Elon Musk has admitted developing autonomous cars is harder than he thought, as the timeline for the latest software update slips again.
Like most automakers, Tesla is working hard to crack autonomous driving, widely seen as the next big evolution for the auto industry. The company’s software has been criticized for being ‘easily tricked,’ and there have been several high-profile deaths involving Tesla’s Full Self-Driving (FSD).
When a user poked fun on Twitter over the latest software update missing its deadline, Musk responded, acknowledging the difficulties involved.
Musk’s admission underscores the challenges companies are facing in their effort to bring the auto industry into the future.
Pony.ai has announced that it has secured $400 million in funding from Toyota to help develop its autonomous driving tech.
Pony.ai claims it is “developing the safest and most reliable autonomous driving technology globally. Having accumulated millions of kilometers in autonomous road testing in the most complex scenarios, we have a solid foundation to deliver autonomous driving systems at scale.”
The investment comes as virtually every major automaker, and numerous technology companies, are pursing autonomous driving tech. So far, Toyota has kept most of its plans under wraps, especially compared to some of its rivals.
Based in both the U.S. and China, Pony.ai’s technology is designed to meet the challenges of roads and driving conditions in vastly different and unique circumstances. A key element to that approach is the company’s Perception module.
“Our Perception module combines the strengths of a heuristic approach and deep learning models to boost performance, while ensuring the safety and operational redundancy of our vehicles,” says the company’s website. “Performance capability is further enhanced by our multi-sensor fusion technology, which intelligently leverages the most reliable sensor data depending on different environmental or driving scenarios.”
Pony.ai’s technology should go a long way toward helping Toyota meet its autonomous driving goals—whatever they may be.
In an interview with Bloomberg Businessweek, Verizon CEO Hans Vestberg talked about Verizon’s recent cloud partnership with Amazon, as well as the transformative effects 5G will bring, especially when paired with cloud computing.
On Cloud Computer Partnership With Amazon
“It is extremely exciting…we spent almost one and a half years with Amazon to do this. So—just to understand what we’re doing with Amazon—we’re bringing the cloud service out to the edge, together with 5G, in order to give super low latency, enormous throughput for applications being developed by developers.”
“This cannot be done. This is the first time in the world where actually we have seen that partnership. Amazon couldn’t have done it by themselves, because they don’t have wireless 5G. Verizon couldn’t have done it by itself, because we are not in cloud service, we don’t have cloud software. The combination of us can create something that is so transformative that, today you basically as a developer you can click on our first 5G edge site in Chicago and start developing an application for 5G with low latency, enormous throughput.”
“Of course…we’re one site right now. Think of us when we have hundreds of them, maybe thousands of them, over time….We can then give 5G experiences of low latency….Autonomous cars, real-time AR/VR, artificial intelligence, all of that can be at the edge.”
“And we’re just seeing the start of it, so that’s why we’re so excited about this partnership and what we launched 3rd of December last year.”
On How 5G Differs From Previous Generations
“Remember, when the design of 5G was done, the idea was this is wireless technology for industries and society….It was of course thought that consumers would get the benefit, but from the beginning was: How can you take away all the cables in the world and have the same performance as you had with cable, being much more agile, having new ways of doing it? That was the idea.
“When I think about 5G, 4G has basically two capabilities: speed and throughput. The phone is better every time you get a new generation. In 5G, eight currencies: battery optimizations, low latency. I mean, just one of the currencies, today I can connect 100,000 devices per square kilometer, tomorrow I can do 1,000,000. There’s never going to be 1,000,000 people on a square kilometer, so it’s done for devices talking to devices, optimizing flows for industries.
“So where are we? I mean, the plan was actually to come out 2020. We came out 2018. I think we’re ahead of the game, but still, from a consumer market, we’re just now starting to massively come into it. As we have said, this year we’re going to launch 20 5G phones….We think that our 5G is so different from others, because the performance on our millimeter wave 5G is just extraordinary. Today I get 2 gigabit per second in my phone! If you have a 4G phone, which you probably have over there, you probably have 40 to 50 megabits per second on Verizon, which is the best network in the country. And here we’re getting 2 gig. You cannot even imagine how much faster that is.”
The world’s leading automotive alliance, Renault-Nissan-Mitsubishi, is spending a billion dollars over the next five years investing in technology-focused mobility startups. François Dossa, Head of Alliance Ventures, a strategic venture capital fund of the company, says that they want to invest money in the new technologies that will make the mobility of the future and the car of the future.
