Engenharia de computação e tecnologias assistivas: recursos de acessibilidade ao computador para pessoas com deficiência motora

Abstract

People with disabilities (PwD) make up about 25 % of the Brazilian population. These people, in most of the cases, have physical impediments that make it difficult, or impossible, to perform certain activities. These activities are often related to living with other people and their impediment causes social exclusion. In order to abolish, or at least decrease, the exclusion of these people, many countries have adopted politics of inclusive education. As computers are a strong educational tool, there is a need for adaptations of their peripherals (mouse and keyboard) so that users with PwD can use them normally. Assistive technologies (TAs) are techniques and tools that aim to assist the execution of these activities and promote autonomy for PwD. The present work deals with the development of three tools to aid the use of computers by motor PwD in upper limbs. The tools were developed based on a collection of information related to the use of the computer with 5 PwDs. The first tool was a low cost adapted mouse (MouseAdapt), made from a 3D printer. The second technology developed was an adapted text editor (AdaptText), which allows the writing of words using only the mouse functions. The third tool developed in this dissertation was a Human Machine Interface (HMI) through the acquisition and classification of electromyographic signals (EMG). The HMI uses 2 pairs of surface electrodes, two acquisition boards opensourses (Shield EKG / EMG Olimex) and one Arduino board for signal acquisition. The proposed HMI still uses machine learning techniques (AM) to classify 3 hand movements to control a text editor. This dissertation also brings evaluative results of the developed tools. The first two technologies were evaluated by observation, for this purpose usability tests were carried out between 5 users with motor PcD. The evaluation of the HMI occurred through inspection tests, which measured the effectiveness of the classifier when classifying signals stored in a database (created in this work) and signals collected in real time and applied to the control of a text editor. Finally, this dissertation encourages reflection on the contributions that technological areas can offer to the development of ATs, mainly in the production of resources adapted for computer control.