Navigant Research Blog

There Are No Self-Driving Cars for Sale Yet

— October 25, 2017

Let’s be absolutely clear about something. As I write these words in October 2017, there are exactly zero self-driving vehicles available for consumers to purchase in America. In fact, Elon Musk’s proclamations and pre-sales of non-existent technology aside, it will likely be at least several more years before an individual can buy a self-driving vehicle. With that in mind, the media needs to stop using the term self-driving in the context of any production vehicle.

Misleading Headlines

In recent weeks, Cadillac has conducted a cross-country media preview of the 2018 CT6 sedan with the first production application of its Super Cruise system. From the event launch in Manhattan to its arrival in Los Angeles 2 weeks later, media outlets including NBC’s Today show, USA Today, Business Insider, and Fortune have referred to Super Cruise as self-driving. In doing so, they are doing a disservice to their own credibility, to consumers, to General Motors (GM), and to every engineer working on automated driving technology.

To its credit, GM itself never calls this self-driving or automated technology. Following the 2014 ignition switch recall, GM instituted new safety review procedures on new products and those changes are reflected in the capabilities and limitations of Super Cruise.

What Is the 2018 CT6 Super Cruise?

This is a very capable advanced driver assistance system (ADAS) similar in principle to Tesla Auto Pilot, Volvo Pilot Assist, and Mercedes-Benz Drive Pilot. Navigant Research’s Automated Driving Vehicle Technology report projects that these types of systems—defined as Level 2 partial automation by SAE—will account for nearly 59 million sales annually by 2026.

I personally spent nearly 900 miles with the system over 2 days. Within its operating domain, it works very well. But the key is that operating domain or definition of where the system can work. This is a supervised partially automated assisted driving system. On divided highways where there are no intersections, cyclists, or pedestrians, Super Cruise can handle steering, acceleration, and braking with the driver taking their hands and feet off the steering wheel and pedals.

That doesn’t make it automated. As the Super Cruise branding implies, this is a more advanced adaptive cruise control. The driver must still watch the road and be ready to take over when the system encounters a situation it cannot handle such as a construction zone, lane merge, or faded lane markings.

A face tracking camera similar in principle to the Face ID system on the upcoming Apple iPhone X watches for facial and eye movements to ensure the driver is alert and attentive, something no other current ADAS does. Meanwhile, high definition navigation maps prevent inappropriate use on surface streets.

Misrepresentation Leads to Unrealistic Expectations

By continuing to call Super Cruise self-driving, media creates unjustified expectations of its capabilities with consumers. GM’s design approach should reduce the sort of misadventures we’ve seen from Tesla drivers on YouTube. However, customers that bought into the media hyperbole may be disappointed with the more cautious assisted driving technologies. Even as the technologies become more sophisticated, customers burned by misleading headlines today may remain skeptical and decide to hold off on future purchases, which damages the overall goal of improving safety on the road.

Patience

It’s one thing to take someone like Elon Musk at his word and assume he’s going to deliver what he promises (which he often eventually does, albeit over budget and months to years late). But by inaccurately portraying what a product can and more importantly cannot do, customer interest (and possibly safety) can be compromised. Truly automated vehicles will get here soon enough. Let’s not rush to mislabel.

 

With Self-Driving Cars, We’re All Cartographers

— December 5, 2016

Connected VehiclesMapmaking used to be the domain of a select group of cartographers that would gather, review, and plot out data onto sheets of paper. The chances that you actually knew a cartographer in the past were probably pretty slim—but not anymore. Today and in the future, virtually everyone is or will be a contributor to the increasingly detailed maps that represent the world we live in.

As our vehicles become increasingly automated, they need ever more detailed maps, and not just the maps we get from Google or Apple on our smartphones. The self-driving car will need much more information. The basics of street names, directions, and building numbers are just the beginning determining a basic route from where a car is to where its user has asked it to go. This data set already exists in every vehicle with a navigation system and a GPS receiver.

