Revolution is a car dealership test driving experience that builds user acceptance for up to Level 4 Autonomous Vehicles (AV). Using a personalised set up process and a curated simulation, Revolution builds users’ confidence by addressing their fears of how the AV handles difficult driving situations.
My role
UX Researcher, UI and Prototype designer.
Duration
Mar - May 2021 10 weeks
Course
Information Architecture IACT 370 Professor Sung Park Professor Ascanio Di Paliano
Revolution also received 4 Silver and 1 Bronze Indigo Design Award.
Topic of Interest
Building a sense of comfort and control for the passengers & drivers of AVs with the help of emotional recognition technology. Inside a vehicle, the amount, quality, and delivery of information is essential not only for a good experience, but protecting the lives of users and their passengers.
We want to rework the pain points users feel in their experience interfacing with the vehicle by utilizing emotional recognition technologies to keep our passengers informed, entertained, and safe.
Initial Problem Statement
AVs with highly intelligent emotional recognition technology are the future. However, most people do not completely understand how this technology works, leading to a lack of trust with autonomous vehicles.
Target Audience
United States residents
Millennials Age Range: 25-45
Current or future drivers of Level 4 Autonomous vehicles
While this space is working with advanced technology that is not yet widely accepted or understood by the population, our expected outcomes for this project is grow Autonomous empathy, a sense of trust and safety in AVs, leading toward building a higher acceptance for AVs.
Section 1
Research
Our goals were to understand how to build trust, confidence, control and a sense of safety for the passengers riding Autonomous Vehicles.
Every year, 37,000 people die in car crashes in the United States. 94% of serious crashes are due to human error.
Risks of an AV
Uncertain responsibility in the case of an accident
Distrust and lack of understanding in automation and technology
AV technology is still imperfect and needs time to mature
Benefits of an AV
Help in decreasing road accidents in comparison to human-driven vehicles
Road traffic can be reduced with fully connected autonomous vehicles
Help in reducing road pollution & making the environment greener
Transparency of actions is key
“In the autonomous mode, the driver should have a sufficient understanding of what the autonomous vehicle perceives to realize its analyses and to make its decisions. The driver must be confident that the autonomous vehicle has the right information to make the right decisions and if not, he/she must be able to take control.”
What people expect
Information
Comfort
Safety
Emotional Recognition Technology in AVs
Multimodal Models are a combination of each of these methods to get a clearer picture of the emotional state of a driver.
Radio waves
Speech Recognition
Thermal Imaging
Facial Expression*
Eye Gaze*
After doing secondary research on different types of emotional recognition, we decided to focus on facial expression, eye gaze, as well as heart rate in our concept moving forward.
Competitors, or enablers
Who is already out there and what are they doing?
Level 3 Personal Vehicle
Steering Wheel and 15” touchscreen
Traffic-Aware Cruise Control
Camera, Sensors
Autopark, Auto Lane Change
Level 5 Fully driverless, Ride Share Service
Steering Wheel and Live Video Feed to see what Waymo sees
LiDAR, Camera, Radar
Live rider support agent
Level 5 Fully driverless, Ride Share Service
Origin - No driver seat
LiDAR, Camera, Radar
Level 5 Fully driverless, Ride Share Service
“Carriage style” car
Predetermined mapped routes
LiDAR, Camera, Radar
Primary Research Methods
The research methods we used to gain insight from both experts in the field, and the general population's opinion on the key features of our project.
Gain a first hand experience driving or riding a level 3 semi-autonomous vehicle.
Observe the how the user experience of the AV is tailored to the driver and passengers. Understand the different functions of the car UI.
7 User Interviews
Objectives
Inquire about different users' perspectives and attitudes toward AVs to understand various mental models.
Understand people's expectations with level 4 AVs and what it would take to feel confident in them.
Assess people's comfort levels with emotional recognition technology in the context of improving the AV riding experience.
Interviewees
7 Participants which included a long-time Tesla User, Tesla passengers, software engineer, industrial designer, tech and car enthusiasts.
3 Expert Interviews
Objectives
Learn about current technology, the fully-realised capabilities of AVs, and how confidence is built with the user.
Learn how difficult scenarios on the road are handled by AVs.
Learn if and how emotional recognition technology is being leveraged within AVs already.
Interviewees
Product Manager at Cruise*
Software Engineer at ZOOX*
Professor Ascanio (Professor at SCAD, ex-designer of Car UI, and Automotive enthusiast)
*Cruise and Zoox are both leading Autonomous Vehicle companies
Cultural Probe at Forsyth park
“Is Self-Driving the Future?” (N = 32)
Objectives
Outreach to a broader pool of people to learn about their views on AVs.
Inquire about people's comfort level with emotional recognition technology in cars.
Learn about ideal car riding experiences.
Study details
2 hours at Forsyth Park
32 participants in total
A mix of qualitative and quantitative questions
Cultural Probe Results
Synthesised Research Insights
After all the data points from both secondary and primary research were collected, we affinitised the data to come up with these key insights and matrix.
Need for Control in AVs
Drivers want to feel in control and responsible for the actions of the car, especially in high risk situations
Riders want the ability to override the AV to take back control in difficult scenarios
Riders want “just enough” transparent communication from their AV to compensate for the lack of control of the vehicle
Fear Regarding Autonomous Vehicles
Users have concerns about edge cases that may throw off the AV
There is considerable technophobia that a computer could never be on the same level as a human brain
Users are aware based on prior experience that software and technology can be fallible at times
Driving Behaviour, Styles and Pleasure
Some users want AVs to emulate human driving style and retain the same judgement calls as they do (passive vs. aggressive)
People enjoy the sensory aspects of driving (feeling the car run, speed, views)
Results in a unique set of emotions that drivers experience with their vehicles
How Might We?
