Hyundai

Overview

During my final semester at UMSI, I took the UX Mastery Capstone course where I got the opportunity to work in a team of 5 in collaboration with Sookyung Cho ( UX Researcher at HATCI, CA) to develop solutions that make driving more accessible through the use of AI. To solve this problem, we developed a WCAG-compliant prototype called “Blue Guardian” a native navigation application for severe weather conditions that uses adaptive intelligence that empowers drivers with safer route guidance using crowdsourced data along with a personalized intelligent voice assistant.

Problem Statement

Solution

User Research

Driving is an essential part of life in the United States, from getting to work to long-distance commutes across states. But when inclement weather strikes, it can lead to dangerous situations, which in turn can lead to anxiousness among drivers.

Interviews

Initial user interviews were conducted with 9 drivers from age groups spanning 18-60+ years to dive deeper into the problems that active drivers face and how they navigate time-critical situations that they felt were dangerous. We also asked the participants a few open-ended questions about their thoughts on AI in general and how would they react to having an AI system actively assist with driving.

Survey

• We also surveyed 55+ people to get more information about drivers’ thoughts and feelings while driving through severe weather conditions.

• Conducting the interview and survey not only helped us get qualitative data, but also the numbers that support the responses received.

Stakeholder Interview

• Interviewing with our stakeholders at Hyundai helped us get more clarity on the problem statement and the aspect of AI we would be working on.

• Designing for the future meant we needed to ideate systems that don’t exist in the market, which was exciting but challenging considering we didn’t have enough knowledge about designing for cars. (This actually proved to be a major hurdle while validating ideas)

Concept Ideation

We conducted brainstorming exercises like Painstorming and Crazy 8s ideation based on our research findings and coalesced them into 30+ ideas that were mapped out using an affinity map, bringing it down to 3 concepts. Based on feasibility and use cases, the concepts were narrowed down to the Blue Guardian Navigation System. The core features that we wanted to develop being:

• Smart routing to suggest optimal routes that minimize the chances of encountering severe weather before driving.

• Multi-modal real-time alerts and alternate route suggestion while driving.

• AI-driven by real-time weather reports and crowdsourced data from credible sources.

• Ensuring the highest standards of accessibility, improving usability and safety.

Wireframes

While our initial research gave us clarity on who we are designing for, we still needed to get clearer picture of our deliverables (low fidelity). Further research was conducted to understand:

• Vehicle HMI sitemaps

• Competitive Analysis for cutting edge technology in vehicles

• In-vehicle testing @HATCI Ann Arbor, to understand the latest production line from Hyundai Motor Company

• Literature on Vehicle Safety, HMI Design and Accessibility in Automobiles

We created multiple variations of the low-fi mocks for the concept test, to understand the structure and quantity of information that users preferred, accessible areas on the center display, and tone of AI assistant. The concept test was conducted with 4 people with these variations and revealed 4 key themes:

Mid-Fidelity Prototype

• While users saw potential in real-time assistance, intrusive alerts and vague risk assessments made them hesitant to rely on it fully.

• Users liked the idea of a system that helps them navigate bad weather safely, but they weren’t convinced it could always make better decisions than their own experience all the time.

• These insights made us realize that the system needs to balance AI-driven recommendations with user autonomy—it should assist, not dictate.

• The mid-fidelity prototype consisted of the following flows.

  • Before driving - Choosing a route

  • While driving - Providing safety updates

  • While driving - Offering alternate routes in congestion

This version also presented more convincing visuals for route information, to help users prioritize safer routes over quicker ones.

Hi-Fidelity Prototype

• After client design reviews with mid-fi screens, we designed high-fidelity prototypes that had more depth and detail, essentially replicating a real-life situation.

• In order to ensure consistency and to streamline UI elements, we created a design system consisting of colors, typography, safe areas, icons, and components.

Usability Testing

Design System

Figma Prototype

Outcomes

Our team presented this prototype to the HATCI teams located in Ann Arbor, MI and Irvine, CA over a virtual conference. We also created a tri-fold poster and presented it at the University of Michigan School of Information Project Exposition.

Reflections

• As a team, we effectively delegated responsibilities based on individual team members' strengths, ensuring a balanced workload and efficient progress throughout the project.

• We leveraged diverse skill sets within the team, aligning tasks with each member’s expertise in research or design to maximize impact and output quality.

• Since the start of the project, we prioritized analyzing research documents, laying a strong foundation for ideation and ensuring user-centered insights guided the entire design process.

• We navigated and resolved conflicts within the team and addressed misalignments between academic and client expectations through open communication and collaborative decision-making.