Everything you need to know about self-driving cars
For a long time, everyone thought that flying cars were the next big step in automotive technology. After wising up and realizing that the world needs new technology to make driving easier and safer, driverless vehicles entered the conversation.
Road accidents claim millions of lives every year and cause tens of millions of severe injuries, with human error being responsible for about 90% of these accidents. Driverless cars are the most realistic way of solving this problem.
Development of these autonomous cars is well underway, and many of the biggest names in the car and tech worlds are competing for and collaborating on self-driving cars. In just a few short years, driverless cars have gone from being a futuristic dream to an inevitable, and in some ways problematic, reality.
How do self-driving cars work?
A fully autonomous car has to obey road rules, respond quickly to unexpected situations and drive safely in all circumstances. In other words, it has to be better than the average human driver.
This level of accuracy is achieved with built-in systems that constantly analyze the car’s surroundings and automatically take the right action. The exact technologies used in self-driving cars vary between manufacturers, and are often closely guarded secrets, but generally all self-driving cars have three key parts.
Today’s technology already has this pretty well covered, but there’s still room for improvement. The main reason all driverless cars need GPS is so that you can tell them where you want to go and they can take you there. The technology itself isn’t all that different from GPS phone apps, but having it built into the car makes it a lot more efficient.
- GPS is extremely accurate and can help driverless cars stay between the lines and on the correct side of the road.
- Driverless cars have to be able to decide between multiple options and choose the best route. Having up-to-date GPS information makes this possible.
In order to plot where it’s going and be prepared for hazards, self-driving cars need sophisticated detection technology. To date, most driverless cars have used a combination of LIDAR (light detection used to measure distance), radar and cameras to see obstacles like guardrails, other vehicles, buildings, pedestrians, lane dividers, road markers and even street signs.
- LIDAR, sometimes known as LIDAR, light radar or laser scanning, involves sending invisible lasers out into the surrounding area. The way the light bounces off the objects helps determine how far away they are and what kind of objects they are. In self-driving cars this is done continuously in all directions to create a realtime map of the vehicle’s surroundings.
- LIDAR needs more work before it’s good enough for fully self-driving cars. It still has a hard time in wet weather and snow as these affect how light is reflected and can also cover up road markings and street signs.
- Radar is used to detect vehicles at a distance, even in poor visibility where LIDAR may fail. Newer vehicles with adaptive cruise control typically use radar to detect cars in front of them and to automatically adjust their speed.
The first ever self-driving car fatality was caused by a LIDAR error in a Tesla Model S, when a car on autopilot drove right into a truck that pulled out in front of it. This happened because it was a white truck and an overcast day. The LIDAR misidentified the off-white truck as part of the grey sky, so the car didn’t hit the brakes.
Cars have had onboard computers and electronic networks for decades now. However, driverless cars usually need multiple highly advanced computers and a much more comprehensive electronic network. This is because it has to analyze a huge amount of information on the fly, such as the current rotation speed and traction of each wheel, LIDAR and GPS information, engine heat, the constantly updating location of surrounding vehicles and much more. At the same time, it will be feeding this information through algorithms to determine whether any actions are required, whether surrounding vehicles are likely to change lanes or slow down and anything else.
- Computers can do things with cars that humans simply can’t. Anti-lock braking systems (ABS), for example, are able to pump the brakes much faster than a person could, while traction and stability control is able to take over individual wheels and control them independently of each other.
- Driverless cars are able to identify a situation, assess it and take action in just a fraction of a second.
- LIDAR, camera, radar, GPS information and everything else is being constantly assessed by the onboard computer. For every moment it’s on the road, a driverless car is always asking itself what the situation is and whether it should be taking any action.
How safe are driverless cars?
Several companies, including Google, are programming their cars to not take any action, such as changing lanes, accelerating or braking, unless there is a 0% chance of it resulting in an accident. Theoretically, this ensures perfect safety as long as all the sensors and electronics work as intended, and as long as nearby human drivers don’t do anything unpredictable.
