top of page
Trust & Safety in Autonomy
It is clear that we will be surrounded by autonomous vehicles. There are many reasons why, hence we dive into those and solve for specific use cases to leverage contextual interactions to increase the trust while keeping human safety in mind.
Iris Autonomous Ride-Hailing
We conducted exploratory research on Autonomous Vehicles and design implications of trust and safety. We interviewed nearly a dozen experts working in the industry and conducted user research and testing with our target audiences - to understand the AV space, rider pain points, needs and safety requirements along the ride. This led us to strategize on ride-hailing of AVs and designing for multiple use cases for the most equitable access and utilization.
Location
San Francisco, CA, USA
Tools
Figma & FigJam
Adobe Photoshop
Rhino 3D & KeyShot 11
Phase I
Exploratory Scenario Research
There are more cars on the road today than ever before, more than the road systems in urban areas can manage.
Public transportation is not adequate and people rely on personal vehicles.
A lot of other factors, including cost, accessibility, safety and productivity.
Majority of the accidents are chalked up to human error. Top causes of crashes, ordered by number of fatalities:
Setting Stage
Estimated
42,915
people died in motor vehicle traffic crashes last year in USA.*
An increase of over
18%
in less than 2 years, 36,096 deaths in 2019.*
Estimated
$277 billion
estimated annual cost of motor vehicle crashes.*
Of the 6.5 million reported accidents
98%
were caused by human error in USA, 2021.*
*Source: NHTSA
Lets go through the journey of a ride, designed for a person with mobility issues.
Onboard
Increasing awareness by showing technology related information & instilling trust.
Setup
Booking the ride- hailing or rental and scheduled or recurring options.
The are trust and safety issues won't go away completely.
But it is imperative to solve for them as ultimately the autonomous vehicle system will be much safer.
Next Steps
Reflection
-
More research & interviews with people facing commute and mobility issues.
-
Work on prototypes working around other accessibilities e.g vision impairment.
-
Explore emergency scenario outcomes in depth.
-
Think about vulnerable population of the society for financial equity and balance with business goals.
-
Testing with users for feedback.
-
Detail the child's onboarding process.
-
Learn more about the tech stack and experiment, prototype and test these experiences.
1
Desk research & Competitive Analysis
Primary Research & Immersion
Analogous Research
Affinity & Journey Mapping
Exploration
2
Scenario Evocation
Archetypes, User Personas
Product Journey & Prioritization
Experience Principles
Strategy
3
Meet Iris - System Design
Ride 1: Busy Parent
Ride 2: Mobility Access
Reflections & Next Steps
Prototyping
THE PROCESS
Due to the speculative nature of the design space, we looked through a broad scope at a wicked problem.
THE CHALLENGE
The vastness of the challenge is daunting. The challenge autonomous vehicles face can be broken down into thinking this way - human lives, interacting with and within a smart metal container full of complicated futuristic tech stack, functionally moving in a dynamic and complex environment of the chaotic unsafe road system.
THE PURPOSE
There are tangible and intangible interactions to solve for - internally riders with vehicle and vice versa, and vehicle with external stakeholders and systems. These intricacies are worth solving for - considering the scale at which the wicked problem is being solved at. Speculating and solving for these pain points in interactions are for the riders would be imperative for the adoption and growth of AVs in the world.
HOW MIGHT WE
-
Create human interactions with autonomous vehicles
SO THAT
-
it increases trust and safety while sharing contextual information
THIS IMPACTS
-
vehicle passengers
-
autonomous vehicle OEMs
-
other cars and pedestrians
-
road safety & traffic policy makers
Capability deficit: When the vehicle isn't capable but the rider trusts the vehicle.
Trust deficit: When the vehicle is capable but the rider is still mistrusting and jumpy.
Users’ trust in the vehicle and evaluations of the vehicle reliability are often misaligned. Such misalignment may be mitigated by acting on user interactions in the HMI, contextual information and the vehicle’s behavioral dynamics.
