Self-Driving Cars: All You Need To Know Today
It has been over 100 years since Karl Benz introduced his first concept of automobiles and invented the first car ever. Since then, humans have achieved wonderful progress in car manufacturing with great improvements which not only make driving safer for drivers but also allow them to enjoy great comfort while driving. For example, we’ve added seat-belt and airbag, or we’ve made engines stronger.
But none of that seems to be a significant change in the automobile industry as the number of casualties keeps increasing over the years. You might be in shock to find out that there are over 1.2 million deaths due to traffic accidents each year, and in America, the number is roughly 30 thousand. Put simply, over 82 people die each day because of traffic accidents.
It is clear that humans tend to become more careless as the technology advances. There are simply too many distractions that keep reducing our attention to the wheel and pedal. Therefore, for the sake of safe driving, people are turning toward self-driving cars with fantastic autonomous features that allow us to drive hand-freely and in a much safer manner.
With all of the achievements being made and technologies being introduced ever since the birth of the first self-driving car by Karl Benz, the following is all you need to know about driverless cars today.
What Is a Self-Driving Car?
Autonomous, self-driving or driverless cars are vehicles that are capable of driving on the road on its own using sensors and an analyzer system.
Technically speaking, humans are now capable of producing cars that can automatically operate and don’t require human interference. But being able to create that kind of car requires a huge amount of efforts and investments. Not to mention that there are many laws restricting how, where, and when these intelligent vehicles can operate.
At the moment, you can purchase your own self-driving car as many manufacturers have already introduced their models. However, your car will be under some restrictions and won’t likely to be able to go everywhere like normal cars.
Terminology and Safety Considerations
Vehicles today come with partly automated attributes – for example, maintaining your car within its speed controls, lane, or emergency braking. Yet, there remain differences between fully autonomous self-driving models and the ones with driver assistance technologies. Reportedly, confusion between those concepts can be the culprit behind deaths.
When it comes to marketing for vehicles these days, the employment of the word ‘autonomous’ is dangerous as car ads make many drivers think ‘autopilot’ and ‘autonomous’ are equal to vehicles with the capability of driving itself. The truth is, they still count on the motorist to ensure safety. Believe it or not, technology alone has not been able to control the car entirely thus far.
When a couple of car manufacturers claim that cars are self-driving, when they are just the partly automated motors, many drivers risk getting overly confident, resulting in crashes.
Autonomous vs. Automated
Self-governing is another word for autonomous. Many previous vehicles have been automated considerably, relying on artificial support in their environment, say, magnetic strips. By autonomous control, we mean satisfactory performance under the environment’s specific uncertainties and the ability to handle system failures without your intervention.
The term ‘autonomous’ is currently more popular in use when compared to ‘automated’ – in other words, we are more familiar using that word. That being said, the latter name is probably more accurate. It refers to operation or control by a machine.
On the other hand, ‘autonomous’ means acting independently or alone. Many vehicle concepts which you are today aware of get one person in the seat of the driver, make use of a communication connection to other cars or the Cloud, and do not independently choose either routes or destinations for reaching them.
While the word ‘automated’ should describe these car concepts more accurately, you can use both of them together as well. For example, according to the Regulation (EU) 2019/2144, “fully automated vehicle” and “automated vehicle” are defined depending on their autonomous capacity:
- “fully automated vehicle” means your car which is manufactured to move autonomously without your supervision.
- “automated vehicle” means your car that is made to move autonomously for a particular time without your constant supervision, yet there still needs your intervention.
You may be wondering, how can a vehicle be closely on par with human intelligence while performing something so complicated? The answer to this question lies in the construction of this type of car:
- Computer vision: It is the way your motor vehicle sees its environment.
- Sensor fusion: Since there are many sensors integrated on the car, you can take every essential data as well as fusing them to deliver a more potent diagram to your vehicle, thus the name sensor fusion.
- Localization: how your vehicle knows its location.
- Path planning: how it thinks about the most favorable direction.
- Control: the way it steers the wheel.
When Were Self-driving Cars Invented?
The idea of these vehicles came far earlier than the giantly accessible Google’s research these days. Their concept was indeed originated at one exhibit at the New York World’s Fair in 1939 – Futurama.
General Motors (GM) held the exhibit to showcase its vision of what the globe might look like in the upcoming 20 years. Included in this vision is the automated highway system guiding self-driving vehicles.
While a world full of robotic cars has not been a reality, modern vehicles do feature autonomous things, like braking systems and assisted parking. Above all, efforts poured on developing fully automated vehicles are ongoing, aiming at making controlling your car more comfortable and safer in the next decades.
Talking back to the 1939 exhibit of GM, the very first self driving car is the creation by Norman Bel Geddes. Radio-regulated electromagnetic fields produced with magnetized metal spikes integrated into the roadway guided that electric vehicle.
By 1958, GM had turned the concept into a reality. Its front end was built-in with sensors named pick-up coils with the capability of detecting the electric running through a wire integrated into the road. The power current was likely manipulated to tell the car to proceed with its steering wheel right or left.
