
Finedrive AI
A stick-type navigation device used by inserting it into the vehicle's USB port
A revolutionary stick-type navigation designed for use by inserting it into the vehicle's USB port, leveraging the car's MirrorLink (phone projection standard).
I have planned the Conversational Voice UX for Fine Voice, the product's voice agent.
Through Fine Voice, users can control the navigation service environment with interactive voice commands.
Additionally, I have overseen the planning of LED feedback scenarios displayed to the driver on the navigation stick.
VUX Design Process
I have understood the technology, services, and policies of MirrorLink (Phone Projection) and derived insights into essential voice interaction methods during driving through user research.
Based on this, I have designed an optimal conversational UX scenario for use during driving. Additionally, I have progressed with an ideal integration scenario between VUX and the Findrive AI product.
Understading
Product & VUX

UX Ideation



Interaction Design

1 Understanding of the product and MirrorLink
2 User Survey for VUX Development
: Defining Service Scope Based on Needs
3 Discovering Design Principles for VUX
1 Defining the Scope of VUX Service Provision
2 Defining Tasks at the Sentence Level
3 Conversational UX Design
1 Overall 대화 Flow map
2 LED 동작 시나리오
1 Overall 대화 Flow map
2 LED 동작 시나리오
1 Overall Conversation Flow Map
2 Usability Evaluation
3 LED Operating Scenario
Understanding of stick-type navigation
FINEDRIVE AI is a USB-type aftermarket navigation system that utilizes the MirrorLink Phone Projection standard.
It functions as a versatile communication device, allowing for the installation and utilization of various applications.
With support for navigation, music, video, and DMB,
it is compatible with MirrorLink-enabled vehicles, providing access to a wide range of MirrorLink-verified apps available on the Google Store.



LED Lens
USB
“Insert a USB stick into the USB port and start the navigation”

Mirrorlink Protocol
based on Phone Projection
Mirrorlink
Common API
Mirrorlink
Common API
Mirrorlink
Common API
Application
Application
Application
MirrorLink
Client
(Head Unit)
User
MirrorLink
Server
(USB Stick)
MirrorLink is a protocol designed to use the car's head unit (Client) as a remote terminal for the smartphone (Server).
It enables the execution and control of smartphone applications directly from the car.
FineDrive has developed a USB stick that serves as a Server, inserted into the vehicle's USB port instead of using a smartphone.
FineDrive's navigation and other service apps have been developed to comply with the MirrorLink standard and integrated into the USB stick.
All apps must be provided through the Application Common API, adhering to the MirrorLink standard app.
MirrorLink and the analysis of the MirrorLink market
MirrorLink is a phone-projection standard that is defined by the international standards organization C.C.C. (Car Connectivity Consortium).
MirrorLink® is an open standard for car-smartphone connectivity which allows users stay connected to their smartphone apps while driving.
Drivers interact with their compliant smartphone apps via any compliant in-vehicle infotainment system (IVI).
The smartphone applications project their user interface (UI) to the IVI display.
Audio apps can also stream and play through the car’s speakers. MirrorLink enables interoperability between a wide range of smartphones and cars.
It also allows all players in the ecosystem to have an equal stake in the booming connected-car marketplace.
MirrorLink® is available today in millions of cars, from 100s of millions of mobile devices, powering dozens of applications.
I attended the Working Group meeting at the CCC Members Meeting to discuss issues related to mirroring link certification (September 2018, Tokyo).
MirrorLink Compatible Smartphone Models
Samsung Electronics, LG Electronics, and other international brands offer MirrorLink on a variety of Android-based smartphones.
Samsung Electronics: Galaxy S series (S5 ~ 10), Note series (NOTE 4 ~ 9), A series (A5 ~ 8)
LG Electronics: G series (G6 ), V series (V20)


MirrorLink Supported Vehicle Models
MirrorLink Applied to All Hyundai/Kia Car Models Released Since 2018
Hyundai/Kia vehicles currently hold a market share of over 80% in the South Korean automotive market
MirrorLink is currently being applied to some models of imported cars such as Volkswagen, Mercedes-Benz, Peugeot
and gradually expanding



It's a technology defined by the international standard organization C.C.C. (Car Connectivity Consortium), providing a communication standard specification that simplifies and conveniently connects smartphones with in-vehicle infotainment systems.

