5 big milestones for autonomous cars to hit: when we think they'll happen

We've all read the stories with varying degrees of hope and trepidation, watched cool (and sometimes funny) videos from the DARPA Grand Challenge, and wondered whether this technology is the "jetpack" we were always promised. But more and more, it looks like autonomous vehicles are going to happen. Ford CEO Mark Fields, for one, believes we only have four years to wait for fully autonomous vehicles to hit the road - and that's more or less in line with Google's estimate of 2017 to 2020 for total road-readiness. It's almost anyone's guess as to when and in what capacity this revolutionary technology will be deployed.

What's stopping, say, Google from deploying a small fleet of driverless vehicles across the North American market right now?

What's interesting about that question is that while aggressive lobbying (mostly by Google) has made road testing autonomous vehicles legal in 4 US states, Tesla is already beta-testing certain autonomous features in its street-legal fleet. All it took was some over-the-air patching, and limited autonomy became available to certain Tesla drivers.

Although this version of the Tesla software requires drivers to keep their hands very close to the steering wheel - with some harrowing results if that warning is ignored - autonomous vehicles are already among us.

So what's left for the industry to overcome before the seemingly magical year of 2020 rolls around?

1) Standardizing the car-pedestrian interface

Sharing the road with conventional cars is enough of a challenge - in fact, Google claims that of 14 minor traffic accidents to date involving its autonomous fleet, none of them were the fault of the car or its driver. But sharing the road with pedestrians presents more unique obstacles. Everyone understands the driver to driver interface - brake lights and turn signals supplemented by the occasional gesture - and driver to pedestrian communication basically works the same way. But if everyone in the car is a passenger, how do you let pedestrians know that it's okay to cross? Google is making strides in that direction. We feel like this will be one of the first major challenges to be overcome - let's say early 2017 for workable prototypes, and production models later that year.

2) Overcoming legal challenges

This is a big one. One might even say the big one. Tesla's new beta patch puts it way ahead of regulators across the country, but sooner or later, challenges like this will cause some kind of consensus to emerge. Odds are that traffic codes will need to be substantially amended to accommodate this new class of vehicle. Our money is on the flat desert highways of Nevada to lead the way in terms of full-state licensure - Google earned the first license for an autonomous car there in 2012. They'll return for the victory lap sometime in 2018.

3) Getting insurance companies on board

Like any disruptive technology, autonomous cars have a serious public relations challenge to overcome. Ask anyone and you're sure to meet diehards who wouldn't dream of riding in one. Even passive systems like lane change assist and forward proximity sensors (let alone parallel parking systems) take a little getting used to. But autonomous vehicle advocates say that the key to reducing risk (and presumably, insurance costs) is reducing driver agency outside of emergencies - and defining the new role that drivers will play is going to fall to insurers as much as lawmakers.

But outside of self-contained, purpose-built roadways, and maybe private taxi companies, it's unlikely that 100% hands-off vehicles will make the 2020 cutoff. Expect liability issues to be ironed out for 95% autonomous vehicles by early 2019.

4) Expanding coverage outside of well-mapped/flat areas

It's one thing to navigate through the maze of one-way streets and poorly parked cars that make up modern cities, and it'll be remarkable if the first generation of autonomous vehicles can even do that. But it's something else entirely to drive along a rocky mountain side road, or follow the sometimes inconsistent directions of an old parking garage. Expect early autonomous cars to have certain geographical limitations, and require drivers to take control near the end of some trips or in certain terrain conditions. Full mastery of the road, if it ever develops, won't happen until sometime after 2020.

5) Making autonomous cars cool

Google's earliest autonomous prototype was a diminutive Smart-like car, and from there it was a Prius. Many drivers' enthusiasm about hybrid savings aside, Tesla's interest in autonomous cars is a big shot in the arm for the technology's image.

But rest assured that it won't be safety features or insurance discounts that get most drivers into a car that requires less driving - the same features that attract drivers today to their favorite vehicles (power, fuel efficiency, appearance, et cetera) will always guide automotive purchasing decisions. Only time will tell what designs will conquer the market!