SCIS Students Present at Research Symposium
On April 27th, three students from the School of Computing and Information Science presented their cutting edge research at the 2016 Graduate and Undergraduate Research Symposium at the Cross Insurance Arena in Bangor, Maine. These students include HariPrasath Palani, Stacy Doore, and Christopher Bennett. The abstracts from their respective research projects may be found below.
HariPrasath Palani, “Multimodal Access to Graphical Information for Blind and Visually-Impaired people using Touchscreen-based devices.”
The proliferation of touchscreen devices has amplified the use of graphical information, most of which is visual in nature. The lack of inexpensive, nonvisual graphical access solutions is concerning for over 12 million blind and visually-impaired (BVI) people in the U.S. (and 285 million worldwide). My research addresses this issue through evaluation, design, and development of a touchscreen-based multimodal interface called a Vibro-Audio Interface (VAI) – that allows BVI users to apprehend and learn graphical information using commercially-available technology. The VAI’s efficacy has been previously demonstrated for conveying graphical information (e.g., graphs and maps) and my work aims to better characterize the underlying perceptual / cognitive factors involved in learning with this new class of information access technology. Two preliminary psychophysically-inspired usability studies addressed the perceptual parameters for detecting and tracing graphical stimuli rendered with the VAI, with results demonstrating that a minimum of 1mm width is necessary for detecting lines using haptic feedback and a width of at least 3mm is needed for effective line tracing. My dissertation research addressing other perceptual parameters needed for nonvisual graphical access will lead to new conceptual models of multimodal information access and development of human-centered design specifications for future multimodal touchscreen-based accessibility solutions.
Stacy Doore, “A Room with a View: Designing Natural Language Interface Structures for Indoor Scene Description.”
We are increasingly reliant upon intelligent agents with natural language (NL) communication abilities to interact with the world around us. While mature technology exists to accurately describe and navigate outdoor settings, the creation of automated NL systems that can effectively communicate spatial scenes within indoor environments presents significant challenges. Descriptions of objects and structures within built environments differ from those of outdoor space due to the lack of metric data and the absence of standard indoor landmarks. This presentation will describe a series of experiments related to spatial preposition use patterns, preferences, and rules within different vista scale indoor environments along with early results. This research contributes to the development of technologies to support mobile device cameras to be used as visual sensors to obtain images of indoor spaces and convert the imagery to verbal descriptions of the indoor space. The formal rules and spatial reasoning structures resulting from this study will allow for the creation of a spatial preposition thesaurus for intelligent systems used within indoor environments. Applications of this research includes indoor navigation support for emergency responders, assistive technologies for elderly and blind and low vision users, and automated customer assistance for location-based retail services.
Christopher Bennett, “Do Cognitive Maps Decay with Age?”
A critical component of effective navigation is the ability to form and maintain an accurate cognitive map, which includes paths, landmarks, and the allocentric global relations between them. Proper cognitive map maintenance can become difficult for older adults as many of the constituent memory structures experience degradation with age. While driving, we rely on our cognitive maps to update our location, the relations between surrounding landmarks, and to plan routes. The present study had younger and older adult participants form cognitive maps using a virtual reality driving simulator and then tested these representations one day and one week later to measure cognitive map decay. Data collection for the current study is ongoing. We believe that the cognitive maps of older adults experience greater decay over time than their younger peers and that certain attributes will be affected more than others (e.g., landmarks will be preserved, as shown by current data but with expected loss of topological fidelity). Understanding how the cognitive map changes over time will allow for development of compensatory technologies to mitigate related safety and orientation problems.
###