Our Sector is Suffering Not an Evolution, But a Revolution
We want to invest money in the new technologies that will make the mobility of the future and the car of the future. We take this very seriously because our sector is suffering not an evolution, but a revolution. We’ve been for years using a car always the same way, your two hands on the wheel and your view on the road, and this is going to change. Many new technologies that we will not develop internally because it would be too long and too expensive and this is why we have to rely on startups.
We are excited with everything related to mobility. In mobility, I can also think of my legs that will be a part of the mobility, so it will be multimodal, not only a car. If I want to go from A to B, I have to be able to use my legs, a bike, a scooter, a car, a train, a van, a bus, and all this together is going to be the new journey that our customer will have. Customers in the future will consider that we are a service company and this is the big change that we have to do. We are good at manufacturing cars and we have to be good at offering a service to our customers.
Should we have our own platform or should we use someone else’s platform? We decided to have our platform, so we’re going to build our platform and I’m in charge of that. On a platform are all the tech bricks that you need to have to offer, ride-hailing, ride pooling, ride sharing, and goods delivery to your customers.
My activity is not to acquire companies, it is to acquire shares in companies. I don’t want to be the owner. If I think of a start-up, if I acquire the startup I kill the startup. We are a company of 475,000 people, that’s a very big company, very kind of heavy company, and we need to keep the fact that in the startup they grow very fast. All these technologies they’re moving very fast. This is why I don’t want to acquire but I want to be part of the shareholder of the company.
What’s Happening in China is Just Amazing
China is the most amazing area in terms of innovation. Number one in the world. What’s happening in China is just amazing. We have been investing in a company called WeRide, it’s a level 4 autonomous startup in China. With level 4 I can tell you, I had a demo with their car in normal traffic conditions and the car goes on the highway and nobody is driving the car. My first reaction I was really dying. I was oh my god and we had 30 minutes and it worked extremely well.
That’s very clear and it’s something that I’d like to say. About 40 years ago the Japanese companies went to the States and to Europe, they copied what we were doing. Then ten years ago, the Chinese did this also. Now I think it’s our time to go to China to look at what they are doing and to copy, because if not they’re going to kill us. We have to be very cautious about China and there are a lot of opportunities especially in deep tech new technologies. With AI, the Chinese know exactly where they are going. They tell you well we have a plan, it’s the central government plan where they want to be in 5, 10, and 15 years.
AI is everywhere, so we have four main activities for investment. First one is the new mobility and this is the decision that we made about the platform. Then we have all related to autonomy, autonomous cars connectivity, and services. and EV, electric vehicles. We are the leader in EV and we want to keep this position, and we need to make a lot of investment and startups are very good at this.
The Key Driver is Really the Technology
Even though I’m not American, I think that money is very important. For me, money is important and it’s also a way to be sure that we invested in the right startups because they will grow and then we’ll make money out of this investment. But this is not the key driver. The key driver is really the technology. We are a strategic fund and so we are here to bring into the Alliance companies and the new technologies that will make this new mobility and this car of the future. That’s really really clear in our mind but we’ll make money also out of it.
About Alliance Ventures:
Alliance Ventures is a strategic venture capital fund of up to $1 billion over the next five years, operated by Renault-Nissan-Mitsubishi, the world’s leading automotive alliance.
The fund, launched in 2018, with a $200 million initial investment is co-located in Amsterdam, Silicon-Valley, Paris, Yokohama, Beijing and Tel Aviv from where it targets technology and business model innovation in New Mobility, Autonomous Driving, Connected Services, EV & Energy and Enterprise 2.0.
Alliance Ventures is the main interface with the Alliance and its member companies for start-ups, incubators, accelerators, investors and the venture capital eco-system. By drawing on expertise and business opportunities from across the world’s leading automotive alliance, we pursue strategic investments at all maturity stages in startups developing disruptive technologies or businesses, focusing mainly on Series A and B, plus follow-on investments to support the start-up’s growth.
Drive.AI has launched a free self-driving car service in Arlington, Texas. “In Arlington, we are launching three different services,” said Drive.AI CEO Bijit Halde. “One for the game day, one for lunchtime service, and one that connects the Convention Center to Texas Live, the entertainment part of the city.”
“Today, we started with three cars with the potential to expand as the need of the community grows,” noted Halde.
