TOPIC 2.
Development of Assistive Technology to Improve
the Functional Ability of Students and Adults with Disabilities

• NutriNet: An Internet Based Self-Directed Multimedia Nutritional Planning and Grocery Shopping System for Individuals with Mental Retardation
• Speech-to-Print-to-Sign System

SPONSORING OFFICE:
Office of Special Education and Rehabilitative Services

Abstract:
For most Americans, understanding and planning daily diets can be a confusing and multi-faceted. The cognitive challenges faced by people with mental retardation create even greater barriers in understanding the complexity of nutritional concepts such as the food groups, caloric values, saturated fats, cholesterol, daily recommended allowances, etc. This has resulted in a high level of dependency on others and all too often poor nutritional health and excess weight in this population. The proposed research will evaluate the utility of NutriNet, an Internet-based multimedia system for self-directed menu planning, grocery list development, nutritional education, and meal preparation. The proposed system will use live video, audio prompts and still pictures or animated graphics to:

1. guide the special needs user through the process of independently preparing a nutritionally healthy menu;
2. generate a shopping list based upon that menu; and,
3. provide an interactive, non-text-based format for understanding various nutritional concepts.

Specific objectives for Phase I include development of a prototype for the Menu Planning and Grocery List Modules. These Modules will be tested against traditional staff-assisted methods of menu planning and comprehension of text-based grocery lists by persons with mental retardation in a pilot study. The hypothesis is that the NutriNet prototype will provide an enhanced format for independent generation of daily menus that are at least as nutritious as menus generated with staff assistance in adult service agency settings, and that a picture based grocery list will be more independently recognizable and useable by the target population. AbleLink Technologies will build upon several previously conducted research projects to develop the NutriNet prototype.

Summary of Anticipated Results and Implications:
The results of Phase I will provide quantitative data demonstrating the technical merit and feasibility of using a multimedia approach accessed via the Internet to independently build nutritionally balanced menus and increased comprehension of corresponding grocery shopping lists. Phase II will focus on further software development of additional educational concepts, special menu planning, multimedia meal preparation modules and ongoing field testing. Substantial commercialization opportunities exist within special education, integrated mainstream home economic classrooms, transition programs, adult service agencies and families of cognitive challenged individuals.

Abstract:
Advances in speech recognition technology show promise for allowing deaf people to receive text generated from speech. Unfortunately, many deaf people do not have a good command of English and find it difficult to follow it in written, as well as spoken form. In the U.S., many of these individuals use ASL, which has a different grammar and syntax than English and no written representation. There is no mechanical or electronic method, at this time, to generate signs from speech.

This project will develop a system for translating speech into signs by using IDRT-generated research in:

1. the application of speech recognition technology to the needs of deaf people; and,
2. the development of a sign language dictionary that allows an individual to identify words through ASL without knowing the English equivalent, using sign parameters.

For this project, sign parameters will be used to control 3D animation. By using animation (as opposed to video), real-time, fluid conversation can be generated in a signed English format. The resulting software will create a system for producing:
Speech —> English Text —> Sign parameters —> 3D animation of signs

Summary of Anticipated Results and Implications:
Phase I will demonstrate the feasibility of this model using 20 signs. Phase II will result in capability to use the system with any sign. This system will enable an individual to speak into a computer and instantly generate real-time seamless signing in 3 dimensions.

[Topic 1] [Topic 3]