Limits of GPS

However, if you’ve ever tried to navigate around urban canyons in places like Manhattan or Chicago, you’ve no doubt experienced the limitations of GPS as the signals orbiting more than 12,000 miles above the Earth’s surface bounce between skyscrapers. Looking at the navigation display and realizing that the car thinks it is several city blocks away from your actual location is not exactly confidence-inspiring.

Even when it works correctly, GPS is only accurate to several feet, not nearly precise enough to safely locate where a car is on the road. Then there’s the problem of navigating around on streets when you can’t actually see the road, such as when it snows. If you can’t rely on GPS for precise positioning and you can’t see lane markers, you need other data to calculate location.

Crowdsourced Maps

That’s where the future of crowdsourced mapping comes in. If you use smartphone-based navigation apps like Waze, Here, TomTom, or Google or Apple maps, you are already contributing to augmenting the map data that is also collected by fleets of sensor-equipped vehicles that drive the world’s roads.

In the near future, the cameras and other sensors that power lane keeping systems and other driver assist features will be feeding information to datacenters where it is aggregated with information from other drivers. In addition to real-time traffic and road conditions, they will be looking for landmarks like bridges, signs, buildings and more, and anything that isn’t already in the high-definition map will be uploaded.

Mobileye is the leading maker of image processing and recognition systems used by automakers for driver assist. In January 2016, the company announced a new product called Road Experience Management that processes images captured by car cameras and sorts out new information. This data is then transmitted and collected in order to update maps. Earlier this year, Ford invested in a startup called Civil Maps that is developing a similar system using cameras and any other sensors on the vehicle that can provide relevant data.

Even when the vehicle sensors can’t see the road, if they can see landmarks, they can triangulate and calculate position to within a few inches. Last winter, Ford demonstrated the ability to do precisely this with its autonomous prototype using a high-definition map generated using LIDAR. The future ability of autonomous vehicles to successfully operate in varied conditions will depend in large part on the contributions that we all make toward improving the quality of maps.

 

Automakers Doing More Rigorous Safety Analysis for Vehicle Automation

— November 23, 2016

Connected VehiclesBack in September 2014 as the ITS World Congress gathered in Detroit, General Motors (GM) CEO Mary Barra announced that in 2016, a new Cadillac model would become available with the semi-autonomous Super Cruise system. With only a handful of weeks left in 2016, we now know that the Super Cruise will debut on Cadillac’s flagship CT6 sedan, but it won’t be arriving until sometime in 2017.

A lot has happened since that announcement, and GM has put a much greater emphasis on ensuring safety as a result of the massive ignition switch recall that began early in 2014. Those process changes have led to some significant upgrades to Super Cruise in an effort to avoid the issues caused by human interactions with Tesla’s similar AutoPilot driver assist system. Navigant Research’s Autonomous Vehicles report projects that by 2020, approximately 13 million vehicles with these so-called Level 2 automation systems will be sold annually.

Geofencing

In the process of evaluating the safety of Super Cruise, one of the key differences that GM has implemented is geofencing. Since Super Cruise is designed primarily as an advanced highway driving assist system for use on limited access roadways, GM is not relying on customers to understand where it does and does not function. Instead, the system will check the navigation map—if the vehicle isn’t on a suitable road, the driver will not be able to activate it. In contrast, Tesla’s operating instructions state that AutoPilot should only be used on divided, limited access roads, but there is nothing in the system to actively prevent a driver from using the system in an urban area or any other roadway that it’s not designed for.

Similarly, Tesla doesn’t really take measures to prevent operators from taking their attention away from the road. Countless videos have been posted by Tesla drivers as they take a nap, read, or even climb in the back seat while using AutoPilot. The research conducted by Bryan Reimer and the Advanced Vehicle Technology Consortium at the Massachusetts Institute of Technology reinforces the idea that even informed drivers will get distracted while using systems like AutoPilot or Volvo’s Pilot Assist.

Improving Safety

Cadillac is installing an active driver monitoring system in the CT6, which will include more prominent alerts if the operator does not remain engaged while using Super Cruise. If the driver does not respond, the car will pull to the side of the road and come to a safe stop.