With the advancement of Connected Car technology, the Car Connectivity Consortium (C.C.C.) plays a crucial role in establishing international standards and regulations for connected car technologies
MirrorLink is currently available in 57% of the Hyundai
and Kia vehicles being sold, with ongoing efforts to
expand its integration to the entire vehicle lineup.
It is also featured in key imported car brands such as
Mercedes-Benz and Volkswagen. (As of 2018, based on
the Korean market standards.
MirrorLink Supported Status
in Domestic Vehicles
(Supply Status as of 2018)
MirrorLink Supported
MirrorLink Not Supported
42% of Domestic Car Models
58%
42%
57% of All Kia Car Models
43%
57%

Select "SETUP" from the vehicle navigation menu.

2. Select "Phone Projection" within the
settings menu.

3. Check the "MirrorLink' option and
connect your smartphone.

4. When connecting the smartphone, the
"MirrorLink" icon is automatically activated.
MirrorLink
Usage
Example
(Hyundai
Motor)
The trend of vehicles equipped with in-car infotainment displays
(AVN/AVC) and supporting phone projection options (Android Auto,
Apple CarPlay, MirrorLink) is on the rise.
The rise in vehicles providing in-car infotainment displays and
integrating phone projection features is anticipated to elevate the
demand for mirroring products.
Sales outlook for Hyundai-Kia vehicles (OEM: Before Market) with Phone Projection capabilities
60,000
45,000
30,000
15,000
2019
2020
2021
2022
2023
AVN (Equipped with an infotainment display)
AVC (Equipped with an infotainment display)
To develop VUX, what functionalities should we offer?
To investigate the functionalities that drivers expect from interactive voice AI during driving,
we initially derived a list of frequently used features in navigation through the analysis of existing user log data. Additionally,
we conducted a survey among users to inquire about the functionalities they deem necessary through VUI (Voice User Interface) during driving.
The results from user surveys were then integrated with the existing user log data to provide a comprehensive overview.


Quantitative Research
[Data is Extracted from the Log Analysis Report]
What features do drivers expect from VUX (Voice User Experience)?
24,690 / 42.6%
Recent destinations
15,647 / 27.0%
Search (Location/Address/Phone Number)
9,952 / 17.2%
Favorites (Saved Locations)
4,002 / 6.9%
Cancel Route
1,024 / 1.8%
Move Map to Explore Location
852 /1.5%
Navigate to Home
829 /1.4%
Volume Up
691 /1.2%
Nearby Search
Others
Log Data Analysis for Atlan Navigation
[Data-driven decision-making]
Collecting and Analyzing User Log Data
[Key User Touch Event Values]
[Log Analysis Report]
[Other Collected Data]
Screen
Number

Screen
Dwell Time

Detailed
Configuration Settings
and Preferences

Executed
Applications

Details of Content Playback

Car movement (GPS)
path and duration
of Movement

Heatmap Analytics
by Cohort Groups

Scroll

Touch
(First/Last)

Swipe

Zoom In/Out

Drag&Drop

Button Value
Screen Coordinates

Occurrence Time

Event Value
Segmentation
by Cohort Groups

Extracting meaningful insights through data analysis














Continuous Collection of User Log Data to a Data Cloud Server
(Collection Conditions: Consent and Network Connection Required)


[Log Data]
Analyzing log data and heatmap information on both a per-time frame (short/long) basis and across Cohort Groups







Quantitative Research
*Kano Model : Evaluating satisfaction among individual value factors within
the framework of user satisfaction criteria to determine the
value for strategic planning
An online survey was conducted targeting a total of 50 participants, including users of aftermarket navigation systems, OEM navigation systems, and smartphone navigation.
Additionally, in-depth interviews were carried out with 8 individuals.
We selected keywords based on the results of both online and offline surveys.
Taking into account customer satisfaction and requirements, we applied the KANO Model to derive elements such as exciting factors, satisfying factors, and essential factors.
Recruiting

Navigation Users
> Users of aftermarket navigation systems,
built-in (pre-market) navigation systems,
smartphone navigation apps
> Users of smart services such as Voice Agents,
IoT, and other smart functionalities."
Raw Data Collection
Compiling all raw data derived from online surveys
and user in-depth interviews

Keyword Selection
Selecting keywords by grouping them based on
major common issues

KANO Model

Customer Satisfaction

Satisfactory
Needs
Requirements
Delightful Needs
Essential Needs
Uncovering potential needs through the
classification of user needs based on keywords
Identifying key tasks based on the three main
factors of Kano model
Delightful
Needs
Satisfactory
Needs
Essential
Needs
Online Survey



Online Survey
(Quantitative Analysis)
> Understanding Navigation
Service and Infotainment
System Usage Patterns through
an online survey
In-depth Interviews

User In-depth Interviews
> Conducting in-depth interviews with
users based on aftermarket
navigation, built-in (pre-market)
navigation, and smartphone navigation


Qualitative Research
Exploring design principles for VUX based on the KANO model
We developed design principles for VUX based on the criteria of essential elements, satisfying elements, and exciting elements.
Subsequently, we proceeded with the development of VUX scenarios, taking into consideration these principles and their respective characteristics.

