Accessibility requirements for human-robot interaction for socially assistive robots

Resumo

Accessibility was expansively investigated in Human-Computer Interaction (HCI), to guarantee that all users with different abilities and needs have equal access to software and hardware products. Thus, many accessibility guidelines are available for designers and developers of mobiles interfaces, web content, etc. In contrast, there are no such guidelines for developers and designers of Human-Robot Interaction (HRI) applications. Hence, the main aim of this thesis is to identify the accessibility barriers in HRI and propose accessibility guidelines for HRI. The diversity in robot types is huge. Hence, we limited the scope of studied robots to Social Assistive Robots (SARs); the interaction components of twenty SARs were studied and classified under: software and hardware components. Then, a systematic review was conducted to search classifications for HRI interfaces, the results were integrated to our proposed classification of interaction components in HRI. A search was performed to identify main HCI accessibility laws and guidelines; six of the studied guidelines were selected to form a basis of our proposal; these guidelines cover hardware, software and user experience areas in HCI. Only applicable guidelines to HRI were selected. To understand accessibility barriers in HRI deeply, the proposed classification of interaction components was mapped to users’ potential disabilities, which produced a proposal of interaction components that could affect users’ potential disabilities during the interaction time. Real and fictional users’ cases of interaction with SAR were evaluated to elicit accessibility barriers. The evaluation has been conducted following the accessibility evaluation methodology for HRI, which was proposed in this thesis, and used in following two heuristic evaluations too. A list of accessibility barriers were identified. The proposed accessibility guidelines for HRI were evaluated and used in three different heuristic evaluations and one user evaluation. The first evaluation was with HRI developers and designers, who agreed by majority that the proposed guidelines helped them implement and design accessible interfaces, and they showed a desire to use them in their future designs and implementations, as none of the developers or designers had applied all the proposed guidelines previously. In the other two evaluations, the proposed guidelines were used to evaluate two different applications of a SAR. Thirteen accessibility problems were elicited from the first application, and eight accessibility problems from the second. For the first application, evaluator’s recommendations were implemented and evaluated through a real user evaluation in a residence home. A list of found accessibility barriers were identified. The final version of the proposed guidelines presented in this document was updated based on participants’ recommendations in the first heuristic evaluation.