GM safety engineers have also addressed the issue of the inevitable mechanical failure. When fully autonomous vehicles arrive, they will require systems that can maintain control during a failure mode until the vehicle is safely stopped. One of the key safety failure modes for a system like Super Cruise is the electrically assisted steering.

One of the optional features on the currently available CT6 without Super Cruise is the Active Chassis Package, which includes a rear-wheel steering system to aid low-speed maneuverability and high-speed stability. This rear steering system will be included on the CT6 with Super Cruise. While the rear steering is not designed to provide the same full maneuvering capability of the normal front steering, it will be sufficient to safely steer the car to the side of the road in the event of a front steering failure.

We won’t have an opportunity to fully evaluate the capabilities of Super Cruise until sometime next year, but it does inspire some confidence that GM is at least thinking about and trying to address both human and mechanical failure modes before putting the system into customer hands.

 

Automakers Need to Start Being More Candid About the Limits of Autonomous Technology

— July 1, 2016

Connected VehiclesWhen was the last time you ever actually read an end-user license agreement or terms of service before clicking “Accept” to install a piece of software or join the latest social network? Odds are that unless you are a lawyer, the answer is never. The technology companies that make these products would probably like it to stay that way. However, in the world of the self-driving car, that is not an acceptable policy. The tragic death of a Tesla Model S driver in Florida highlights the need for all automakers to be more open and transparent about the limitations of autonomous technology.

Revolutionary (When It Works)

It seems that barely a day goes by when we don’t get a breathless press release from an automaker, supplier, technology company, or Silicon Valley startup about the amazing progress that they are making on self-driving technology. You can already go out today and purchase vehicles from a number of brands that promise at least partial autonomous capability, and full autonomy is being targeted by the end of this decade. While Tesla Autopilot, Volvo Pilot Assist, and other similar systems seem truly magical when they work as advertised, there are far more scenarios where these systems do not function at all.

Unfortunately, we have not seen Tesla CEO Elon Musk stand on a stage and tell people not to use Autopilot in the city, on curving rural roads, or in the snow. GM CEO Mary Barra stood on the stage at the 2014 ITS World Congress in Detroit and promised a Cadillac with hands-off Super Cruise capability in 2016. I’ve experienced prototype systems from Toyota and Honda and driven production systems from Tesla and Volvo, and when they work, they are incredibly impressive.

I am an engineer by training and technology analyst by trade, and I have a much greater understanding than the average consumer about how these systems work. As a result, I can never truly relax with these systems because I’m always on the lookout for the failure mode, and they are numerous. Unless very explicitly told, the average consumer will be so excited by the prospect of turning over control to a computer that they will not pay any attention to the warnings that Autopilot is very much in beta before enabling it. Volvo doesn’t even give that warning before allowing Pilot Assist to be engaged.

Mainstream Customers

Tesla is fortunate that many of its existing customers are early adopters that expect technology to be imperfect, although most of them probably don’t expect to be at risk of injury when it fails. When the Model 3 arrives and mainstream consumers try Autopilot and find its limitations, they aren’t likely to be as forgiving, and the same is true for every other automaker offering autonomous features. Navigant Research’s Autonomous Vehicles report projects more than 4 million autonomous-capable vehicles to be sold by 2025. Those customers need to know what the systems can do—and, more importantly, what they cannot.

We don’t yet know all the details of what happened in the tragic crash in Florida. Similar accidents where one vehicle crosses a highway divider happen all the time, and fatalities occur when humans are in control. What we do know is that we are far from a time when we can just sit back and relax and let the computer do the driving. Every company involved in this space needs to be far more upfront with consumers about this technology can do or risk poisoning the market.

 

Blog Articles

Most Recent

By Date

Tags

Clean Transportation, Digital Utility Strategies, Electric Vehicles, Energy Technologies, Policy & Regulation, Renewable Energy, Smart Energy Practice, Smart Energy Program, Transportation Efficiencies, Utility Transformations

By Author


{"userID":"","pageName":"Advanced Driver Assistance Systems","path":"\/tag\/advanced-driver-assistance-systems","date":"12\/11\/2017"}