How long does it take to the service area ?
Where's the nearest gas station ?
Let's go to Oxford Circus
Give me nearby traffic information

How can I help you ?
Delightful Needs
Enjoyment / Interaction
Beyond driving-related features,
offering satisfaction by engaging
in diverse conversations on topics
that the driver desires
Essential Needs
Multimodal Interaction
In addition to voice, providing
touch interaction allows for
effective handling of exceptions
such as noise
Essential Needs
Short-cut Access
Despite limited touch depth,
an opportunity area can be
provided through voice
Essential Needs
Providing commonly used
key features
Providing key features
used in touch interaction
Satisfactory Needs
Full control coverage
Enabling voice control for
all apps beyond navigation
Satisfactory Needs
Context-aware &
Usage pattern
Providing appropriate suggestions
and responses based on an
understanding of driving situations
or usage patterns
Essential Needs
Clear feedback
Offering a clear AUI and dynamic
GUI for situational awareness
during speech and recognition



Definition of the Scope of VUX Coverage
Based on the results of user surveys, I defined the task list that we aim to provide.
The task list was derived within the FineDrive Navigation IA, and the main task flow that can be provided through VUX was structured.

Nearby
Search
Gas
Station
Parking
Lot
Popular
Dining Spot
Gas
Station
Parking
Lot
Popular
Dining Spot
Search Near Destination
Remaining
Distance
Toll Fee
Route information
Delete Set Waypoints
Delete waypoint
Zoom In/Out Map
Zoom In/Out
Search for
the destination
POI Search
(POI: Point of Interest)
Destination Detailed
Information
Home
Work
Favorites
Recent
Destinations
Address
Search
Waypoint Detailed
Information
POI Search
(POI: Point of Interest)
Address
Search
Change
waypoints
Add waypoints
Waypoint Detailed
Information
POI Search
(POI: Point of Interest)
Address
Search

Navigation: Atlan Auto
Weather
information
The current
location weather
Local
weather
Destination
weather

Weather
Built-in
Apps
Navigation
Fine Music
DMB
Other apps
Fine Photo
Turbo GPS


Built-in apps
from Findrive
Media Control
Music Apps
Video Apps

Media
(Third-party apps)
Traffic
Information
Traffic information
for the current location
Highway traffic
information
Destination traffic
information

Traffic
Screen Categorization Within the Scope of VUX Coverage
Navigation guidance
(Built-in Navigation)
Voice Agent activation
Voice Agent : Voice Command re-request
Voice Agent : Destination search results
Fine Music
(Built-in App)
Fine Music > Album list
(Built-in App)
DMB (Digital Multimedia Broadcasting)
(Built-in App)
Fine Photo
(Built-in App)
Turbo GPS
(Built-in App)









Designing conversational UX
AI should support both single and multi-turn conversations.
A multi-turn conversation is a conversation with multiple turns in which AI assistant asks questions and the user responds with the answers.
[Basic VUX Guideline]
Intent
Required entity
Optional entity
Fine Voice's
Wake word
"Hi, Fine, please show me the fastest route to Samsung Station."


What's the hottest stock lately?


Sorry, I don't have
stock information.
Utterance Design
Prompt Design

Out of Service
The Fine Voice Agent understands the user's intent, but if the request falls outside the supported service scope, clear guidance is provided for the unsupported service.



Thanks for the quick response!

Utterance Design
Prompt Design

Multi-turn : Add-listen and Answer
After completing the task, provide casual greetings based on the user's intent.


I'm glad I could help!
If you have any more
questions or need assistance,
feel free to ask.



Could you please
repeat that?

I'll put on the sports channel
Show me.

Sports.

Utterance Design
Prompt Design



Multi-turn : Re-listen
If the Fine Voice Agent fails to fulfill the entity or intent, it inherits from the preceding utterance.
* Multi-turn : A type of conversation where the user and the Fine Voice Agent engage in dialogue with two or more exchanges after the initial Fine Voice Agent call.


Remove the waypoints

First one!


Which waypoint would you
like to remove?

Delete the first waypoint and
guide to the second waypoint.


