QIONGZHI (KAREN) XU
← ALL WORK
HMI DESIGNINTERNSHIP2026IN PILOT

KION Group LiO

Designing the touchscreen for LiO, Linde's autonomous guided vehicle, so operators always know where it's going, what it's doing, and what to do next.

TIMELINE

May – Aug 2026

ROLE

Product Development and Design Intern

TEAM

Product managers, engineers

LOCATION

Linde Material Handling, Xiamen

LONG STORY SHORT

I led UX/UI design for an autonomous guided vehicle and its warehouse management system, in a team getting ready to launch.

01

Design

Designed and prototyped the HMI end to end, from first setup to the everyday dashboard.

02

Research

Studied warehouse workflows and competitors, and turned usability problems into product requirements.

03

Cross-functional

Worked with PMs and engineers, and built AI workflows that helped engineering and business teams share knowledge.

THE CHALLENGE

How do we make the robot interface simple, clear, and usable?

An AGV runs on its own, but people still set it up, give it tasks and step in when something changes. Standing next to a moving forklift, an operator needs to understand three things at a glance: where the robot is going, what it is doing, and what to do next.

COMPETITIVE ANALYSIS

How other robots talk to people

Before designing, I studied competing AGVs in the Chinese market, tried operating existing AGVs myself, and talked with engineers about how they use the machines day to day. The same problems kept showing up.

~6 months

for engineers to learn to use an AGV with an older interface. And that was for the professionals.

FINDING 01

Designed by engineers, for engineers

Most interfaces were built around the machine, not the person using it. They showed technical settings instead of the task in front of the worker.

FINDING 02

Setup is a wall of steps

Getting started took a long chain of steps. For anyone who is not a professional, that was hard to follow and easy to get wrong.

MY RESPONSE

Guide people, one step at a time

LiO walks first-time users through setup with a clear stepper, lets them skip what is optional, and gives every screen one obvious next action.

DESIGN PRINCIPLES

Guidance

"What should I do next?"

Every screen has one clear next step, and setup shows the whole path with a stepper.

Feedback

"Is everything working?"

Live run time, distance, speed and task progress, plus colour that changes as the robot moves.

Orientation

"Where am I?"

A simple route map with named pick, drop, charging and parking points.

DEFINING THE PRODUCT

26 screens, written down before they were drawn

Before designing, I wrote the HMI UI requirements: every screen, its purpose, who uses it, and the one action it asks for. Errors and pop-ups were planned as their own states from the start.

5

Startup

Power on, system check, errors, returning users

13

Onboarding

Map the space, set points, run a test

6

Daily operation

Create, start, follow and finish tasks

2

Settings

Manage and share maps and routes

DESIGNED FOR First-time usersReturning usersOperatorsWorkers at stationsTechnicians

USER FLOW

Mapping every path, and every question

The first full user flow covers first-time setup, returning users and daily tasks. I pinned open questions for engineering beside it, like how to make sure operators park the AGV in the right area when switching from manual to auto mode.

User flow diagram branching from power on into first-time setup, returning users and daily task paths

THE FLOWS

From unboxing to the first task

Setup screen with language selection and a four-step progress bar

01

Setup

Language and Wi-Fi come first, with a stepper that shows every setup stage ahead.

Pairing screen scanning for AGVs and listing AGV 1, AGV 2 and connected AGV 2

02

Pair AGVs

Scan for robots nearby and connect one or several to the same tablet.

Mapping screen with a route of pick, drop and intersection points

03

Map the floor

Drive the route and tap to mark pick, drop and intersection points. Charging and parking points can be skipped.

Button pairing screen with a 3D button and assign point and assign route actions

04

Pair physical buttons

Assign a hardware button to a point or a route, so workers can call the robot without the tablet. This step is optional.

Dashboard with a tooltip explaining the map, step 4 of 4

05

Dashboard onboarding

A short guided tour points out the map and task controls on first launch.

Dashboard showing the AGV approaching a pallet with task progress from P1 to P2

06

Run a task

Pick a start and end point, then follow the robot live on the map and in the progress bar.

A CLOSER LOOK

Docking you can read at a glance

As the forklift lines up with a pallet, its path changes from red to amber to green. Continue Task only lights up once the robot is in place, so an operator never has to guess whether it is safe to go on.

Docking view with a red path, 3 metres from the pallet, Continue Task disabled

3 M Far off, path in red

Docking view with an amber path, 2.5 metres from the pallet, Continue Task disabled

2.5 M Getting closer, amber

Docking view with a green path, 0.5 metres from the pallet, Continue Task enabled

0.5 M In place, ready to continue

DESIGN SYSTEM

One turquoise, tuned for the floor

I set up the HMI foundations as tokens. Turquoise is the main brand colour, sitting on a dark base, with green, yellow and red kept for status only, and contrast checked for every colour.

Brand · turquoise

turquoise-100
#E5F8FB

turquoise-500
#42D6F0

turquoise-600
#25A3BA

turquoise-700
#0B7184

turquoise-900
#124A54

Feedback

green-500
#35CD05

yellow-500
#FFAD00

red-500
#FF0000

red-700
#CC0000

Neutrals

black
#000000

coolgrey-900
#222222

coolgrey-700
#444444

coolgrey-400
#9B9B9E

coolgrey-100
#EEEFF3

white
#FFFFFF

Typography · Noto Sans

Title XXL

64 PX · BOLD

Title XL

48 PX · BOLD

Title M

32 PX · BOLD

Title S

24 PX · REGULAR

Body S

18 PX · REGULAR

OUTCOME

IN PILOT

LiO launched at Automation Summit 2026, and is now in pilot.

LEARNING TIME · OLDER INTERFACES

~6 months

for engineers to learn the machine

LEARNING TIME · LIO

< 1 day

the goal LiO is designed for, for engineers and non-experts alike

WHAT I TOOK AWAY

Simple beats impressive

LiO runs on complex technology.

Good design isn't about looking good. It's about turning a complicated technical process into something anyone can use, even someone who has never touched an AGV.

Two users, two jobs

Engineers set it up. Workers just use it.

Engineers care about mapping, routes and device management. Workers at stations care about starting tasks, checking status and finishing the job. Clear steps for setup, and everyday actions where they are easy to find.

The map is the hero

Operators read the screen while walking.

We reworked the information priority of first-time mapping again and again. In the end, the live map became the visual centre, with tips and buttons placed where a quick glance can find them.

NEXT PROJECT Infinite Packaging, Amazon →