers have to guide them patiently with the aid of limited learning devices. The advancement of technology in twenty-first century should provide more design of great learning devices. However in developing countries like Malaysia, there are limited locally made assistive devices to suit the language used and the local culture. There are more than 20,000 people who are categorized under vision disability in Malaysia. The percentage of visually impaired people who master Malay language, as the national language in Malaysia, is low. The main purpose for this research is to develop a Malay language learning tool for blind children. This research work involves the implementation of Hamming Distance Technique (HDT) and simple Genetic Algorithm (GA) in spell checking and word suggestion mechanism. Besides spell checking, this system has a complete, step by step learning method with audio output. The learning contents are built using MATLAB. Moreover, it is linked with a tactile feedback module that is built using C language and microcontroller, to provide Braille display functionality. Also, this research involves developing a database for 10,000 Malay root words. This number of words is more than enough for kindergarten level. The simulation results indicate that the algorithm is able to suggest a word, based on the design settings. It depends on the size of word. The longest word, which is 6 ALP, has the slowest word suggestion time, at around 10 seconds for the worst case scenario. The feedback from two surveys is positive with 100% satisfaction on the overall performance of the prototype.