Halde says that “they will be the same type of cars we had in Frisco, Texas, part of the Dallas-Fort Worth area.” He noted that it’s just a beginning and that the small size of the launch is only one way of looking at it. “We want to be very careful and deliberate because safety is of utmost concern. Also, user acceptability is a core concern. We start small and grow fast.”
Riders will pay nothing Halde says and anyone in that area can access this service. Halde added, “In Frisco, we only had a mobile app. In Arlington, we have a mobile app and walk up kiosks where you can type in your name and phone number to get picked up.”
Halde explained that the company feels there are three factors to success with launching a self-driving ride-hailing service. One, can we safely deploy? But driving is not just one problem. The other factors are driving in the city and driving on the freeways. Drive.AI considers those as two distinct problems.
“We want to make sure that we take a small problem and solve it and then grow from there,” said Halde. “We don’t want to push the technology when the people aren’t ready.”
Driverless vehicles have the ability to literally change the world by making driving safer, more energy efficient, more accessible, and many will be happy to hear… eliminate congestion and gridlock. The government today made an important first step in truly making this possible.
“Today is an important moment at the Department of Transportation,” announced Anthony Foxx, US Secretary of Transportation. “We have issue record recalls, we still have too many people dying on our roadways and we have too many Moms and Dads stuck in traffic losing productive time with their families. In the 50 years of the Department of Transportation there has never been a moment like this.”
He added, “A moment where we can build a culture of safety as new transportation technology emerges that harnesses the potential to save even more lives and that will improve the quality of life for so many Americans. Today, we put forward the first Federal policy on automated vehicles. The most comprehensive national automated vehicle policy that the world has ever seen. It is a first of its kind.”
“It is taking us from the horseless carriage to the driverless car,” says Foxx. The policy is effective today, but the agency welcomes ongoing dialogue and will make changes as time goes on. “The focus on this technology will always be safety.”
The New Driverless Vehicle Policies
The new policies by the National Highway Traffic Safety Administration (NHTSA) will also let those “drive” without a drivers license, just like they do currently with Uber, Lyft and taxis. The government differentiates rules and regulations for cars requiring a driver and those that don’t.
If you were wondering, driverless cars will not have to have steering wheels or brake pedals. The agency says they have been charged with creating a path toward fully autonomous vehicles.
The 15 point assessment is designed to recognize that driverless vehicles are a rapidly changing and emerging technology. It does however, let the industry see a roadmap for how the government will deal with the regulatory environment for autonomous vehicles. Their goal is to build a safety culture now around autonomous vehicles, instead of as an afterthought.
The bottom line is that the NHTSA is extending its rulemaking authority to driverless vehicles.
Autonomous Vehicles Will End Drunk Driving
Also speaking during the announcement was the National President of Mothers Against Drunk Driving (MADD), Colleen Sheehey-Church, saying “over ten thousand people continue to die each year needlessly due to drunk driving.” She added, “A fully autonomous vehicle would stop a drunk drive simply because they can’t physically drive the vehicle.”
“I would also like to point out the driverless cars can do much more than simply stop drunk driving, these vehicles could potentially stop most of the traffic deaths in our country,” says Sheehey-Church. “A driverless car is not distracted, it ensures that the occupants are traveling at appropriate speeds and it would avoid pedestrians and bicyclists.”
“While improving safety, a driverless car would also create new mobility opportunities,” she said. “Older drivers who may be shut in or unable to drive may be able to drive at night again. Members of the disabled community who may not be able to drive could now have new opportunities for transportation like never before.”
“To that end, MADD is proud to support the new proposal on autonomous vehicles,” she said.
Watch the HAV Press Conference here:
Overview of Federal Automated Vehicles Policy
The Obama Administration today has released the first set of guidelines for fully autonomous vehicles called the Federal Automated Vehicles Policy. The 8 page policy release predicts a driverless car future that will create safer roads and many more energy efficient transportation options. Although the main focus of the new policy is about highly automated vehicles (HAVs), there are portions that also apply to lesser levels of automation such as the driver assist systems found in Tesla’s and other high end cars.
“We’re envisioning a future where you can take your hands off the wheel and the wheel out of the car, and where your commute becomes productive and restful, rather than frustrating and exhausting,” said Jeff Zients, who is Director of the National Economic Council and Assistant to the President for Economic Policy, in announcing the new policy.
The government sees autonomous vehicles as a way to leap current hurdles for the 4 million Americans who are living with a disability as well as older people who have difficulty seeing at night. They also view it as a way to make our society more fair and just, where vehicles are made assessable for all. They even believe that blind people will eventually be able to use driverless cars to get around, with innovative technology that will be developed to assist.
