President Obama reently announced plans to earmark a whopping $US4 billion for autonomous vehicle research. These funds will be dispersed to pilot programs all over the US during the next decade -- but where and how the money is spent will determine just how big a step forward Obama's plan really is. Autonomous cars have driven millions of kilometres, but most of these kilometres have simply been devoted to improving the technology enough to get the cars street-ready. Obama's funding plan is a chance for the US Department of Transportation (USDOT) to allocate funds to programs with very specific goals that will help shape the future of transportation.
But there's another important part: This is a way to create highly visible pilot programs which can help expose as many people as possible to autonomous tech -- and help illustrate the many benefits of self-driving vehicles to a potentially sceptical public. So what should the $US4 billion fund? We've got some ideas.
Build more self-driving buses
Forget about single-passenger cars completely for a second. The most important autonomous technology is the kind that can transport multiple people at a time, coordinate with existing transit connections and help get more cars off the road.
Self-driving buses are in so many European cities already
The first self-driving buses on US soil will be deployed at a Northern California office park later this year. The USDOT announced a partnership with Mobileye, the Jerusalem-based startup that powers most of the autonomous buses worldwide. There are plenty of cities in Europe that already have these things on the road. Let's see an autonomous public transit prototype for a US city in 2017.
Promote autonomous trucking, shipping and deliveries
Here's the other big idea that's just not being championed enough: Why are we still letting humans handle shipping?
Future Truck is a giant Mercedes that can drive itself
Much has been made about Daimler's self-driving semi-trailers, but there needs to be just as much attention paid to tiny self-driving carts that can make local deliveries in a tight urban grid. Perhaps a fleet of self-driving delivery trucks to run in a separated-grade track around cities. It's a lot more realistic than drones.
Champion tech that's based on machine vision
Most of the self-driving tech out there right now uses a combination of cameras, lasers and GPS -- which requires rigorous training plus constant hardware and software upgrades. But there's a different autonomous vehicle that relies on machine vision and artificial intelligence -- basically a robot with environmental awareness. Carnegie Mellon roboticist Hans Moravec created this "evidence grid" technology in the 1990s, and a company named Seegrid paired it with real-time stereo-perception to guide autonomous forklifts and light trucks around warehouses.
Seegrid's Vision Guided Vehicle (VGV) tech moves heavy materials in warehouses
Seegrid's CTO Mitchell Weiss told me that their technology offers a more flexible autonomous system, and already has a proven track record when it comes to operating in close proximity with humans. Seegrid has performed tests in open environments in partnership with the military -- now it's time to get this alternate path to autonomy on the streets.
Build a city without parking
When autonomy arrives, cities will need about 80 per cent less space devoted to cars. Mostly because they don't need as many places to park. In the new Mother Jones, Clive Thompson offers a stunning vision of an autonomous future, where neighbourhoods have handed some of that auto infrastructure back to the people, and erased the arcane parking minimums imposed by city planners.
SFMTA's smart parking shows a pre-autonomy future with no circling cars, ever
What will this look like? Thompson uses San Francisco's high-tech parking meters as one example -- no more cars circling looking for vacant spots, meaning less vehicular congestion and dramatically decreased emissions. Cities need to have a benchmark like this. One of the pilot programs needs to focus on how exactly a bustling urban district would convert itself into this post-parking reality. Cities like LA, which have about 3.3 spaces for every car right now, will be almost unrecognisable.
Help on-demand ride apps eliminate car ownership
We know Uber has big autonomous tech plans. And Lyft is working with GM to produce a new type of autonomous vehicle. But what this really means is that "rideshare" apps will eliminate drivers -- and no one will need to own the car anymore.
GM's Mary Barra announces the partnership with Lyft at CES
The future will see fleets of dedicated Uber cars -- or Waze-navigated cars -- that you can summon. And the navigation systems will be smart enough to pick the most efficient routes and minimise travel times. This is one of the most important pilot programs to get right, because if people feel confident in this kind of system, autonomy is an easy sell. Which means car ownership (and the number of cars on the road) will go down.
Launch the self-driving car road show
One of the biggest problems around autonomous tech is the secrecy. This mysterious vibe has been perpetuated by companies like Google, which only recently started handing over data about crashes, for example. Transparency will let people see what self-driving tech is all about. We need to create a huge interactive experience that allows people to poke around the technology themselves. And who better to do this but Google?
It's time for Google to take a prototype of its car out for public inspection
Google needs to not only take their cars to the people, it needs to let the average person feel what it's like behind the wheel. Perhaps there's a detailed explanation of the safety testing in the form of a giant interactive exhibit. Maybe there's a Google Cardboard VR experience. And when it comes to all that data that these companies have been so recalcitrant about giving to cities, let's just say this: For any entity trying to get funding from USDOT, you must hand over your data. No exceptions.
Focus on safety innovations
Safety, safety, safety, safety, safety. Notice that I've hardly mentioned any traditional car manufacturers so far. That's because the only benefit to consumers that any automaker should be talking about in the context of autonomy is safety. If any automaker gets any of this money for single passenger vehicles, it can only be in the name of safety.
Volvo's concept was created as part of its Vision Zero initiative to end all traffic deaths
Volvo's self-driving concept car is an example of how automakers can bridge the gap between luxury and safety. Until this recent announcement, Volvo was forced to do its testing and development in Sweden. This is true for many automakers. But now that the US National Highway Transportation Safety Administration can make exemptions for safety-focused innovations, these companies will hopefully have an easier time bring us new ideas.
Illustration by Sam Woolley