KINGSTON, R.I. – July 22 – The College of Arts and Sciences computer science staff celebrates a recent win for Best Demo Paper at the 26th IEEE International Conference on Mobile Data Management (MDM). At the event held from June 2-5 in Irvine, California, co-writers Professor Abdeltawab Hendawi, Gyanko Issah Yussif (Ph.D. student), and Marwan Abdelatti (Postdoc) walked away with the award for their paper, Harnessing Crowdsourced Mobile Data And LLM for Dynamic and Accessible Pedestrian Routing.”
The paper investigates how real-time, user-contributed data combined with large language models (LLMs) can transform pedestrian navigation on urban sidewalks, especially for those with differing mobility needs.
“Winning the Best Demo Paper Award at IEEE MDM 2025 is deeply meaningful for our team. It affirms our conviction that AI can be a powerful force for equity and accessibility, and it inspires us to continue advancing intelligent systems that prioritize not just innovation, but a human-centered approach,” said Hendawi.
The idea was sparked by a personal experience: after sustaining a foot injury and multiple fractures, Hendawi relied on crutches for several weeks. Navigating city sidewalks became a daily challenge. Every step and surface mattered—he frequently had to reroute due to blocked sidewalks, uneven paths, or a complete lack of accessible options. That frustration highlighted both the urgent need for better pedestrian support tools and the widespread nature of the problem.
For example, in 2020 alone, New York City settled 13,741 slip-and-fall claims, totaling $1.03 billion. Clearly, inaccessible and hazardous sidewalks are not just a personal challenge but a significant public concern.
To address this, the team’s mobile application builds upon their previous Prego system, which enabled multi-preference, personalized routing for roadways. The new core system extends those capabilities to pedestrian navigation. It integrates LLMs to provide contextual assistance and adaptive, personalized routing tailored to individual mobility needs.
A key feature of the platform is its community-driven model: users can report current sidewalk conditions, hazards, and blockages, contributing to real-time, dynamic maps. The system also supports agentic interactions via the AI assistant, allowing users to issue voice commands to reroute, submit reports, or inquire about accessibility conditions. This gives pedestrians greater control and agency over their navigation, improving safety and independence for all.
Since 1999, the MDM series of conferences is a prestigious forum for the exchange of innovative and significant research results in mobile data management. The conference provides unique opportunities to bring researchers, engineers, and practitioners together to explore new ideas, techniques, and tools, and exchange experiences.
Comprising both research and industry tracks, it serves as an important bridge between academic researchers and industry researchers. Along with the presentations of research publications, it also serves as a meeting place for technical demonstrations (demos), advanced seminars, panel discussions as well as Ph.D. forum and Industrial forum to cater Ph.D. students and industrial developers.