All New Tesla Cars Now Equipped to Achieve Full Autonomy—Someday

Elon Musk says vehicles in production will have full self-driving capability by 2018. But there are roadblocks in the way.

Tesla announced Wednesday evening that all cars currently in production -- including the Model S, Model X and soon-be-released Model 3 -- will be equipped with full self-driving hardware.

These cars will soon be able to check their owner’s calendar, find out where they want to go, plan a route and drive to the assumed destination, with zero human input. At least that’s the aim.

Tesla’s existing Autopilot technology offers semi-autonomous driving, where the person behind the wheel still has to keep an eye on the road. The company bills the technology as advanced Level 2 automation, based on the National Highway Traffic Safety Administration and Society of Automotive Engineers’ recently unified standards. CEO Elon Musk said that all Tesla cars built from now on will be able to reach Level 5 autonomy -- the highest level, where the vehicle can cope with all situations automatically during an entire journey.

Early Thursday morning, Musk tweeted a video of a Tesla car navigating a complex urban environment. The driver has no hands on the steering wheel. Musk added that Tesla’s electric vehicles will be able to recharge on their own using the automated “snake” charger.

An increasing number of automakers are making cars with some degree of autonomous functionality. Several manufactures, including Ford Motor Co. and BMW, have pledged to release fully autonomous cars in the 2021 timeframe. Google currently operates a small fleet of fully autonomous cars, but is still in the testing phase after years of technology development. Tesla’s plan to bring autonomous cars to market puts it miles ahead of the competition.

But there are some roadblocks standing in the company's way.

For one thing, while Tesla vehicles are now being outfitted with fancy hardware, the software component of Tesla’s self-driving technology doesn’t exist and has yet to be approved by regulators.

“It will take us some time into the future to complete validation of the software and to get the required regulatory approval, but the important thing is that the foundation is laid for the cars to be fully autonomous at a safety level we believe to be at least twice that of a person, maybe better,” Musk said on a call with reporters.

Vehicles equipped with new hardware will also initially lack some of the autonomous capabilities Tesla vehicles have with the first-generation Autopilot system -- including automatic emergency braking, collision warning, lane holding and active cruise control.

Those features will be validated and installed over-the-air through Tesla’s software updates. Musk said new cars should reach the same level of autonomy the Autopilot system currently offers by early next year, and will continue to be updated with new features every few months. Following that timeline, vehicles under production today should reach full autonomy by 2018.

Autonomy will come at a price, though. The purchase of a new Tesla car will now come with the option of Enhanced Autopilot or Full Self-Driving Capability for an additional $5,000 or $8,000, respectively. The earlier Autopilot system will no longer be offered on new cars and existing vehicles without the new hardware will not be able to achieve full autonomy.

The safety debate

Reaching the highest level of self-driving capability is an incredibly complex undertaking. Highway driving at consistent speeds is one thing; driving in cities and on smaller roads with more hazards is another. Google, for instance, filed a patent describing how to navigate odd circumstances like getting around a herd of cows. Self-driving technology needs to be prepared to handle any type of situation, no matter how random or improbable.

There’s also an ongoing debate about the best technology mix needed to achieve full autonomy. Specifically, Google’s self-driving vehicles use lidar -- a laser-based functionality. Tesla’s Musk believes lidar is “unnecessary” -- a position that spurred debate after Model S driver Joshua Brown died in the first fatal self-driving car accident in May.

Tesla’s new hardware suite includes eight surround cameras that provide 360-degree visibility; 12 updated ultrasonic sensors, allowing for detection of both hard and soft objects at nearly twice the distance of the prior system; and a forward-facing radar with enhanced processing that can see through rain and fog. The cars are also equipped with a new onboard computer with 40 times more computing power than the previous version. 

Electrek reports this is a notable shift from the hardware products Tesla was previously working on -- now with fewer radar antennas and a greater focus on cameras to take advantage of its new "Tesla Vision” product.

Tesla said the May accident occurred because the glare of the sun interfered with the car’s cameras, so it couldn’t see the truck turning left in front of the Model S. The incident led to a falling out with Israeli autonomous vehicle component maker Mobileye, which told Reuters it thought Tesla was “pushing the envelope in terms of safety” with its Autopilot system.

In September, Tesla announced that the eighth version of its Autopilot firmware would rely more heavily on radar than it has before. Wednesday’s announcement puts Tesla’s technology focus back on cameras.

Can't lose the public's trust

Robbie Diamond, president and CEO of Secure America’s Future Energy, a Washington, D.C.-based nonprofit that runs an autonomous vehicles task force, warned that technology needs to be deployed wisely in order to succeed. He called for striking a balance between adding public safety measures and preventing the implementation of onerous regulations that will stymie innovation.

“It's encouraging that automakers and technology companies are advancing autonomous vehicle technology at such a rapid rate, because it promises to save tens of thousands of lives every year, create mobility access for millions, and sever our dependence on oil,” Diamond said. “However, this dizzying pace of technological advancement creates risk if the technology is deployed recklessly, because the worst outcome is to lose the trust of the public.”

In September, NHTSA released guidelines for federal policy regulating autonomous vehicles. The guidelines are also expected to inform states and help prevent a patchwork of policies cropping up around the country. In California, for instance, where 18 companies including Google and Tesla hold permits to test autonomous cars, the Department of Motor Vehicles is crafting additional regulations that companies fear could slow technology development.

Studies show self-driving cars could eliminate 90 percent of all vehicle accidents in the U.S. Governments and the people they represent need to be convinced this is true in the real world.

"Technological capabilities are improving rapidly, but a very real roadblock on the path to autonomous vehicle deployment is consumer acceptance and trust,” said Avery Ash, autonomous vehicle market strategist at the transport analytics company INRIX. “These vehicles bring potential for tremendous safety gains for consumers, but that will only be widely accepted with clear evidence that these systems consistently outperform human drivers.”

A report by AAA published earlier this year found that three-quarters of drivers are afraid to ride in a self-driving car. But the same study found that drivers who own vehicles equipped with semi-autonomous features are 75 percent more likely to trust the technology than those who don’t, showing that more experience with the technology builds trust.



Still, there’s no guarantee consumer confidence and policy will keep pace with Musk’s plan to launch a fleet of fully autonomous vehicles in just over a year's time.