Recent advances in large language models (LLMs) demonstrate strong capabilities in contextual understanding and adaptive dialogue, creating new opportunities for in-vehicle conversational AI agents. In this research, I explore how such agents can be designed inside autonomous vehicles (AVs) to enhance the travel experience for older adults (65+).

I conducted a mixed-methods study (N=23; focus groups and interviews), in which focus groups identified accessible AV designs with conversational interfaces. Later, a virtual reality (VR) simulator and an LLM-powered chatbot were used for the interviews. Findings reveal that older adults prefer conversational agents with appropriate anthropomorphic characteristics (i.e., pacing, tone, and frequency) and value proactive social engagement and emergency support.

Brainstorming interfaces for unfamiliar technology like AVs can be challenging for non-designers. In this study, I built AV-Sketch, an in-situ (on-site) simulator that enables the creation of automotive interfaces while immersed in VR, allowing active participation in the design process. I conducted a participatory design study (n=15) using PICTIVE (Plastic Interface for Collaborative Technology), which led to the design of AV-Sketch. I further tested the tool in a design session (n=10) to assess users’ experiences in designing AV interfaces. 

AV-Sketch enabled better contextualization of design ideas in real-world scenarios, with improved spatial considerations and dynamic aspects of in-vehicle human-machine interfaces (HMI).

Many developing countries in the Global South face unique transportation challenges, including inadequate infrastructure, high traffic congestion levels, and limited public transportation access. Additionally, women and girls face significant safety risks (e.g., sexual harassment and assaults) when using public transportation, which is a significant barrier to women's mobility.

In this research, I conducted a mixed-methods study (focus groups and co-design) with Bangladeshi women (n=23) to study their daily struggles and concerns in public transportation. The qualitative findings informed the design of AVs to ensure greater mobility and enhanced safety in the Global South and other marginalized communities. 

Older adults (65+) are expected to be one of the largest user cohorts of AVs as they support aging in place through freedom of mobility. As aging is associated with sensory and cognitive declines, such as poor vision, hearing impairment, and dementia, they have unique needs and challenges while interacting with AVs as passengers. 

I led all project milestones, including the literature review, study design, VR prototyping (Unity, C#), study execution, data analysis, and translating findings into actionable interface design insights. Through this work with older adults (N=25), I distilled user input into inclusive design recommendations for AV interiors to easily access and ride autonomous vehicles without experiencing anxiety and frustration.

Despite widespread attention from academia, industry, and policymakers, Islamophobic posts are increasing online. There are significant advances in content moderation within HCI; however, social computing scholarship in this area remains constrained by Western interpretations of justice. This research aims to understand the issues of Islamophobia on social media and explores effective technological and policy solutions to address this problem. 

In this work, we developed an online tool for reporting and moderating Islamophobic tweets, using HCI principles with an ML-based pre-analysis score that estimates the likelihood of Islamophobic posts. I led the user-centered study design and conducted usability studies, contextual inquiries, and interviews with 32 participants to assess the tool’s effectiveness in addressing Islamophobic content. 

This research investigates the emerging domain of interaction between pedestrians and AVs to design accessible interfaces for vulnerable road users. This includes pedestrians with reduced mobility, like wheelchair users, and individuals with hearing impairments. My work has pioneered accessibility research on pedestrian safety in the automotive domain and is among the most highly cited in this area.

I conducted participatory design studies with interaction design experts (N=5) to explore potential interface ideas. Later, I performed iterative co-design sessions with wheelchair and hearing aid users to understand their unique perspectives and challenges in making crossing decisions as pedestrians. Based on the findings from the qualitative analysis, I developed VR prototypes (Unity, C#) simulating HMIs to establish a better communication medium between AVs and road users.

This project introduces External Automotive Displays (EADs) to provide visualizations that can share context and user-specific information and offer opportunities for direct and mediated interaction between users and automobiles. I conducted a study with interaction designers to explore design opportunities on EADs. Based on the design study, I prototyped four EADs in virtual reality and presented a preliminary design critique of the prototypes.

Academic Contributions:

[1] [AutoUI 20]: Designing external automotive displays: VR prototypes and analysis