case 02 · Master Digital Design, HvA
Feel the way: vibration-first wayfinding for blind travellers
At every fork in the tactile path, your phone vibrates which way to go, taken from the trip you already planned. No scanning, no stopping, no asking a stranger.
Drop screenshot here
fig. 1
Set up once, then the phone guides from your pocket: choose how to be guided, a turn at the junction, guiding on.
problem
At an unfamiliar station, blind and low-vision travellers struggle most with finding the right stairs, exit or platform. Asking a stranger works, but it costs energy and independence.
my role
I led the haptic language and the app flow inside the NS app, synthesised the co-creation session and interview, and built the Wizard-of-Oz demo with the team.
result
A working haptic prototype tested with a fully blind traveller. Her feedback cut the design from 6 vibration cues to 3 cues plus voice.
01 · context
The fork is the hard part
Tactile guidance lines lead you through a station, but where they split, nothing tells you which branch goes to your destination. Many stations have one set of stairs in the middle of a long platform: pick the wrong direction and you walk all the way back.
Drop screenshot here
fig. 2
Our laser-cut junction tile (MDF prototype by Hannah Wolff) and the cue it triggers on the lock screen.
the brief
One physical input and one sensory output, demoed in a realistic transport scenario.
Our question: how might we help visually impaired travellers orient themselves at a station with less effort than asking a stranger?
02 · research
What two travellers taught us
We ran a co-creation game with Steijn (26, lost his sight about three years ago) and used an interview with a traveller born blind (31, by a fellow team). I clustered both transcripts into insights.
asking costs energy
Both rely on strangers at complex stations → guidance must be faster than asking.
the ears are full
Constant speech competes with VoiceOver and the station itself → vibration first, voice only on demand.
confirm in real time
“North” is useless; “left, here” is what helps → one cue, at the moment of the decision.
03 · key decision
A checkpoint, not another app
I proposed making the tile a checkpoint and taking the destination straight from the trip already planned in the NS app. At the station there is nothing to choose, read or scan.
We renamed the project from “Invisible Cane” to “Line Link”: it links the line you already follow to your next step, and it never replaces the cane.
04 · solution
One continuous experience inside the NS app
Set up once, plan a trip as usual, then put the phone in your pocket. Designed to WCAG 2.2 AA: 20 pt text, 64 pt buttons, VoiceOver labels, no meaning by colour alone.
journey
Your destination comes from your trip
Plan your train as usual. Line Link reads the platform, transfer or exit from the journey and starts guiding when you get off the train.
Drop screenshot here
fig. 3
Plan a trip, switch Line Link on: the destination is already known.
at the junction
The tile marks the junction, the phone gives the direction
A Bluetooth beacon in the tile tells the app you reached a junction. The app checks your route and sends one vibration; a time-sensitive notification carries the same instruction in text for VoiceOver.
Drop screenshot here
fig. 4
One vibration plus the same words on the lock screen: “Turn left”, “Turn right”.
haptic language
Six cues built from three “letters”
Tap, hold and flutter. Turns use only taps and holds; anything that starts with a flutter is information, not a turn. No cue is the start of another, so you never act on half a pattern.
Drop screenshot here
fig. 5
v1: learn the vibrations. Turns first, information starts with a flutter.
choice + demo
Your senses, your settings
Travellers choose vibration only, vibration + notification, or vibration + a short voice hint, and can practise every pattern. In the demo a teammate triggered each cue from a laptop (Wizard of Oz, Pulsar haptics) while the phone stayed in the pocket.
Drop screenshot here
fig. 6
Onboarding choice and guidance settings, with vibration strength and practice.
05 · test
Tested with a fully blind traveller
On demo day a fully blind participant tried the prototype while a teammate triggered each vibration from a laptop. She liked the concept, but six cues were too many.
6 → 3
vibration cues after testing
1
fully blind tester on demo day
3-5
travellers for the next test
what worked
- She liked the concept
- She would learn three cues
what changed in v2
- Vibration only for left, right and straight
- Everything else as a short voice hint in one earbud
- Replay without the phone: double-tap the earbud or say “Repeat”
Drop screenshot here
fig. 7
v1 → v2: six vibrations became three, everything else is spoken. v2 is not tested yet.
Drop screenshot here
fig. 8
v2 haptic language: three patterns built from a tap (120 ms) and a hold (600 ms), with a 250 ms gap.
06 · edge cases
Designed for when things go wrong
One rule for every edge case: vibration only means a turn. Everything else is spoken in the earbud and also shown as text, so VoiceOver users and travellers without an earbud get the same message.
wrong way
“Wrong way, stop. Turn around.” The stairs to track 7 are behind you.
lost the line
No tile for two minutes: stand still and feel for the ridged tiles, or call NS Travel Assistance.
earbud off
Turns still vibrate. Voice hints arrive as notifications until you reconnect.
track changed
“Track changed.” Guiding switches to the new platform, keep going.
Drop screenshot here
fig. 9
Wrong way, lost the line, earbud disconnected, track changed.
Drop screenshot here
fig. 10
Checked for VoiceOver order, a light high-contrast mode and text at 150%.
07 · design system
A small, accessible design system
Colour, spacing and radius variables with a dark and a light mode, text and effect styles, components with variants. Every screen is bound to it, which is why the light theme took one switch.
rules
Nothing below 16 pt, 64 pt buttons, never colour alone.
haptic tokens
Tap 120 ms, hold 600 ms, gap 250 ms. No pattern is the start of another.
Drop screenshot here
fig. 11
Tokenised components with variants, and the haptic tokens.
reflection
What I’d take to the next project
I designed vibration-first because of one person’s strong need. Our demo-day tester showed the opposite. In accessibility there is rarely one user, so I now design the choice first, and test the learning curve.
← all work
next: Virtual Try-On →