The Japanese, then in 1977, enhanced upon the idea, making use of a camera system relaying data to a computer for the road images processing. Still, this car could only move at speeds under 20 mph.
One decade later, the Germans created the VaMoRs. Cameras outfitted this vehicle so that it could travel itself at up to 56 mph safely. Since technology advanced, so did automated cars’ ability to notice and react to the environment.
How Do Self-Driving Cars Work?
An autonomous car needs essential sensor and navigation systems so that it can see what’s happening around. Then you need to program an intelligent analyzer system so that your car can process the data coming up before making accurate decisions on how fast it should run, whether there is an obstacle to avoid or not, and so on.
Although the concept of self-driving cars can be quite simple and we might have advanced self-driving cars technologies to pull it off, it is still a long, challenging process requiring lots of money and efforts. Not to mention that autonomous cars are not capable of performing complex analysis in certain situations like what a human brain can do. So, don’t think that these smart vehicles will allow you to place your hand off the wheel and look for a quick nap.
For self-driving cars to fully understand what’s happening on the road, most of them must come when through these two stages:
First, the car must be able to see what’s happening around it, which is called the perception stage. While a human can glance at objects to receive data about it, a self-driving car will need a whole set of different equipment like radars, cameras, ultrasonic sensors, and LiDARs.
In case you haven’t know what Lidar is, it is short for light detection and ranging which is a crucial feature for any driverless car. In fact, these devices are doing a pretty good job of interpreting input data to create a 3D model of everything surrounding it. This gives it a clear vision of the road and reduces many misidentifications.
Predict and Planning
However, the traffic is not that simple! Your car must be able to predict and plan next moves of not only itself but also surrounding vehicles. Sometimes this can be quite simple if the car is on a clear road with no object ahead of it. But there are a lot of scenarios where self-driving cars are not totally prepared for yet.
For example, if you’re going to make a turn and notice a vehicle moving in the opposite direction. Your car must be able to calculate that vehicle’s current speed and light signal to identify whether it should slow down or keep its current speed. Moreover, because human isn’t a machine, we often rely on our senses more while driving, which means there will be situations where the cars don’t know how to react to the real environment properly.
With that being said, prediction and planning are still the biggest obstacles that self-driving car developers have to deal with.
The Current State of Self-Driving Cars
So, when will self-driving cars be available? Are they expensive?
Every year we see companies, such as Tesla, Mercedes, Toyota, Nissan, Google (Waymo project), introducing their new concepts of self-driving cars in the CES (the International Consumer Electronics Show). But this year, there have been talks about how these cars will be used and manufactured, which is a good new because they might consider manufacturing them commercially.
In 2020, we might start seeing some of these models hitting the road. But still, it’s important not to get ahead of ourselves as these vehicles will probably available in certain areas only and will take time to be popularized.
All self-driving and security features will probably be updated and improved frequently under the form of online subscription. For instance, Tesla have delivered software updates for their driverless cars.
Meanwhile, you will gradually realize that the cars you drive every day are becoming super smart and convenient. The Tesla’s Autopilot has already allowed cars to work automatically with little interference from drivers.
Our future self-driving cars would probably still be under control at a certain level by their manufacturers. And to create a network of driverless vehicles, it is clear that these cars should be connected to each other at a certain degree for synchronized reactions in most cases.
Self-Driving Cars Pros and Cons
Autonomous smart car scans the road and automatically stops at the crosswalk in the city
Although it is clear that self-driving cars will give us many advantages while driving, there exist some controversies. Here are the pros and cons of autonomous cars:
- The interior of driverless cars could become a leisure room where drivers can spend their time relaxing and doing productive works rather than driving. Also, people will have more time to talk and don’t have to face forward at any time.
- It is estimated that over 80% of car accidents in the USA are caused by the driver’s fault. Once the traffic is operated by self-driving vehicles, the chance of having accidents will be greatly reduced.
- Drunk or drugged drivers will no longer be a problem with self-driving cars.
- Because every car will follow the rules obediently, this will reduce the amount of time and money the government has to spend to maintain a certain amount of security forces.
- The built-in sensors on autonomous cars can be more accurate than human senses which come handy while driving in the night or on tough terrain.
- People who find parking difficult will enjoy their freedom since the car will drop them at the destination and go park itself.
- Disabled people will have the chance to go wherever they want without having to take public transportation or ask their relatives for a ride.
- The self-driving car technology is still new and won’t be available for everybody. It is estimated that self-driving car cost would cost at least $100,000. So, it would take quite some time for self-driving cars to become popular.
- In a world full of autonomous vehicles, taxi drivers and truck drivers will likely to lose their jobs because there is no need for them anymore.
- Computer glitches and malfunctions can still happen even with the most advanced technology. And once it happens, the consequence can be even more severe than a simple car crash. Hackers can exploit these flaws in order to gain control over your car and cause serious damages.