Multi-turn : Slot-filling
Type of conversation where the Fine Voice Agent interacts with the driver for more than two exchanges:
> If, after the agent is called, essential entities are not fulfilled, the agent asks the driver for the necessary
information.
Utterance Design
Prompt Design




Feeling bored during highway driving?
Shall I play some of your favorite dance
music to liven things up?
Play some R&B, my usual favorite.


Sure, I'll play some of the R&B music
you often listen to on Fine Music.
Utterance Design
Prompt Design

Multi-turn : Feed-forward (Context aware)
> Fine Voice Agent proactively suggests based on the user's driving context comprehension.




Tomorrow in Seoul, the sky
will be cloudy, and rain is
expected in the afternoon.

Tomorrow in Busan,
it's expected to be clear all day.
Tell me the weather in Seoul tomorrow



Multi-turn : Slot-merging
> If the Fine Voice Agent is called and the entity or intent is not fulfilled, it inherits from the preceding utterance
Utterance Design
Prompt Design
Busan!




Sure, playing it for you right away!


Your frequently listened 9 O'Clock
Radio News is about to start.
Would you like to listen now?

Single-turn : Feed-forward (Usage pattern)
> Based on the user's usage patterns, the Fine Voice Agent proactively suggests options.
Utterance Design
Prompt Design



Tell me about the traffic
conditions near the destination!


The traffic flow near the
destination is currently good.

Single-turn
FThe type of conversation where, after calling Fine Voice Agent, the dialogue concludes after the agent
responds to the user's request
Utterance Design
Prompt Design

Conversation Flow map
The suggested concepts are separated into detailed units of functions. For a Fine voice agent,
each function consists of a logical flow of conversation and value-processing logic.
So, I designed the detailed conversation scenarios that work with the corresponding screen for the Fine voice agent to deliver the concept to a real service.
The scenarios are designed to provide a good experience for users involved in the conversations.

Fine Voice Agent says :
Where would you like to go?

Fine Voice Agent says :
“There is currently one waypoint.
Please set additional points directly
on the screen for use."

Fine Voice Agent says :
“This feature is only available
during navigation guidance.”

Fine Voice Agent says :
"Which numbered place
would you like to go to?"


Fine Voice Agent says :
“I'll guide you through
the fastest route.
Please say 'Start Guidance'
after confirmation.”


Fine Voice Agent says :
“Guidance started.”




User says :
“Start guidance!"

Fine Voice Agent :
Run Slot Filling
Required : Destination
Additional : Guidance Options >
Fastest Route, Minimum Time,
Toll-Free Roads, Shortest Distance,
Priority on Highways)

Fine Voice Agent :
Search for a destination
Input : The name of the destination

Fine Voice Agent :
Search for a route
Input : Gangnam Station, Exit 10
Check : The current navigation
options that are set
No waypoints
One or more waypoints
Multi modal
If not during route guidance



User says :
“Hi, Fine, please add a waypoint
on the way to the final destination."



User says :
“Let's go to Gangnam Station!"



User says :
“Second one."
No.
Status
LED color
LED Operating Scenario
LED
1
Engine Start
(including ACC activation)
Blue
Illuminates with a rapid clockwise
rotation for 5 cycles

2
Engine off
Blue
Illuminate with a rapid counterclockwise
rotation for 5 cycles as the LED gradually
fades away

3
Voice recognition
activation
Blue/White
With a 'ding!' sound, the color changes
to white/blue


4
Command
processing
Blue/White
Rotate rapidly clockwise and returned

5
Voice response
Blue/White
The color has faded and returned

6
Data network
connected
Blue/Orange
Rotate five times clockwise

7
Data network
not connected
Yellow
Rotate ten times clockwise

8
Device error
occurred
Yellow
Flash rapidly until the error is resolved

9
SW Update
Rainbow
Rotate clockwise until completion

10
Factory reset
Rainbow
Rotate clockwise until completion

11
Recognition of music
files on the SD card
Blue
Illuminates with a rapid clockwise
rotation for 10 cycles

12
Launch the app
Blue
Illuminates with a rapid clockwise
rotation for 3 cycles


LED operation scenarios
Finedrive AI offers a variety of LED display methods based on the product's operational status to assist users with intuitive recognition.
We detailed and proposed specific operational scenarios for LED displays based on different situations.
I proposed an interactive LED prototype that responds to different operational statuses for the product.
This prototype is designed to visually represent the product's various operating statuses through dynamic color changes.
Interactive Color-Changing LED Prototype
[The key screens of the navigation where log data is extracted]
Copyright © 2024 Min Suk Kang. All rights reserved.