The Current State of Autonomous Driving

The Current State of Autonomous Driving

There’s no denying it: autonomy is a hot area in the automotive industry. It’s not just automakers who are working hard to develop this advanced technology, as many tech firms are also burning the midnight oil to be the first to patent exclusive designs.

You’ve probably heard plenty about self-driving cars and how they’ll revolutionize the way we get around. After all, autonomous technology won’t engage in road rage, break traffic laws, or get tired late at night. But there’s plenty of controversy around cars driving themselves. After a highly publicized incident in Arizona where a self-driving car ran over and killed a woman crossing the street at night, suspicions about autonomous cars has risen.

Not helping with things is eager tech and automotive companies which like to make their systems sound perhaps a little better than they actually perform. There have been numerous accidents with drivers who wrongly assumed they didn’t have to pay any attention to what their car was doing since they were under the impression the autonomous system would manage everything.

Following is a summary of where autonomous driving technology is at the moment. Everyone, whether you own a car with any autonomous features or not, should be aware of how far the tech has come as well as where it’s heading.

Levels of Autonomy

Most people wrongly think an autonomous or self-driving car is one that can navigate city streets without any human assistance. While autonomous cars theoretically can do that, the reality is that there are several different types of levels of self-driving cars.

To really understand what’s going on with autonomous cars, you first need to know about these widely-recognized levels of autonomy for self-driving cars. There are six, and each level decreases the involvement of a human driver in the process.

There are different levels of autonomous vehicle technology

Level 0

We reached this level a while ago. The automated systems in the car will provide safety warnings, such as an impending frontal collision. In addition, systems can intervene to prevent an accident, but they don’t maintain sustained control of the car.

Level 1

Often referred to as the “hands-on” level, the driver shares control of the car with the automated system. One prime example of this kind of tech is adaptive cruise control. The driver steers the car, keeping it in the lane, while the autonomous system maintains a preset following distance with any other vehicle out in front.

Parking assistance systems are another common Level 1 autonomous feature. The driver controls the accelerator and brake pedals, but the autonomous system manages steering into a parking spot.

With Level 1, the driver has to stay fully engaged and be ready to assume full control of the car at any given time.

Level 2

Called “hands-off” by some, Level 2 involves the automated system taking full control of the car’s operation. This includes not only acceleration, but also braking and steering. However, the driver still needs to be monitoring the car and its surroundings since it might be necessary to take over at any given moment.

A prime example of Level 2 is Tesla’s Autopilot. While drivers don’t need to be managing anything about the car’s operation, the system does require the driver to grip the steering wheel periodically to indicate attentiveness.

Level 3

Referred to as “eyes-off,” drivers using Level 3 autonomy can actually not pay attention to the road and surrounding cars, leaving the system to manage everything. But, while the human driver can watch a movie or do other things not requiring them to be watching the road, they must be prepared to take over the operation of the car at any moment. The car will provide a quick warning that the driver needs to take control again, but it could be quite brief.

The Audi A8 has a Traffic Jam Pilot system that’s Level 3. It manages driving in slow-moving traffic, but once the car starts moving over 60 kilometres per hour, the human must assume control again. Also, the technology only works on highways where there’s a physical barrier separating oncoming lanes of traffic.

Level 4

What some people call “mind off,” with this level of autonomy you could safely go to sleep or even sit in a seat other than the driver’s seat. The system will only drive the car in specific locations, thanks to geofencing technology, or in specific driving conditions. If the system runs into trouble, it can safely pull off the road and park if there’s a problem, which is why a human doesn’t need to be paying attention.

Level 5

This can be lovingly referred to the “steering wheel optional” level. That means a human driver isn’t necessary at all, because the autonomous system can manage everything bout navigating roads. A robotic taxi that transports people to their destinations in a city without any human oversight is a good example of this tech.

Waymo Service

Waymo, which was founded in 2009 by Google, is a self-driving technology company. In the past several years, the company has formed alliances with several automakers, including Jaguar Land Rover and Fiat Chrysler Automobiles.

Something amazing happened in October of 2018. Waymo became the first company in the United States to receive a permit for operating a full-driverless car on public roads.

With that historic ability, Waymo has started offering commercial rides to the public. For now, the service is only available in the metro Phoenix, Arizona area. For now, only what Waymo terms “early riders” can hail rides from the driverless minivans. Those early riders participated in the company’s research program, so they already have some experience with autonomous tech.

While the Waymo service is restricted right now, the company anticipates offering it to more members of the public in the near future. There’s no definite timeline for a rollout, or announcements of it being made available in other areas of the US or Canada. Still, this is a huge step forward for autonomous car technology. Researchers will be able to learn countless things from this service as it runs day after day, week after week, transporting people to all kinds of locations around the Phoenix area.

