Multimodal Interaction and Performance Validation of a Raspberry Pi-Based Grade 1 Braille Notetaker

Abstract

Electronic Braille notetakers can support independent reading and writing, but their interaction logic must coordinate chorded input, refreshable output, file operations, cursor feedback, and audio prompts within a compact device. This paper presents the design and functional validation of a Raspberry Pi 4-based Grade 1 Braille notetaker developed as an extension to the UTM eBraille platform. Fifteen mechanical switches provide six-dot Perkins input and command navigation, while a 24-cell Braille module, speaker, custom main circuit board, and Braille-cell interface board form the output and interconnection layers. The software is organized into file manager, reading, editing, trash, settings, and search states. Bidirectional conversion routines translate keyboard chords into text and map text characters, capitalization indicators, numeric indicators, and cursor positions to Braille-cell data. Functional tests verified basic characters, capitalized letters and words, numbers, cursor movement, text insertion, save reminders, and file-property presentation. A repeated entry test used the ten-character sequence “abcdefghij” over 20 trials. Mean entry time was 0.3233 s per character, equivalent to 3.1134 characters/s, and mean character accuracy was 97.5%. The results demonstrate that a state-based multimodal interface can combine essential note editing and file management on an embedded Braille platform. Current limitations include text-only files, uncontracted English Braille, an external speaker, and the absence of participant-based usability evaluation.