The policy guidelines which were developed over several years are a work in progress and will be updated annually with the goal of keeping the regulations up-to-date with the rapidly evolving technology.
Components of the Policy
Vehicle Performance Guidance for Automated Vehicles: The guidance for manufacturers, developers and other organizations outlines a 15 point “Safety Assessment” for the safe design, development, testing and deployment of automated vehicles.
Model State Policy: This section presents a clear distinction between Federal and State responsibilities for regulation of HAVs, and suggests recommended policy areas for states to consider with a goal of generating a consistent national framework for the testing and deployment of highly automated vehicles.
Current Regulatory Tools: This discussion outlines DOT’s current regulatory tools that can be used to accelerate the safe development of HAVs, such as interpreting current rules to allow for greater flexibility in design and providing limited exemptions to allow for testing of nontraditional vehicle designs in a more timely fashion.
Modern Regulatory Tools: This discussion identifies potential new regulatory tools and statutory authorities that may aid the safe and efficient deployment of new lifesaving technologies.
Vehicle Performance Guidance
The policy creates a 15-point Safety Assessment which outlines objectives on how to achieve a robust design. It allows for varied methodologies as long as the objective is met:
Operational Design Domain: How and where the HAV is supposed to function and operate;
Object and Event Detection and Response: Perception and response functionality of the HAV system;
Fall Back (Minimal Risk Condition): Response and robustness of the HAV upon system
failure;
Validation Methods: Testing, validation, and verification of an HAV system;
Registration and Certification: Registration and certification to NHTSA of an HAV system;
Data Recording and Sharing: HAV system data recording for information sharing,
knowledge building and for crash reconstruction purposes;
Post-Crash Behavior: Process for how an HAV should perform after a crash and how
automation functions can be restored;
Privacy: Privacy considerations and protections for users;
System Safety: Engineering safety practices to support reasonable system safety;
Vehicle Cybersecurity: Approaches to guard against vehicle hacking risks;
Human Machine Interface: Approaches for communicating information to the driver,
occupant and other road users;
Crashworthiness: Protection of occupants in crash situations;
Consumer Education and Training: Education and training requirements for users of
HAVs;
Ethical Considerations: How vehicles are programmed to address conflict dilemmas on
the road; and
Federal, State and Local Laws: How vehicles are programmed to comply with all
applicable traffic laws.
Model State Policy
The policy emphasizes that states will continue with their traditional responsibilities for vehicle licensing and registration, traffic laws and enforcement, and motor vehicle insurance and liability regimes while also carving out a new Federal role for autonomous vehicles. The goal is to not have states stepping all over themselves with a hodgepodge of rules, making it impossible for self-driving cars to drive between states.
The Federal responsibilities include setting safety standards and enforcing them, investigating safety issues and managing recalls, public education on driverless safety and communicating future guidance to the public in order to achieve national safety goals.
The Feds also created a regulatory framework model for states to follow in order to create a consistent approach to governing autonomous vehicles:
Application by manufacturers or other entities to test HAVs on public roads;
Jurisdictional permission to test;
Testing by the manufacturer or other entities;
Drivers of deployed vehicles;
Registration and titling of deployed vehicles;
Law enforcement considerations; and
Liability and insurance.
Current Regulatory Tools
Especially interesting is the governments forward looking approach in trying to make existing laws work to allow the use of driverless vehicles. This will be done via government agency reinterpretation of existing laws, using Letters of Interpretation, basically stretching them as far as they can go without changing their intent.
The policy is also going to use its current power to provide limited exemptions to vehicle manufactures to test new designs of cars that are not currently allowed. For instance, all cars must have a steering wheel, except that you don’t need one in a driverless car and it could even add danger because people could bump into it. Exemptions will allow manufacturers to bypass “buggy whip” rules that aren’t applicable in a vehicle that nobody is driving.
They have also created a path to more permanent ways to bypass old safety and design rules using a petition for rulemaking. This allows manufactures to adopt new standards, modify existing standards, or repeal an existing standard.
Modern Regulatory Tools
The new policy identifies new tools that could be created under current law while also laying the foundation for new laws requiring Congressional action. Within this section the policy is a first step toward reinventing laws and regulations of the world’s likely driverless future revolving around safety issues, software updates, regulation processes, record keeping and data sharing.