We used How Might We questions to frame the design challenge.
How Might We Foster users’ sense of control and inspire confidence with an AVs ability to handle any scenario it may encounter?
How Might We Overcome users' initial apprehension and conditioned fears concerning AVs?
How Might We Utilise our technology both non-invasively and consensually so it doesn’t feel creepy?
How Might We Communicate the AVs actions and system state to users without overwhelming them with too much information?
Section 2
Design
By defining the core problems clearly and identifying our target audience, we shifted into ideating a concept that would cater to fulfilling their emotions and our business goals.
Concept 1
Test Drive Experience
A showroom test driving experience that focuses on building awareness and acceptance for AVs in the pre-purchase touchpoint.
The test drive experience simulates different scenarios with LCD screens on the windshield & windows. In the process users learn about the AV features and how the AV functions through the In-Vehicle Infotainment (IVI) System and Instrument Cluster.
Rationale
By exposing people to an immersive test drive, it provides them with the knowledge of how an AV works and how it is keeping them safe, reducing their fear and building confidence in the AVs.
Expected Outcome
Feature, Function, Task.
Concept 2
Redesigned AV Ecosystem
A showroom test driving experience that focuses on building awareness and acceptance for AVs in the pre-purchase touchpoint.
Focussing on the user experience of an AV, the proposal would include the key touch points that aid or directly give the user control of the vehicle. We would design how emotion recognition would create a feedback loop with the vehicle, the infotainment system and voice user interface to communicate, and the smart key fob.
Emotional Recognition Technology
In-Vehicle Infotainment (IVI)
Voice User Interface (VUI)
Smart Key Fob
Rationale
Users have a negative bias and fear towards AVs because of misinformation and lack of knowledge. Users feel unsettled and anxious in an AV due to the lack of control and human presence in the car. This can also be an isolating experience that induces further anxiety in a user.
Interaction Model
Feature, Function, Task.
While both concepts emerged to be strong ideas, we primarily chose to move forward with concept 1 as the base, then to push the innovation of it as far as our timeline and technical skill and understanding would allow us. Our next step was to build out the scenarios, task flows, screens and visuals for the test drive route and experience.
User Scenarios for the test drive
Scenario 1
User goes through the AV settings and customizes the display layouts, emotional recognition and AV driving style to match their preferences.
Scenario 2
User interacts with the IVI, navigates to a destination and the AV helps the user switch lanes when there is a construction site on the road.
Scenario 3
User interacts with the IVI, navigates to a destination and the AV helps the user pull over for an emergency vehicle passing by.
Task Flow Scenario 1
Set Up
Task Flow Scenario 2
Construction Site
Revised Interaction Model
User Testing Overview
We developed our concept and began creating the design artefacts to test the ideas with.
Objectives
Testing onboarding interactions and user comprehension of copy and technical settings.
Testing homepage layouts, placement of information and labels.
General questions on written copy, setup, Emotion Recognition Technology, homepage, iconography and alerts/notifications.
Tasks
Users would start off the onboarding process, and then proceed to set up their driver profile, customizing screen layout, ERT preferences and AV performance settings.
Users would then navigate to the Homepage, set a location and begin experiencing the IVI interactions.
User Testing Round 1
7 Participants
Closed & Open Card sorting
Results
AV current speed, autobrake, seatbelt indicator, transmission mode and hazard lights are needed behind the steering wheel display
Users want time, temperature & current location on both displays
Users prefer to have seat adjustment, windshield wipers, blinkers and lighting to be controlled via tactile buttons in the car
User Testing Round 2
5 Participants
Wizard of Oz
Results
Noted under the Lofi Wireframes below
Lo-Fi Wireframes - Set Up Screens
Set up process was easy to follow
Offer a layout with a bigger map when not in self-driving mode
Make driving style preferences more accessible through the landing page
Provide more context to how Emotion Recognition Technology is used within the AV
Differentiate between in-car (emotional recognition) and out-of-car (safety) notifications
Mid-Fi Wireframes - Set Up Screens in Light Mode
Appreciated different layout choices
Less text is good
Lo-Fi Wireframes - Navigation & Alerts
Users want the physical experience of the static tactile buttons
Navigation was easy to follow
Need more control over alerts and notifications
Notifications need to be more visible
Mid-Fi Wireframes - Navigation & Alerts in Light Mode
Visual elements feedback
Some of the copy and terms aren’t clear.
Some features need better explanation, suggested through visuals to easily absorb info.
Too much copy on certain screens, it is hard to digest.
Some greys are too light, not enough contrast.
Final Site Map
After further evaluation by users and an expert, Prof. Ascanio Paliano, we were able to define the site map of the car IVI center display, behind the wheel display, and tactile buttons.
User Testing Round 3
6 Participants
Wizard of Oz
Results
Driving
Driving felt natural and very realistic
The speed felt natural and speedometer readings felt accurate
Braking needs to be more powerful
Environment
Users got lost a few times and didn’t know where to go
Environment felt very bland and empty
Add more relatable environments like city areas
Add signage for directions
Scenarios
Construction scenario works well to make users feel more confident an the AV’s capabilities
Add more scenarios so that people can get a feel for the range of the AV’s capabilities
Section 3
Final Pitch
The full test drive experience with Revolution
Onboarding set up
Preference set up
Home Page and Navigation
Navigation enroute
Emotional Recognition and Notifications
Behind the wheel display
Default Instrument Cluster
Set Speed and Speed Limit
Navigation and Construction Alerts
Alternate Screens on Toggle
Route planning and scenario renders
To view a more comprehensive process of the physical model and the code for the unity set up, the Rookies Award page goes into detail from my team member's work on the fantastic test drive experience.
