LBtrans-Bot: A Latin-to-Balinese Script Transliteration Robotic System based on Noto Sans Balinese Font

Abstract

Balinese script writing, as one of Balinese cultural richness, is going to extinct because of its decreasing use. This research is one of the ways to preserve Balinese script writing using technological approach. Through collaboration between Computer Science and Balinese Language discipline, this research focused on the development of a Latin-to-Balinese script transliteration robotic system that was called LBtrans-Bot. LBtrans-Bot can be used as a learning system to give the transliteration knowledge as one aspect of Balinese script writing. In this research area, LBtrans-Bot was known as the first system that utilize Noto Sans Balinese font and was developed based on the identified seventeen kinds of special word. LBtrans-Bot consists of the transliterator web application, the transceiver console application, and the robotic arm with its GUI controller application. The transliterator used the Model-View-Controller architectural pattern, where each of them was implemented by using MySQL database (as the repository for the words belong to the seventeen kinds of special word), HTML, PHP, CSS, and Bootstrap (mostly for the User Interface responsive design), and JavaScript (mostly for the transliteration algorithm and as the controller between the Model and the View). <em>Dictionary</em> data structure was used in the transliterator memory as a place to hold data (words) from the Model. The transceiver used batch script and AutoIt script to receive and trasmit data from the transliterator to the GUI controller, which control the Balinese script writing of the robotic arm. The robotic arm with its GUI controller used open-source mDrawBot Arduino Robot Building platform. Through the experiment, LBtrans-Bot has been able to write the 34-pixel font size of the Noto Sans Balinese font from HTML 5 canvas that has been setup with additional 10-pixel length of the width and the height of the Balinese script writing area. Its transliterator gave the accuracy result up to 91% (138 of 151) testing cases of The Balinese Alphabet writing rules and examples document by Sudewa. This transliterator result outperformed the best result of the known existing transliterator based on Bali Simbar font, i.e. Transliterasi Aksara Bali, that only has accuracy up to 68% (103 of 151) cases of the same testing document. In the future work, LBtrans-Bot could be improved by: 1) Accommodating more complex Balinese script with trade off to the limited writing area of robotic system; 2) Enhancing its transliterator to accommodating the rules and/or examples from the testing document that recently cannot be handled or gave incorrect transliteration result; enriching the database consists of words belong to the seventeen kinds of special word; and implementing semantic relation transliteration.