PhD Dissertation Proposal by Smitha SHESHADRI | Rethinking Indoor Localization and Tracking: A Conversational Approach to Positioning

Rethinking Indoor Localization and Tracking: A Conversational Approach to Positioning

ABOUT THE TALK

People spend much of their time in large indoor environments such as work and school campuses, shopping malls and commercial complexes, transit hubs such airports and train stations, and cultural spaces like museums and art galleries, where the ability to localize and track user movement is critical for navigation and context-aware services. While GPS supports positioning outdoors, it cannot be used indoors, and existing alternatives either require costly infrastructure or suffer from issues such as drift and scalability. This proposal investigates a different perspective: treating the user as a sensor. Instead of relying solely on devices, the work shows how people’s natural descriptions of their surroundings can serve as lightweight, conversational inputs to achieve indoor positioning. Three strands of research form the basis of this proposal: (1) Conversational Localization, which studies whether user-provided descriptions can support one-shot indoor positioning; (2) Conversational Tracking, which integrates conversational cues with inertial data to correct drift in continuous trajectory estimation; and (3) Modeling Conversational Positioning, which investigates modelling techniques to project system performance across varied environments without the cost of deployment. Together, these investigations aim to establish conversational interaction as a practical, infrastructure-free paradigm for indoor localization and tracking.

SPEAKER BIOGRAPHY

Smitha SHESHADRI is a PhD candidate in Computer Science. Her research focuses on Human–Computer Interaction and Indoor Spatial Intelligence, with an emphasis on conversational, user-as-sensor approaches to localization and tracking.