Air Traffic Controllers For AVs
I was reading today about Lyft setting up their own AV unit with a more
open twist – allowing automakers and tech companies to use their AV
vehicles, share data, and maybe even share technologies. They are
claiming that they are pursuing an open strategy as a way to bring safety
benefits and environmental benefits to market faster.
I frickin’ love this.
It reminds me of a presentation I gave to the Megatrends conference on
connected cars back in April. The presentation was “The Future” and I
claimed that in the future manual vehicles will be banned from public
roadways. Leaving off truck and freight movement (which I think is where
AV will be first commercialized), I talked about three primary business
cases – municipal fleets, robo-taxis, and chauffeured subscriptions.
Municipal fleets, in my talk, covered AV buses, while robo-taxis were
ride-hailing and ride sharing, urban vehicles. Chauffeured subscriptions
are for individuals subscribe to a particular brand of vehicle, like Audi
Silver Car or Book by Cadillac, and swap our vehicle types as needed. For
example, children getting to school could summon a shared van, a commuter
needing to have a conference on the way to work could request a private,
single ride, sedan, while a family on vacation to Tahoe would need a SUV
with 4-wheeled drive.
It was a dramatic way to get to the two, major type of platforms that
could develop when AVs are the majority:
1. Entertainment platform
2. Logistics platform
CB Insights along with Ernst and Young project incremental media industry
revenue of $20 billion. That doesn’t count the revenue for the plumbing
to deliver the content! The entertainment platform (ie. content plumbing)
has several needs depending on business case. A municipal fleet would
primarily need to keep passengers informed on the ride and maybe offer
enhanced connectivity and/or content that can be consumed on a private,
small screen. A robo-taxi, shared, or single rider, could also offer
content on a small screen, maybe enhanced connectivity and power, but will
also be primarily concerned with informing passengers about the route.
Chauffeured subscriptions could have radically different entertainment
needs. The family on the way to Tahoe will need not only information on
route and options for pit-stops, but also for the right type of
entertainment, likely delivered on a shared screen in the vehicle. Or
different content on different screens. The commuter having a meeting
will need to make sure sensitive, private information is not retained by
Some of the technology to power an entertainment platform include things
such as security, identity management, digital rights management, payment
gateways, and broadcast capabilities.
The platform for AV logistics is different. I once heard Laura Merling,
from Ford, describe it as Air Traffic Controllers for AVs. I love the
analogy for many, many reasons. When you think about it, the AV delivery
platform will need the ability to track inventory of vehicles available
for rides with positions and status of people requesting rides. Ideally,
it will be fully automated dispatching to manage ride requests and
vehicles to riders and (hopefully), ride sharing.
A little aside here about transportation networking companies and
ride-sharing misnomer. I was at the AV Summit in San Francisco and a
researcher was showing that TNCs are not ride sharing services at all.
They apparently suffer from dead-heading, which has nothing to do with
Jerry, but rather refers to the situation where a driver has no passengers
and is cruising around. Anyhoo, not only is dead-heading a problem, but
the vast majority who use Uber and Lyft are single riders. It would be
great if we could figure out how to ensure the majority of AVs are
actually shared. Maybe with better logistics and eliminating the creep
factor, we can get there. But now back to the main point.
The AV delivery platform should be able of incorporating with or part of
managing a municipal fleet of buses, coordinating rides for robo taxis, or
used by chauffer services. The requirements are similar with the platform
will needing to manage destinations, directions, trip tracking, vehicle
maintenance, and payment processing. It will need to communicate with
riders and passengers.
The technologies deployed on an AV delivery platform will need a control
client in the vehicle to obtain location, manage communications and
services delivery, mobile app will be helpful for sharing information with
riders and passengers, allowing people to summon a ride. The AV logistics
platform needs a services back end that could include services catalog.
Think of a services catalog as the set of microservices such as knowing
where the vehicle is located and where it needs to go. It will need a
transactions engine and customer care tools, an OTA manager will be key to
updating software in the vehicle remotely, adding new services, patching
security bugs, etc. Besides service back end, a core back end will be
necessary to effectively manage services delivery would include event and
notification managers, command and control components, the code to handle
authentications and authorizations. Of course, the AV logistics platform
will need ways to easily connect to automakers and OEM systems, and/or
municipal systems, ideally through API gateways, and other third party
systems, like entertainment services.
Part of Lyft’s motivation to open their vehicle distribution and logistics
network is to make a standardized system for AV to use. This is a huge
play and I’m not surprised in the least. I’m not sure a particular
automaker could be the AV air traffic controller. It feels like too much
of a public good or assuming government will not get involved, then a
neutral third party where all automakers can participate. It would not be
helpful if Ford, GM, and Toyota all had different systems for deciding
which vehicles go where, when, and how. Even if it starts out that way,
the economies of scale and public interest are so large, that I could
totally see it being, if not Federal like real air traffic control, then
at least state managed. Not that I’m sure that is a good idea!
It seems like Apple, Google, and large media companies will fight to the
death to control the majority of AV’s entertainment platform. You could
say that Uber and Lyft has a leg up on everyone for vehicle logistics, but
you never go if they have the vision, the culture, the money, and the
staying power for when the majority of the vehicles on the road are
I do like what I read about Lyft’s announcement, especially the virtuous
cycle of data gathering for making high-definition maps for the vehicles
to use and for the platform that run autonomous vehicles. And yet another
reason why economies of scale will rule this day.