Autonomous cars explained

In one of our previous articles, we have briefly discussed the pros and cons of autonomous cars. Today we wanted to dive deeper into the topic and explain what exactly autonomous cars are, point out differences in their types, understand how they work and what challenges they face.

First of all, autonomous cars are vehicles that are able to replace humans in the process of driving. They operate by sensing the environment and don’t require any action from human drivers in any case.

What’s the secret behind these smart, fully autonomous vehicles? To put it briefly, it’s all in sensors, complex algorithms, actuators, and powerful processors required for executing software. Thanks to the sensors installed in various parts of the car, they are able to create and remember a map of their surroundings: positions of other vehicles nearby, traffic lights, road signs, locations of pedestrians. They are also able to acknowledge curbs in the road and park on their own.

All of this sensory input is processed by special software and on this basis, instructions are created and sent to the car’s actuators responsible for controlling the acceleration, braking, and steering. How is it possible that a vehicle is able to follow traffic rules and avoid obstacles on their own? The answer is in, among others, hard-coded rules and object recognition.

According to the Society of Automotive Engineers (SAE), at present, there are 6 levels of driving automation and those were accepted by the U.S. Department of Transportation.

The levels are as follows:

0 - no automation - whole control of the car is completely manual,
1 - driver assistance - single automated units,
2 - partial automation - automated steering and acceleration, but controlled by human,
3 - conditional automation - most driving done by the vehicle but still controlled by human,
4 - high automation - the vehicle does all driving tasks, there is still the possibility for a human override
5 - full automation - the car is capable of all driving tasks and requires no human interaction

You may have noticed that in the SAE’s terminology there is a use of the word “automated” instead of “autonomous”. That’s because the meaning of the word autonomy is more complex than the electromechanical point of view and it could give you the idea of a self-aware vehicle that is smart enough to do whatever they want, for example, to change the destination.

Autonomous and self-driving are two names usually used interchangeably but they do have a difference between one and another. The difference is in the level of human interaction required - self-driving cars are considered Level 3 or Level 4 where they are capable of driving themselves but still need human supervision and are subject to geofencing - alerting the driver about going out of the planned area. The Level 5 fully autonomous cars can go anywhere.

We have written about some downsides and concerns regarding autonomous cars in our previous article on this topic. Some of those are still relevant, and here are some more. This is why autonomous cars aren’t cruising the streets yet.

Autonomous cars use LiDAR, which acts as a so-called eye of self-driving vehicles. It provides them a 360-degree view of the surrounding helping them to drive themselves safely. The problem with LiDAR is that it’s quite expensive and there is a problem with the balance between range and resolution. It is unknown what will happen if multiple autonomous vehicles drive the same road simultaneously.

Another thing - we are certain that in good weather autonomous cars will act predictable thanks to a good view of road signs and lane dividers. But what about bad weather conditions? Snow and any kind of soiling such as oil, debris or whatever trucks may drop from their loads - they will obscure the street markings for the sensors and cameras. There is no solution for that either.

Some more things obvious for cars controlled by humans, but unknown in the case of autonomous vehicles: how will they perform in tunnels or crossing a bridge, what happens in heavy traffic and how will they share the road with traditional cars?

Lastly, there is a big question regarding the laws and regulations, especially between states. How will those apply to autonomous cars and will they differ in other states. If yes, will crossing states with autonomous cars be even possible? Maybe there will be a new option that would allow the cars to recognize the state’s rules?

That’s a lot of challenges that autonomous cars have to face before they enter the market at all. On a better note, there are benefits as well. Firstly, there is an unlimited amount of ways in which autonomous vehicles could be helpful to humans, they are just like robots but they can drive themselves to another destination.

We wouldn’t be ourselves if we didn’t mention the positive influence on lowering CO2 emissions. If autonomous cars become reality, zero-emission vehicles will drastically lower pollution, and that’s why it’s important to keep searching for answers to the most intriguing aspects of autonomous cars.

Our previous article about autonomous cars: ​https://zeremi.com/articles/autocars
Source: synopsys.com