Data sharing is an area the self driving industry may not be too happy about. They are likely to focus their army of lobbyist on Congress to make sure they aren’t giving up their proprietary data that they have spent millions obtaining.
Google is almost ready to put its self-driving cars on public roads. According to the company, the “next step” in the project involves putting its prototypes on the roads of Mountain View this summer.
Google’s self-driving cars will be capped at 25mph, and each will carry a safety driver who has access to a removable steering wheel and pedals so that they can take control if they have to.
“When we started designing the world’s first fully self-driving vehicle, our goal was a vehicle that could shoulder the entire burden of driving. Vehicles that can take anyone from A to B at the push of a button could transform mobility for millions of people, whether by reducing the 94 percent of accidents caused by human error (PDF), reclaiming the billions of hours wasted in traffic, or bringing everyday destinations and new opportunities within reach of those who might otherwise be excluded by their inability to drive a car,” says Chris Urmson, Director of the Self-Driving project.
According to Google, its self-driving fleet has spent the equivalent of a human life test driving.
“We’ve been running the vehicles through rigorous testing at our test facilities, and ensuring our software and sensors work as they’re supposed to on this new vehicle. The new prototypes will drive with the same software that our existing fleet of self-driving Lexus RX450h SUVs uses. That fleet has logged nearly a million autonomous miles on the roads since we started the project, and recently has been self-driving about 10,000 miles a week. So the new prototypes already have lots of experience to draw on – in fact, it’s the equivalent of about 75 years of typical American adult driving experience,” says the company.
Earlier this week, a report claimed that in Google’s previous testing, four out of 48 self-driving cars were involved in accidents. Google confirmed this, but stated that none of those four were the fault of its automobiles – two accidents occurred when the cars were in control, and the others while humans were controlling them. According to Google, all of the accident were low impact, having occurred when its cars were traveling below 10mph.
Still, Consumer Watchdog called for the company to make its self-driving car accident reports public.
According to Google, its self-driving cars have logged over a million miles without ever causing an accident.
Will there ever come a day when it is actually illegal to get behind the wheel of a car and drive it yourself?
Tesla founder and possible Terminator Elon Musk says yes, this is a distinct possibility.
Musk made his comments at the Nvidia GPU conference, where he said it’s just too dangerous. “You don’t want a person driving a two-ton death machine,” he said.
Musk’s argument is self-driving cars will eventually be shown to be so much safer than actual, human driving that the public will have no choice but to outlaw the latter.
“[Self-driving cars are] going to become normal,” he said. “It’s like an elevator. We used to have elevator operators, and we developed some simple circuitry … The car is going to be just like that.”
“We’ll take autonomous cars for granted in a short period of time,” he said. “It’s going to be the default thing and it’s going to save a lot of lives.”
Of course, this is something that’s far from certain – and even if it happens it would be years and years from now. In the grand scheme of things, companies like Google and Musk’s Tesla are just beginning to focus on this technology. Google is admittedly far ahead of the curve, and its autonomous vehicles are getting very close to road readiness.
But Musk has said that Tesla will eventually be the leader in self-driving cars.
Likely taking some flak from auto enthusiasts, Musk walked back his comments a bit on Twitter – making it known that a driverless future is not really something he wants to be a part of.
To be clear, Tesla is strongly in favor of people being allowed to drive their cars and always will be. Hopefully, that is obvious.
Try using this excuse the next time you get pulled over for speeding.
According to Google, exceeding the speed limit is actually safer than strictly obeying the law in some circumstances. For this reason, the company’s autonomous vehicles are allowed to break the law, ever so slightly, if the situation warrants it.
Dmitri Dolgov, the lead software engineer on Google’s driverless car project, says that speeding is safer than sticking to the speed limit when all the cars around you are going much faster.
Of course. That’s called the little old lady driving on the interstate theorem.
Because of this, Dolgov says that Google’s driverless cars are programmed to go 10 mph over the speed limit “when traffic conditions warrant,” according to Reuters.
Google’s autonomous cars doing 55 in a 45 shouldn’t worry you. What could worry you is how Google’s autonomous cars will react when other cars, you know, driven by dumb, fallible humans, make mistakes. The time is fast approaching, too. Driverless cars should hit UK roads in January of next year. Google is already testing them all across Mountain View in the US, and is hoping to start operating them in other states (pending approval of course) soon.
Hearts and minds will need to be won, however, as about half of Americans are still a bit wary of robot cars on the road.