Anuj Kapoor | Searcharxiv

arXiv2026

Insight: Enhancing Mobile Accessibility for Blind and Visually Impaired Users with LLMs

This research paper addresses the limitations of current mobile accessibility services like TalkBack, which provide manual gesture-based sequential feedback to BVI users. Motivated by the promise of large language models (LLMs), this paper introduces Insight, an Android accessibility service that provides natural language interaction and real-time summarization of the screen. The paper performs a within-subject experimental study with users to compare Insight and TalkBack on usability factors. Results show Insight reduced mental effort and task time, and was preferred because of its dialogue interface, but users felt the need for interruption management. Results show LLM-based interfaces can significantly improve mobile accessibility, and describe the potential of hybrid solutions combining gesture and dialogue modalities towards more inclusive design.

cs.HC↗

arXiv2020

Sensor-Based Spreader Automation for Reducing Salt Use and Improving Safety

Over 30 million tons of deicing salt is applied on U.S. roads annually at a cost of roughly $1.2 billion and with significant negative environmental impact. Therefore, it is desirable to reduce salt use while maintaining winter road safety. Automatic adjustment of application rate in response to road, weather, traffic, and other conditions has the potential to achieve this goal. In the US, salt application rates are typically pre-set manually based on roadway classification, cycle time, desired level of service (LOS), and expected traffic, road, and weather conditions. The operators can temporarily change the application rate manually based on their experience and observations. Current spreader automation mostly involves adjusting discharge rate in response to spreader speed, although pavement temperature sensors are likely to be adopted in the future. This paper explores extending spreader automation for adjusting salt discharge rate on curves, inclines, and other areas with historically high accident concentrations. First, we propose the use of gyroscopic sensors and inclinometers on board the spreader to adjust discharge rate in real time in response to curves and inclines. The second part of the proposed solution includes the installation of roadside Radio-Frequency Identification (RFID) tags that can communicate the length and the severity of the hazard zone to an RFID reader on board the spreader to automatically adjust discharge rate accordingly. Our proposed methods have the potential to reduce salt use and operator fatigue, while increasing winter road safety without adding excessive cost or complexity to the existing spreader systems in the US.

eess.SY↗