How do autonomous vehicles work?

We have been seeing driverless cars and flying cars in various science-fiction films for over half a century. However, until now, this was seen merely as an unrealistic fantasy that might, at best, become a reality in eons. A few years ago, we could scarcely have imagined that self-driving cars were an invention just round the corner.

With technology, the line between science fiction and reality is becoming increasingly blurred!

Autonomous or robotic vehicles are those that do not require human intervention to be able to navigate a map and reach a destination without any mishaps. So far, no autonomous vehicle has been officially launched on the market, but companies such as Tesla and Google are working to make this a reality sooner rather than later.

How does an autonomous vehicle work?

Artificial intelligence (AI) is the foundation of an autonomous vehicle’s independence. Through techniques such as machine learning and neural networks, developers can create autonomous driving systems by analysing vast quantities of images. 

Neural networks detect patterns in the data which will be incorporated into the machine learning algorithms. These will be used to train the autonomous driving system to identify objects such as people, bicycles, animals, road signs, physical obstacles, crash barriers, etc.

Here’s how we envisage an autonomous car might work:

• The lead passenger in the car chooses a destination and the vehicle’s software calculates the best route.
• A rotating lidar sensor mounted on top of the car scans the surrounding area within a range of 60 metres and generates a dynamic 3D map around the car.
• The sensor on the rear right-hand wheel monitors lateral movement to position the car on the three-dimensional map.
• They are fitted to the front and rear bumpers radar systems that calculate distance with any potential obstacles.
• The AI software connects to all the sensors and collects information from Google Street View and the in-car cameras.
• It will be AI that make the decisions which a human driver might take following extensive deep learning training.
• The autonomous vehicle’s software will consult with Google Maps the presence of points of interest, road signs and traffic lights along the route you have set.
• There would be a function allowing a human driver to take control of the vehicle at any time if they so wished.

Waymo: the autonomous vehicle company

Back in 2009, Google decided to invest in the development of autonomous driving with a project called Google Self-Driving Car Project (Google’s Self-Driving Car Project). It gained such prominence that, in 2016, the project became a technology start-up under its current name, Waymo.

The Waymo vehicle is currently in the testing phase, specifically as a taxi (known as a ‘robotaxi’). In this highly ambitious project, they have had to combine a vast amount of data collected from sensors, lidars and cameras so that the vehicle can independently identify, in a fraction of a second, all the relevant elements around it and predict their possible movements.

The most amazing thing about the deep learning techniques Waymo is using is that the more one of its taxis drives, the more it learns and improves.

Advantages and disadvantages of automatic cars

Autonomous vehicles have their staunch supporters, including many tech giants keen to develop a groundbreaking product, but they also have their critics. And this is hardly surprising, given that the concept of autonomous driving must overcome a number of challenges before it can be considered entirely safe.

Advantages

Some of the benefits of autonomous driving are:

Reduction in accidents

One of the major milestones that companies such as Waymo are aiming for is to drastically reduce road accidents. To this end, self-driving cars would be fitted with sensors and advanced systems that enable them to avoid collisions and make safer decisions on the road, thereby minimising human error.

Greater efficiency and smoother traffic flow 

One of its major advantages is its ability to link different sources of information, including the cloud. Autonomous vehicles could communicate with one another and with the road infrastructure, thereby facilitating traffic coordination and reducing congestion on roads and in cities. 

Access to transport for people with disabilities 

With self-driving cars, society could make it possible for people with visual and hearing impairments to become part of the world of driving. Self-driving cars are therefore more accessible, as they offer greater autonomy to people with visual or mobility impairments, providing them with greater independence and access to transport.

Time-saving and convenience

Drivers would also become passengers, so there would be no need to constantly keep an eye on what is happening on the road. As the driver’s constant attention is not required, passengers can use the journey time to work, rest or entertain themselves.

Lower environmental impact

Autonomous cars can drive more efficiently, which can lead to a reduction in fuel consumption and, consequently, lower emissions of polluting gases. The development of autonomous driving goes hand in hand with that of new eco-sustainable solutions such as electric or hydrogen-powered cars. Furthermore, when selecting the route to take to reach a destination, they would always choose the shortest and least polluting route.

Disadvantages

High cost

Like all disruptive technology that comes onto the market, self-driving cars won’t exactly be the cheapest on the market. 

At first, very few people will be able to afford the first prototypes of self-driving cars to come onto the market. This limits their accessibility for many consumers. However, over the years, their price is likely to fall.

Dependence on technology 

Autonomous cars rely on highly complex and sophisticated systems, which means they are vulnerable to technical faults or cyber-attacks that could jeopardise passenger safety. 

This is particularly because autonomous vehicles would rely on a continuous Wi-Fi connection, just like any other computer, through which cyber attackers could attempt to take control of the car.

Infrastructure adaptation

For efficient implementation, changes to the road infrastructure and greater investment in the installation of sensors and supporting technology would be required. 

We need to rethink how autonomous vehicles coexist on the roads with conventional vehicles, and amend road safety regulations to take account of this new form of driving.

Ethical dilemmas

Although self-driving cars are programmed to prioritise safety, they may face situations where the decision to save one group of people means putting others at risk. 

Resolving these ethical dilemmas is a complex challenge that may exceed the capabilities of the deep learning which aims to simulate the human mind, without actually possessing a comparable capacity for autonomy.

Job losses

As autonomous vehicles become more common, they could replace jobs held by professional drivers, such as bus drivers and taxi drivers. 

And if autonomous driving extends beyond cars to include other vehicles such as trains, lorries and underground trains, the future of professions such as train drivers, hauliers and pilots would be left in the balance.

Where can I buy a self-driving car?

For now, You cannot buy any fully autonomous vehicles, but within a couple of years you’re likely to see this acclaimed invention make its debut at the most forward-thinking dealerships.

However, we are already seeing some features that could be seamlessly integrated into autonomous vehicles. For example, blind-spot monitors, emergency braking systems, automatic gear changes and other driver-assistance systems have already found their way into some cars – and there are more to come!