However, it’s important to keep in mind that this technology is new and the kinks are still being ironed out. There have been accidents involving driverless cars, including one fatality, but many of these have come down to driver error. The future of driverless cars promises fewer accidents and safer roads, but we’re not quite there yet.
What are the pros and cons of driverless car technology?
- Safety from faster reaction times
- Efficiency and accuracy of driving even at faster speeds
- Freedom, especially for disabled people
- Job losses and economical impact
- Extra costs from added features
- Unanswered legal & ethical questions
- Human error will still be a factor
What are the advantages of driverless cars?
The biggest and most obvious advantage of self-driving cars is safety. Autonomous vehicles outperform human drivers across the board. They have:
- Faster reaction times. A human driver will usually be the slowest part of a car. Driverless vehicles can identify hazards and take action before a person even sees it.
- Smarter decisions. In an emergency situation you don’t have enough time to think things through. Driverless cars, however, do have enough time to analyse a situation and decide what the best move is.
- Better drivers. Machines don’t get fatigued, distracted or drunk, which cuts out three big risk factors right away. They obey the speed limits, follow road rules and pay attention to warning signs, and it’s safe to assume they have no interest in road rage, tailgating or drag racing. Some people are simply bad drivers, but computers are consistently good.
Another clear advantage is efficiency. By itself, an individual driverless car can be more efficient in almost every way, but having the roads full of autonomous vehicles would change everything.
- Better fuel efficiency. Individually, a self-driving car can average better fuel efficiency thanks to being able to maintain more consistent speeds and brake and accelerate in an optimal way.
- Consistent behavior. Driverless cars are more predictable, keep consistent speeds and can drive much closer to each other than people can, letting them reduce congestion.
- Faster speeds. Autonomous cars can safely travel at much higher speeds than people can.
- No more traffic jams. A driverless car network would allow for optimization of traffic flow. For example, a highway lane for driverless cars only might look like an unbroken chain of vehicles, all driving at exactly 90 mph with only centimeters between each one.
April 2016 saw a successful test of self-driving truck platoons. Six convoys of self-driving trucks left factories all around Europe and drove themselves to Holland. Each convoy consisted of two or three vehicles, all connected by wireless devices, and one lead truck that determined the route and speed of all vehicles in the convoy.
This test successfully demonstrated that driverless vehicles can adjust to different road rules when crossing national borders and that there are benefits to be had from networking multiple driverless vehicles together. It also showed real efficiency gains. The trucks didn’t need to stop for food or rest, maintained consistent speeds across enormous distances and were able to improve energy efficiency by slipstreaming, or following another vehicle closely.
More freedom for disabled people
When cars first hit the market all those years ago, they promised a new type of freedom. Unfortunately this didn’t really help anyone who was blind or otherwise disabled and unable to drive. Driverless cars bring a whole new promise of freedom, which is more important than it might seem.
- Elderly drivers are enormously overrepresented in car crash statistics, largely because reflexes simply get slower with age. They’re also more likely to be in a position where they have no choice but to drive, even if it’s unsafe. As the elderly population grows, this will become an increasingly serious issue. Self-driving car technology is both an effective and realistic solution to this problem.
- Being able to live independently is important when adapting to disability, but sometimes being physically unable to drive is an insurmountable obstacle. Driverless technology means a lot of people will suddenly be able to support themselves where they previously couldn’t, alleviating the pressure on their support networks.
Did you know?
Tesla drivers had safely used the self-driving autopilot feature for a combined 160 million kilometers before it resulted in a fatal accident. By contrast, human-driven cars tend to have a fatal accident every 96 million kilometers.
What are the disadvantages of driverless cars?
Like most revolutionary technology, driverless vehicles will have a big economic impact that we’re largely unprepared for. One of the best ways to get an idea of what to expect is by looking at certain critical industries. Consider trucking, for example.