Autonomous taxi services have already started

Yandex has launched its own driverless taxi service for the public in two Russian cities. The Russian tech company faces a challenge Waymo doesn’t: more inclement weather. Since it virtually never snows and often doesn’t rain in Phoenix, the Waymo autonomous vehicles don’t have to navigate through wet roads and deal with covered lane markers. Already, Yandex showed off its autonomous taxis in Las Vegas for the Consumer Electronics Show 2019.

It’s important to point out that both the Waymo and Yandex taxis aren’t just vacant vehicles riders hail and use by themselves. There is a human driver behind the wheel, ready to take over when the autonomous drive system needs assistance. In other words, these are not Level 5 or even Level 4 cars by any stretch of the imagination. Still, it’s an important step forward for the tech.

Sensors

Perhaps the most important and least-understood aspect of autonomous cars is the onboard sensors. They are absolutely essential for a self-driving car to reliably navigate a changing environment. While you could teach a car where the lane lines, curbs, and traffic lights are in a city, that’s not enough for it to navigate through traffic without accidents.

Other cars on the road are constantly changing positions. Pedestrians might cross the road at pretty much any time. Cyclists can be on the shoulder. Traffic lights change. Emergency vehicles will turn on their lights and sirens to move quickly through traffic. Construction can alter lane patterns temporarily. A sudden accident in the lane you’re in can require emergency navigation. As a human driver, you rely on your sense of sight and hearing to react to the changing environment. For self-driving cars, they must have sensors which are sophisticated enough to perceive everything a human can, if not even more.

LIDAR

Many believe that LIDAR is the most essential sensor for self-driving cars. Through this sensor, the autonomous system has a 3D scan of the surrounding environment, allowing it to react to changes more like a human would. Since it works using lasers, the precision of LIDAR is phenomenal.

There’s one huge obstacle with LIDAR sensors: they’re quite expensive. This means the cost of developing any self-driving vehicle is high. Several big tech companies are working on a solution to the cost problem, but for now it doesn’t seem to be forthcoming.

Cameras

We all know and love cameras since they’re on smartphones and many other devices. While cameras provide images of what’s surrounding a self-driving car, the autonomous computer must use machine learning to make sense of the data. Those images are in 2D, so computer power is necessary to transform that into a 3D model.

We’re already familiar with cameras on cars and elsewhere

Radar

Once a technology used mostly for airplanes, radar has become popular for cars. Most blind spot and rear cross-path detection systems on modern cars use radar sensors. They aren’t affected by inclement weather like heavy rain, fog, or even bright sunlight, which is a huge advantage over cameras. Even better, radar can “see” things that are otherwise obscured, like an approaching car blocked from view by a cement wall.

V2X

The ability for a car to communicate with other cars, infrastructure in a city, pedestrians, or a local network is V2X. Among the advantages to this kind of sensor is that known conditions from other sensors can be communicated to nearby self-driving cars, such as an accident, construction zone, etc. The more eyes that are providing data about what’s going on in an area, the better.

We haven’t seen this kind of vehicle sensors used much by automakers, but that appears to be changing. Ford says by 2022 all cars and trucks will have C-V2X capabilities, thanks to its partner Qualcomm. Expect other automakers to react by also offering V2X in their lineups.

Software

Then there’s the software side of things in an autonomous car. All that data from sensors must be processed, which is often done through algorithms and artificial intelligence. These work to make sense of everything surrounding the car, which is essential for successful navigation.

There’s also a planning module, a necessary component for making decisions as the car travels. It comes loaded with established policies and information about traffic laws. To properly plan a route, the system uses maps of the area.

This is an area of considerable research and growth for autonomous cars. While sensor issues certainly are a catching point for self-driving tech, software development is as well. Once better software has been created, which will be an evolutionary process, achieving higher levels of autonomy should be possible.

Is Level 5 Out of Reach?

Despite all these advancements in autonomous drive technology, some people in the automotive and tech industries speculate that Level 5 autonomy will be impossible to reach.

The reality is that for now Level 5 is purely theoretical. There are no cars existing at this time which can drive without the need for human intervention at some point.

To drive through cities and even in the open country, cars need more than just all the right sensors and advanced software. They also must be able to use judgement and exercise conjecture. In some situations, wisdom is necessary. These are all human qualities and we have no idea if a machine can be programmed with them or learn such skills.

Most driving is pretty routine and mundane. You navigate busy, well-marked roads to commute from home to work and back again every day. Machines can do that, and likely better than you, because it’s simple and they don’t get bored. But if a sinkhole were to suddenly open up in the road a hundred metres ahead of your car, as a human you’d know to hit the brakes and/or steer around it. An autonomous car might not know what to do, or worse wouldn’t perceive the hole and continue ahead. The same goes for if you see a brightly-coloured ball bounce into the road. Your mind will rightly conjecture that child could come chasing after the ball, so you slow down and wait. Would a self-driving car know to do that?

Get Your Car Loan Approval

Related Articles

Apply for your Car Loan

Apply Now and Experience a Whole New Way to Buy a Car

Apply Now