Desk Research
Insights
-
Measured with Positive Risk Balance - adoption when autonomy is safer than a human driving a car.
-
Communicate values of the AV experience - visual, auditory and haptic. Communicate to passengers in both subtle and explicit ways.
-
Create a personalized experience.
-
Acknowledge user preferences.
-
Help passengers engage safely with their vehicle.
-
Assist passengers in anticipating what’s next.
-
Clearly message vehicle actions before they occur.
Breakdown of Trust
-
Predictability
-
Dependability
-
Faith
Trust Considerations
-
Accuracy
-
Reliability
-
Understandability
-
Message Framing
-
Message Coordination
-
Familiarity
Primary Research - Qualitative Interviews
Key Takeaways
-
Safety - a lot of AV OEMs are very close to a disengagement rate that would make them safer than manually driven vehicles.
-
Trust is subjective - to person and to situation, long building process.
-
There would definitely be an increase in accessibility in mobility for persons with disabilities.
-
Interactions would be of 2 types - inside and outside the vehicle. External include intangible forms of communication were used to.
-
There are hundreds of scenarios to design and solve for - hence the delay in release and trust deficit in adoption.
-
Humans would still need to feel control despite giving up key operations.
-
Accessibility - Recognize the need to make this technology available to the disabled ASAP, in order to allow their emancipation.
-
Need to reduce learning curve & make people aware.
10 Industry Expert Interviews
1 Hour (avg.)
UI Designer (Tesla)
Internal Designer Experiences Lead (Waymo)
UX Designer (Waymo)
Experience Designer (Cruise)
Industrial Designer (Aurora)
UX Designer (Zoox)
Prototyping & Interaction Designer (ex Nio)
Experience Designer (ex Uber ATG)
Car Designer (Kia)
AV Racing Club (CMU Autonomous Racing Club)
Topics of conversation
"HUMAN LIVES, WITHIN A COMPLICATED SYSTEM, IN TURN WITHIN A COMPLEX & DYNAMIC ENVIRONMENT"
Primary Research - Quantitative Survey
31 responses - 14 Questions
Primary Research - Immersion
Our Mentor, Professor Neil Torrefiel was able to arrange for us to get a ride in a Waymo to experience Autonomous Vehicles first hand.
The most interesting part was the Product Lead for in-car experiences was testing the new features and safety affordances during our ride. It led us to learn how Waymo was iteratively prototyping incremental evolution of the interactions and features in the AVs.
Insights
Finding the patterns - Affinity Map
Human Behavior
Design for pseudo-control and personalized experiences in free time
Onboarding and Learning Curve
Awareness crucial for adoption and it's incremental over time
Business Models
Private, hailing, sharing, goods, delivery
Trust
Sincerity, competence, reliability, within tolerance of risk
Safety
Competence testing, physical affordances, support, control outcomes
Interactions
Internal with vehicle and external with stakeholders on the road
HMI & Communication
Transparency, intent, utility, multi-modal
Standardization
Homogeneous experiences and expectations
Accessibility
Possibilities to improve mobility and caregiving, and be very equitable
Progress - Miles Tested
Competitive Teardown
Competitive Analysis 2X2
Archetypes
We started looking at who we are designing for, and some archetypes stood out as the the problems solved for them are more inclusive and impactful.
The Commuter and the disabled are use cases which would solve for a lot of edge cases as well as structure a lot of the general uses.
Stakeholder Mapping
Industry Trends
Residual, Dominant & Emergent Industry Trends
Key Takeaway
The important aspect to look at here is that the Autonomous Vehicles are at the cusp of emergent trends, while being backed by industry dominant tech companies and car manufacturers. Other tends in that zone are mobility access, smart HMI and evolution of road systems using networked communication. These are indicators for where the business investments and scientific innovation are headed in the present for development in next decade.
Scenarios
Scenario Evocation
Storyboarding
We listed down the numerous scenarios along the ride worth solving for. At one point we listed over 150 scenarios across a dozen touch-points. We categorized them based on whether the interactions were internal and external to the vehicle.
