Waymo x Raleigh
Building public trust in self-driving cars
Driverless cars are about to enter our already chaotic, information-packed roads. How we handle communication between these vehicles and the humans around them during this transition period, when the technology is new, unfamiliar, and untrusted, may well determine the success of automated vehicles.
Our team has been developing and prototyping an open-source universal communication library for self-driving cars to use when interacting with humans. The system we propose will make these vehicles safer and more transparent in their thinking, which in turn makes their behavior more predictable and ultimately more trusted.
We are pairing this with a rollout strategy and communications plan for an effective introduction to a major urban market. The Raleigh-Durham research triangle offers a perfect test market as a tech-forward city with congestion issues and a multitude of diverse audiences.
The body language of a human driver creates predictability and trust on the road. How can it be effectively replaced in a self-driving car?
An Open-Source Communication System for Automated Vehicles
Symbols are displayed when the car is communicating an action to those around it. These symbols are agnostic to display format.
Building Public Trust and Symbol Familiarity
Despite what advertising tells you, travel is usually pretty sucky. Autonomous vehicles remove the stress from vehicular transportation and gives individuals agency over how they spend their time during automotive travel.
Inviting Public Participation with Smart 311 Integration
No matter how much testing is done prior to the public roll out, there will inevitably be occasions where self-driving cars or other elements of automated smart city infrastructure get confused and behave in ways that the public finds confusing or irritating.
Autonomous vehicles that behave unexpectedly should be recorded transparently and with the full involvement of the public.
To that end, Google can provide the backend for smart 311 apps to be deployed in its host cities. These apps, already appearing in many cities with smart infrastructure, allow citizens to receive live information from officials and to make reports on city damage, overflowing trash, and other headaches.
We propose including reports on self driving car confusion to this system. Because this app is ultimately managed by the host city, the data collected on self-driving cars is to be publicly available and shared with any AV manufacturers operating there.
Elements of gamification encourage engagement, with the reporter receiving badges for quality reports, and the city itself receiving a score for reports cleared.
We began our process by surveying the complicated landscape of companies involved in self-driving cars, and the variety of patents and concepts centered around car/pedestrian communication.
The sheer variety of communications systems threatens to cause confusion and frustration in the public, a major barrier to public trust.
Moreover, the most common proposed solution — large LED displays with text instructions — have behaved poorly in testing.
We began with the USDOT transportation guidelines as laid out in the Manual on Uniform Traffic Control Devices (2009, rev. 2012), with the hypothesis that similarity to familiar road symbols would aid understanding.
From here we developed a dynamic, modern version of these familiar two-dimensional visuals and began testing our assumptions.
Brit Kern - Experience Designer
Mitchell Moss - Copywriter
Colin O’Shea - Art Director
Zachary Vono - Experience Designer
Preparing for Humanless Driving
Our team continues to expand and test the symbol library to include more situations and perception indicators. We look forward to releasing an updated system incorporating guidelines for how to deploy the symbols across a variety of car makes and models, and to adding audio cues to increase accessibility.