Trucking will switch to driverless vehicles as soon as possible.
This is because driverless trucks will earn much more and cost less. Driverless trucks don’t need to take breaks, can travel a lot faster, use less fuel and are much safer. Any trucking company that doesn’t switch to driverless vehicles will be at a huge disadvantage next to its competitors, so most companies will be keen early adopters.
- There are about 173,000 truck drivers in Australia. Almost all of them could be in danger of losing their jobs depending on how quickly it’s taken up in Australia.
- Filling stations, roadhouses and many other businesses depend on truckers for a regular flow of customers. Many of these will also go out of business.
- Trucking is, by a large margin, one of the most dangerous jobs in Australia. A driverless fleet means trucking companies can save a huge amount on workers compensation and employer’s insurance.
- Trucks are a crucial part of Australia’s business infrastructure and efficiency gains here will benefit most other industries. Despite the downsides, and the loss of hundreds of thousands of jobs, the cost of not switching to driverless trucks could be even higher.
This type of pattern will be seen across many areas. Driverless cars promise much more efficiency and cost-effectiveness than human drivers and any business that doesn’t make the switch will find itself at a competitive disadvantage. This means that the mass uptake of driverless vehicles for business purposes could sweep across the world before they ever become similarly popular for personal use.
The downsides will most likely be severe and largely unavoidable, particularly if you work in an affected industry.
Five driverless car problems that still need fixing
There are still a few problems that need to be overcome before driverless cars are ready to replace human drivers.
- The cost. All the technology that’s packed into driverless vehicles doesn’t come cheap and will bump up the price of a car considerably, particularly while the technology is still relatively new. Manufacturers want a self-driving car that they can sell to a mass market audience, and that means finding a way to lower the cost. This will happen naturally as the technology gets more affordable, but the cost is still high at the moment.
- Legal issues. If two self-driving cars crash into each other, who’s fault is it? Is the occupant of a self-driving car a driver or a passenger? Are you legally required to keep your eyes on the road while riding in a driverless car? Do driverless cars really need a steering wheel? There are a lot of legal questions around driverless cars and most of them have no clear answers. A lot of legal and legislative work still needs to be done in almost every state and country before driverless cars are able to hit the roads en masse.
- The roads. The cars themselves are only part of the picture. One of the burning questions in the industry is whether driverless cars should be designed for existing roads, or whether roads should be redesigned for driverless cars, such as by placing sensors on lane dividers or creating special driverless car lanes. With today’s technology we could already have fully functional, fully self-driving cars out in public, but only on specially designed roadways. Whether these infrastructure adjustments will ever happen determines what the future of driverless cars looks like.
- Ethical problems. Driverless cars need to be programmed to make life or death decisions without human intervention. What if your car has no choice but to decide between killing you or a pedestrian? These kinds of situations will be rare, but given the sheer number of cars around the world they will happen and they need to have programmed responses ready for such situations. Surveys around these questions have shown that people recognise sacrificing oneself as the nobler option and approve of cars that make decisions this way, but this may because they have no intention of buying one of these cars themselves.
- Human error. Somewhat ironically, some of the hardest driverless car problems to solve lie in the inseparable human aspect. A major one is that a lot of people simply enjoy driving or otherwise don’t want to give up control to a machine. Having even one human driver among robots will make everything much less efficient and much more dangerous, but people will do it anyway. Another human problem is the simple errors they may make, such as jerking the steering wheel while sneezing or accidentally pressing buttons. As long as people still have some measure of control over the car, these factors will continue to cause accidents and deaths.
What self-driving car technology can you get in the US?
Fully autonomous vehicles aren’t on the roads yet, except for testing purposes, but there are many assisted driving technologies already out there. These systems are the stepping stones to driverless tech, and are quite widely available in the US.