We started looking at the story of a rider. How we could solve for the pain points of the riders through al the phases of the ride and empathize with the users.
Vehicle Strategy
Ride-Hailing
This wasn't a simple decision. We could be designing for privately owned vehicles but our advisor and industry experts broke down the logic for us - if there aren't human drivers then there could potentially be carparks full of autonomous vehicles waiting to service anywhere within minutes and they would be cheaper than owning a vehicle at scale as it's not dependent on paying humans for driving.
1. Accessibility
Based on our research we realized that most autonomous vehicle companies are currently either in prototype stage or focused on early adopters and have not started designing for accessibility as a core function of the autonomous vehicles.
Vehicle Behaviors
Business Model
User Strategy
2. Busy Parent
We see a surge of ridesharing services for parents to help them with mobility for their children. We realized this is a gap - potentially saving hours of travel time daily for the parent. And also, a parent using a service for their child's mobility would be the epitome of trust.
Personas
Key Takeaways
The vehicles aren't capable of door to door mobility, but that doesn't mean that there aren't issues to solve for while getting in and out, and during the ride to ease the life of people with disabilities.
Key Takeaway
The most important part of this journey is that a service designed to alleviate their pain points and helping them gain back their time would help them spend more time with family - the right way.
Experience Principles
How might we
Create an enhanced autonomous riding experience that is accessible
So that
it increases trust and confidence while sharing contextual control and information
Empowered
To augment the user in more effortless craft. Insights and Information bring confidence to users actions.
Convenience
To be available to all user types, at any time, any place to promote independence.
Productivity
Take control of your time. Work, play, or sleep.
Customize as per your preferences.
Safety
Action and reassurance before data. Ongoing support provided by the operator.
“MAKE YOUR LIFE EASIER WITH IRIS”
We identify and deliver moments that matter most on your travel journey to create utility, convenience, value and effortlessness.
Experience Principles
Phase III
Prototype - 2
Accessibility in a scheduled rental trip
Scenarios
Product Strategy
Service Map
We divided up the product experience based on the phases of the ride. Across these phases we looked at the numerous scenarios worth designing for with goals and insights in mind, while also thinking of outcomes based on the designed touch-points.
Board
Face detection and unlock for easy access and opening of Wheelchair ramp.
Identify
Directed Sound
AR Windshield
External lights
AR Wayfinder on app
Safety
Communicating onboarding safety precautions for that type of ride.
Book
Different services for different users.
Accessibility features and booking adoptions.
Personalize
Customizing vehicle behaviors, controls and interior modes.
Ride
Home screens housing ride status, infotainment, control options.
Intention
Communicating vehicle behaviors, intent and status in case of hiccups along the way.
Emergency
Emergency controls and protocols.
Ride End
End of trip safety and feedback options.
Setup
Getting the family's (especially the children's) details for ease of use in multiple cases and then looking at types of rides that can be booked.
Phase III
Prototype - 1
Unaccompanied child & family scheduled multi-stop ride
Now lets go through the journey of a ride, designed for a parent to help their unaccompanied child in transit.
Onboard
Increasing awareness by educating riders about the service features and the tech behind it.
Book
Different types of rides for different use cases.
Family options to add custom trip specifications like unaccompanied child, bike rack, etc.
Board
Different services for different users
Identify
Helping the child identify and board the vehicle.
Personalize
Customizing vehicle behaviors, controls and interior modes for parent to monitor remotely. Home screens housing ride status, infotainment, control options.
Board
Communicating onboarding safety precautions for that type of ride.
Emergency
Emergency controls and protocols for parents monitoring remotely.
Intention
Communicating vehicle behaviors, intent and status in case of hiccups along the way. AR windows to show information of landmarks around to intrigue the children.
Ride End
End of trip safety and feedback options.
Arrival
Ride arrival status with child mode options like child monitor.
Analogous Research
Phase II
Strategy
bottom of page