Date of Award
Spring 4-22-2026
Document Type
Thesis - Restricted
Degree Name
Master of Science (MS)
Department
Computer Science
First Advisor
Sheikh Ahmed
Second Advisor
Rochelle Mendonca
Third Advisor
Sabirat Rubya
Fourth Advisor
Suzanne Burns
Abstract
This thesis develops and evaluates myAccessibleHome for accessible home environments. This application is a privacy-first, self-guided mobile software tool that does home safety and accessibility screening and provides necessary recommendations and guidance. This work can be positioned as a step towards AI-based home accessibility. This work is motivated by the need to provide support to independent living and to expand access to rehabilitation and assistive technology. myAccessibleHome guides users through a very user-friendly interface to complete a screening that gets various kinds of information from the user to get a clear idea about the scenario of the users regarding accessibility at home. Then, based on that user’s responses, it prepares and summarizes the result of the screening. With that, it also provides personalized recommendations, helpful resource guides, and a shareable report. If the users want, they can share both the screening data and the generated report, but if not shared, no data ever leaves the user’s device. There is no server for the application, so all the operations are done on the user’s mobile device, and all the data always stays there unless explicitly shared by the user. A feasibility study of the application with IRB approval collected a baseline profile. It also collected a two-week usability snapshot and then a six-week confidence follow-up. All the data collection was self-reported. For both usability and confidence, most of the parameters are highly rated by the users, but some also got low ratings. Those low-rated areas are the future improvement scopes of the application. The thesis makes a total of four contributions. The first one is that it develops a privacy-first, on-device, and self-guided workflow, which performs home safety and accessibility screening and provides necessary action guidance. Secondly, it provides a complete screening-to-action pipeline, which is also organized and consistent. Then it evaluates the system as a research contribution using two data snapshots. One is a two-week usability snapshot, and another one is a six-week confidence snapshot. Lastly, it shows why the system is a step toward an AI-based solution. It already has structured and standardized inputs, deterministic rule-based scoring, and prioritization. It also has structured outputs and evaluation signals that will be able to support any later AI work.