- At the ethical aspect, there will be situations where your vehicles will face dilemmas where they have to choose to cause one accident to reduce the risk of multiple ones. If that happens and you’re the one who is going to suffer the most, will you even buy it the first place?
- The manufacturer will probably provide your cars with frequent updates to further enhance its safety, but they could collect the information of location and other user’s data.
- The process of collecting data, predicting, and planning is still a major challenge for driverless car manufacturers. For example, reading human hand signals or road signs is not always easy for the AI because its thoughts cannot be as flexible as ours.
- Until we are able to create a traffic system consisting of all driverless vehicles and help them communicate with each other, it will never be 100% safe for self-driving cars to operate.
Autonomous Car Classifications
With many types of cars being introduced as smart, self-driving, self-parking, or coming with driver assistant features, many of us will find it confused on autonomous car classification. In this section, we introduce 6 levels of cars in the autonomous classification chart.
These cars have no control over your driving at all. While it may have some sorts of warning system in the front and back so that you won’t hit your neighbor’s car when parking, everything is still manually performed by you and your car won’t be able to stop by itself at any situation.
If your car is capable of having certain control over itself at a minimum degree, it belongs to this category. Typically, it could be the radar-based control that automatically slows you down once it detects obstacles ahead, and then returns to the original speed when the road is clear. Or the lane departure warning system that tracks your movement to ensure that you stay inside the proper lane.
These are capable of driving on themselves in a few situations either by wheel handling or acceleration control. If your car is able to nudge itself to the left or right in any degree, or it comes with a self-parking feature, it belongs to this category. But remember, the driver must always keep their hand to the wheel to ensure everything is fine, which isn’t something we would want.
Level 3 has a decent intelligence to control the car but will still require the drivers to keep their eyes on the wheel. These cars are capable of performing self-driving on a normal road with clear vision. And you can even set the path and destination for it thanks to the navigation system.
But what is weird here is that these level-3 cars don’t have any kind of false safe, which doesn’t really make sense because the ultimate purpose of autonomous cars is to help drivers relax in the back of their vehicle and know that they’re safe.
With this type of car, you won’t have to put your hands on the wheel 24/7, but it still needs to be able to interfere at any second. It is kind of pointless, so we don’t often see this type of car.
That’s also the reason why manufacturers often choose to go straight to level-4 cars, which offer the same non-human driver mechanism but with better intelligence so that they can auto-resolve in tricky situations. For instance, if there are accidents or traffic problems ahead of you, your car must at least be able to warn you about it and automatically park itself on the side of the road.
A good example of this level is the Waymo car – a Google-backed self-driving car. It comes with powerful and effective sensors, helping to auto-resolve in most situations regarding the traffic. Also, the Waymo can perform complete self-driving with no human inside of it, but this is still restricted to certain areas and you won’t be likely to see these cars very often in the near future.
Our final goal when it comes to making autonomous cars is, of course, the ability to have 100% control of the car regardless of what situation it encounters. Driver interventions will be unnecessary with this type of car, and in fact, we find many concepts of future self-driving cars from famous manufactures having no wheel at all.
Despite the facts that famous self-driving car companies (Google, Tesla, Apple, Uber, etc.) are trying to develop level-5 autonomous cars, we’re not sure when and where these vehicles will be introduced. Probably won’t be anytime soon because there are still many controversies against robot-driving cars, and the manufacturers will have to convince consumers that their cars are 100% safe and won’t be hacked or altered by any means.
How Safe Are Self-driving Cars?
The good news is that though completely automated cars have been still some way off, a lot of new modern vehicles are already getting the most out of semi-autonomous driving technology.
For the not-so-good news, self-driving motors are not perfect; there are related risks when you are driving with them, including:
- The risk of drivers getting distracted and counting on the self driving automobiles to deal with unexpected situations, reducing any necessary reaction time in the matter of an incident.
- The inability of some infrastructure today to respond to errors, mistakes, or deficiencies caused by automated means of transport can result in accidents.
- Resulting accidents due to unwelcomed malfunctions with software systems.
- The cybersecurity threat of the auto getting hacked to steal the motor vehicle, its contents, or the private information of the owner.
Are Self-driving Cars Legal?
Since earlier in 2012, up to 17 states, as well as the District of Columbia, have been debating legislation in terms of authorizing automated vehicles on their roads. That being said, only Washington, D.C., Florida, Nevada, and California have indeed enacted the laws. For sure, it may not mean that self driving cars are not legal to run on other states’ roads, only that this means of transport is not expressly authorized.
The laws of the states tend to be silent as to whether or not a human is in control of the steering wheel. Most of them require you to operate the vehicle, yet they are not clear about precisely the number of the car’s tasks you must undertake as a driver and which your vehicle can take. While most states’ regulations assume you will hold the wheel, owing to such a legal vagueness, autonomous cars may technically be enabled to operate over the roads as long as you are in control of them.