I imagine the day my car will transform into my very own chauffeur. All I will need to do is just hop in, press an on-screen button, sit back, and relax as it takes control to drive me to any destination of my choice. It will stay on its lane and ease between oncoming and parked cars. It will also be on the lookout for pedestrians and traffic lights. And I will be free to read the newspaper, reply to my emails, get a headstart on my job, call family and friends, take my breakfast, watch a movie, watch the news, or simply relax. I will just be spoiled with free time that I wouldn’t otherwise have if my hands were on the wheel.
You see, ever since the self-driving car revolution came into place, I have been keenly following its development. Fully autonomous vehicles are placing better and safer driving outcomes on the table. Reports project safer road systems that will see traffic deaths reduce by more than 2,500 by 2030, according to McKinsey.
How Autonomous Vehicles Work
I have gathered quite a bit of information on how these autonomous technologies work. With zero input from a human being, these futuristic driving machines are equipped with state-of-the-art hardware and software systems, including computer vision, sensor fusion, control, localization, and path planning, and they’re able to sense and navigate the world around them.
Embedded vision systems allow these cars to see the world around them using high-resolution camera images. Sensor fusion systems incorporate data from sensors to form a deep understanding of the environment surrounding them. Localization systems allow them to know their positioning using GPS technology. Path-planning systems enable them to come up with a route or course, and, lastly, control systems allow proper steering and direction.
Driven by Vision
Self-driving cars are designed and built to accurately predict the behaviour of pedestrians, cyclists, and other drivers through advanced sight. Devoid of such advanced vision systems, we might as well forget autonomous cars. Think about it: A driverless car would never know that car in front has braked, or even tell the difference between a tree and a pedestrian without high-tech vision systems. And so it’s true that they have the capability to make the roads much safer. With innovative vision systems, you can bet on the fact that they can’t suffer distracted driving, which is a leading cause of car accidents around the world.
Side, Rear, and Forward Facing Cameras
For a start, autonomous vehicles employ a number of high-resolution video cameras. With a 24 cm baseline and a design that complements other cameras, these cameras are mounted at the front, sides, and the rear to provide the optimal view of objects surrounding the car in a 360-degree range, watching out for any telltale signs on the road ahead, including other vehicles, braking vehicles, crossing pedestrians, traffic lights, cyclists, and more. They work just the same way you do while driving with your eyes on the road. The only slight difference is that these ones have a 360-degree range and multiple angles set up to provide critical information regarding any scene into which the car is driving.
Joining forces with the cameras, LIDAR (light detection and ranging) is another innovation advancing self-driving technology. It’s the box-like structure you see mounted on the roof of a self-driving car spinning at approximately 600 rpm when the car is on the go. LIDAR offers 360 degrees of continuous visibility, providing vision superpowers for any autonomous vehicle. It helps the vehicle map and navigate its environment and know ahead of time that there’s a stop sign or there are traffic lights meters ahead. By firing millions of laser beams every second, LIDAR sensors enable a 3-D visualization environment for the vehicles.
As a result, self-driving cars are able to differentiate objects in their surroundings — say a person walking or a cyclist, their speed, and direction — with irrefutable accuracy. We are talking an accuracy of ±2cm of all objects in close proximity to the vehicle. Remarkable predictability, navigation, and high-resolution object tracking capability truly make LIDAR a critical vision component that helps autonomous vehicles detect and avoid crashes before they happen.
In addition to LIDAR sensors, autonomous vehicles also employ radio direction and ranging (RADAR) technology. Radar sensors encompass radio waves, infrared, and optical antennas positioned on the sides, front, and the rear end of a car to facilitate object detection and tracking. The sensors perform extremely well in thrilling weather conditions and accurately detect and track objects that are about 200m away. They process how fast a car or an object is moving — and in which direction. In fact, many newer vehicles come installed with radio technology referred to as vehicle-to-vehicle technology, allowing them to communicate with each other or with other objects on the road. For self-driving cars, both LIDAR and RADAR sensors greatly complement each another.
For highly performing vision systems, mapping is very imperative. As such, GPS technology is also part of the equation. By incorporating altimeters, tachometers, and gyroscopes for accuracy, GPS fetches the car location, at which point LIDAR, RADAR, and cameras are mounted. Mapping technology empowers self-driving cars to steer when the road conditions are too difficult for sensor navigation. Additionally, GPS helps the car move through the commotion while joining a highway or a traffic circle or crossing a bridge.
Central Processing Unit
So, how are all these vision systems integrated together? Well, captured data and images are quickly transferred to a central computer positioned in the car trunk. The computer processes the data in real-time and matches it with stored data to allow the vehicle to plan, steer, accelerate, control speed, and avoid obstacles.
Would you feel safe in an autonomous car? Some people get hesitant about trying out a new technology. Even further, some people are even doubtful that autonomous cars are actually going to make their way into the mainstream. Personally, I don’t feel any mistrust with self-driving cars. I mean, semi-autonomous cars are already here with us. Guess what the near future holds? Fully autonomous cars.
In fact, their future shines bright, seeing that companies including Uber, Tesla, Google, and more have already released prototype versions of fully self-driving cars. Associated automotive engineers are busy making certain that they are safe and performing as expected through lots of pilot projects and testing. Additionally, experts say that the hardware and software systems that enable these vehicles to reliably and safely drive themselves continuously improving. Globally, the industry is growing by over 15% annually and is estimated to be worth £900 billion by 2025. If you ask me, fully autonomous cars are knocking on our doors, and they will inevitably will make life easier and roads safer.