My role

• Role: Project lead / Duration: 2.5 months (2018)
• I set the design direction, worked on the research and design with the team, reviewed every deliverable and presented the work to Li Auto.

Key decision: Don’t put every function on four screens at once. We reduced the HMI to its essential building blocks (account, car, services, voice control, notifications and apps) and designed the first release around them.

Project background

China's new energy vehicle industry was booming, backed by government policy and new infrastructure, and local companies were racing to build the next generation of cars. Li Auto (then called Lixiang) had a clear focus: families, at a carefully chosen price. Our brief was to explore what a next-generation HMI could look like across four screens, with internet services and voice control built in deeply.

Approach

Research: interviews with drivers and different kinds of passengers, then personas, scenarios and journeys, to find the key driving situations and the needs in each.

Chart placing four driver groups on axes from tool-like to emotional and from innovative to traditional — Four driver groups mapped by what they want from a car

Timeline of family stages from couples to children aged 8 to 12, with seating diagrams and car needs — How a family's car needs change as children grow, with the strongest switch point marked

Slide with portraits of 19 focus group participants, 2 ride-along users and 4 internal interviewees — Who we interviewed in Beijing and Shanghai: drivers, ride-alongs and colleagues

Table listing 25 interviewees with their cars, cities and interview formats, plus summary counts — All 25 interviewees at a glance, with the cars they drive

Market research slide with logos grouped by car makers, technology trends, services and consumer research sources — Market scan across car makers, in-car technology, services and consumer reports

Opportunity areas turned into questions about driving, family life and brand

Dark slide of a 20-day research plan with information sync, market research, interviews and ride-alongs — The 20-day research plan, from stakeholder talks and interviews to design ideas

Interaction design: the interaction structure and visual style for the four-screen HMI, with a prototype and a demo video.

Slide showing proposal 1 with three dashboard screens in a single and a full-screen immersive mode — Proposal 1: screens that follow each family member's account, with a full-screen mode

Slide showing proposal 2 with three dashboard screens using light card widgets and full-screen views — Proposal 2: lighter, card-based screens that connect visually across the dashboard

Table comparing proposal 1 and proposal 2 by core idea, framework and effect on perception — The two proposals compared side by side: person-centred versus space-centred

Table classifying voice assistant dialogues into user-triggered and system-pushed types with example follow-up actions — How the voice assistant responds, whether the driver asks or the car speaks first

Diagram of four screens with roles: driving centre, vehicle control centre and two interaction centres — The role of each of the four screens, from driving info to shared interaction

Visual design: the final visual design.

Final visual design of three dark interface screens on a car dashboard showing speed, navigation, calendar and music — The final design across the dashboard: driving info, family hub and passenger screen

Takeaways

How do you fit everything into a car?

Internet-connected electric cars invite big ideas, and four screens was a bold statement. The obvious question was what all that screen space was for. A screen that size opens up exciting options, but anything that pulls a driver's attention is a safety risk.

Look at what has to appear on screen and you find four separate systems: the dashboard (speed, fuel, lights), the central system (settings, connectivity, notifications), entertainment (media, games) and hardware controls (air conditioning, seat heating). For drivers to understand them, the design has to be precise and the information well structured. Most buyers have driven for years and have habits. A system that strays too far from what they know, including smartphone conventions, could put them off buying.

There was a business reason too. Going all-in risks spending heavily on features the market isn't ready for. It's wiser to find the core services people actually need, launch those with a solid plan, and give customers time to adjust. So the short answer is: you don't put everything into the car at once.

To set the scope, we mapped a system structure with only the essential building blocks (account, car, services, voice control, notifications and apps) and how they relate, shown on the right of the diagram below. It kept the first release focused.

How do you push the design further?

One of the most inspiring people I worked with in three years of consulting was Li Auto's director in charge of the whole vehicle, hardware and software. Across many discussions and workshops, he kept asking the right questions and didn't stop until everyone had really thought before answering. Those questions pushed the team to examine and refine the solution until it held together. Below is a record of how our thinking changed:

We moved from the practical level (user and market research, concrete problems) to an abstract one (turning everything into design elements), which strips things down to what is essential. How does element A connect to element B? What happens to that connection in situation C? Does the element change under different conditions? Working this way gave us a car system that is logical, well supported and simple.

Work done @ ARK Innovation Consulting / 2018

Works

Li Auto: a four-screen in-car system