The good news is that you can get these features without breaking the bank, and some might even save you money in the long run by preventing accidents and lowering your insurance costs. You might want to consider getting a car with these features if:
- You’re a fairly inattentive driver who gets ticketed a lot. These systems can keep you driving safely and legally even if your attention wanders a bit.
- You’re in the market for a brand new car. It might be easier to find a car with these features than without them.
- You aren’t confident in your ability to parallel park, you often need to park in tight spots or you just want to make parking easier for yourself.
Car loans for driverless cars
If you’re buying a car with financing, these autonomous features are still worth considering, not only as valuable features for your car, but also for how they might help your insurance costs.
Keep in mind that some car loans might require you to take out a comprehensive car insurance policy. Because these cost more, discounts can help you save. Driver assist technologies that demonstrably lead to safer driving, such as adaptive cruise control, are recognized by some insurers who may reduce your premiums if you buy them.
Current autonomous technology
|Technology||What it does||Example cars||Advanced details||Costs|
|Adaptive cruise control||Cruise control that also adjusts your speed to match the car in front||High-end cars from most major manufacturers within the last two years||Earlier versions require a lot of human input, but recent models can take over for extended periods of time||About $2,000 extra, depending on the system|
|Automatic parking||Takes over entirely and parks the car for you||Recent models of the Ford Focus, Audi A6, Jeep Cherokee, all new sixth generation BMW 3 Series, most new Mercedes-Benz models and many more||Within the last two years cars have become able to safely park themselves in most situations||About $1,000 extra|
|Hands-off steering||Makes the car steer itself; the closest thing to a fully driverless car||Specific models of newer, higher-end cars like the Audi A7 and BMW 5 Series||Cars are generally able to drive themselves in good, clear conditions, but have difficulties in more complex situations||$3,000 to $6,000 extra, depending mostly on which other features it comes bundled with|
|Automatic lane departure warning||Warns you if you drift over lanes, can automatically steer you back||Many different models and manufacturers have varying forms of this||It generally depends on having clearly visible lane markers, but newer versions have more control than older ones||Starts from about $1,500 extra, depending on whether it’s a warning system only or if it also has automatic steering|
What will driverless car insurance look like?
Newer model vehicles are often cheaper to insure than older ones. This is because spare parts are a lot easier to get and more widely available. Paying more for a newer vehicle with driver assist technology might save you money in the long run.
No one is entirely sure what will happen to car insurance when fully autonomous cars hit the roads in large numbers, but it’s clear that it will need a lot of adjustments.
- Who’s at fault? Ordinarily, car insurance companies will pay out if their customer is at fault, or will claim expenses from the other party’s insurance if their customer wasn’t at fault. This blame game is much harder with self-driving cars because the manufacturer might be more at fault than any of the vehicle occupants.
- How safe are driverless cars? Self-driving cars will be safer than manned vehicles, which could lower insurance premiums considerably. However, this will only happen when it has been proven beyond doubt that all driverless cars are generally safer. Difficulties in gathering good data and differences in the quality of driverless systems might delay this. Don’t expect reduced premiums for driverless vehicles until they’re relatively common and well established.
- Will manufacturers get involved? There’s a good chance that car manufacturers will become much more closely involved with car insurance. In the transition period when fully driverless cars start flooding the market (it’s thought that this will happen between about 2021 and 2030), manufacturers might start offering their own car insurance policies for specific vehicles to make up for the lack of good options elsewhere.
- car insurance will become more standard. Black box car insurance measures a driver’s safety with a black box device in the car, which is then used to determine their premiums. Although common elsewhere, this type of insurance is still a rarity in Australia. However, with driverless cars, these devices could become invaluable for insurers who need to determine the quality of self-driving systems and find out who was responsible for an accident.
- Insurers often cover themselves with the general requirements that you must pay attention while driving and obey all road rules. This will need to be amended in the age of driverless vehicles. If not, the occupant of a driverless car might be automatically liable for any accident simply because their hands weren’t on the steering wheel at the time, even though they didn’t need to be.