What are Self-Driving Cars?

Written by Yuvraj Dhillon

With the growing popularity of automation, humans are getting more and more inclined towards Artificial Intelligence that can easily learn to navigate complex paths without human intervention.

A self-driving car or an autonomous vehicle is a vehicle that is capable of sensing the environment and moving on its own with little to no human input. This means that it does not require humans to steer the wheel, apply the brakes, or even press the accelerator. In theory, a fully autonomous vehicle would only need the input of the destination, and it would reach the instructed destination while the person could be sitting in the car taking a nap, texting or watching a movie.

But how can the car accomplish this? Well, most of the cars attempting full autonomy use LIDAR (Light Detection and Ranging) system, which is very similar to the SONAR(Sound Navigation and Ranging). SONAR uses ultrasonic sounds to calculate the depth of the ocean floor and in the same way, LIDAR uses light waves to identify the dimensions of the object. Thus, the car’s software can then take that data, calculate the distance of the object from the car, and identify the object in order to successfully steer around it.

Okay, imagine this. You're inside a moving self-driving car, on a road in the city, and suddenly, you see a person standing in front of the car. The LIDAR system would automatically recognize that object as a human using the data obtained from the light waves, and that picture would prompt the software to steer on a trajectory that would enable it to safely navigate past the person.

The problem with the LIDAR system, however, is that it’s way too expensive. Companies like Uber and Waymo (a subsidiary of Google), have been testing the LIDAR system. It has proven to be effective, but the cost still makes it tougher for self-driving cars to go mainstream. Therefore, Tesla, another automotive company, introduced cameras around the front and rear, instead of LIDAR, which makes it cost-effective and viable. It's just as effective as the LIDAR, if not better. By having the cameras provide images from multiple angles, the car can easily detect the object in front of it.

“We will have more than one million robo-taxis on the road, A year from now, we’ll have over a million cars with full self-driving, software... everything.” -Elon Musk

So the only challenge in building a truly autonomous self-driving car would be that there wouldn't be enough data for the software to identify objects like humans, animals, trees, etc. Luckily, the programming that Tesla puts into its software is like an AI that’s learning every day, and in another 2 or 3 years the software inside self-driving cars may have enough data to figure stuff on its own, literally requiring no human intervention. And this is truly amazing, as that would prevent so many road accidents from ever taking place. then The road drives would be way more comfortable, safer and enjoyable. Also, robotic taxis might pave the way for cheaper transportation. Since no human driver would be required, the rental taxi might pick up and drop customers on its own, reducing the logical need to ever even buy a car.

Waymo, a subsidiary of Google, has been experimenting with its own self-driving taxis.

Tesla, Waymo, and Uber have not pushed further than a level 2 self-driving car, meaning that a human still needs to be on the steering wheel and needs to pay attention to prevent the car from going on the wrong track. There will be mistakes, and there will be failures. But without trying, humans can never get to level 5 of full automation. So only a tad bit more patience is needed. Tesla has been regularly updating its software, and this technology is not far away. The future of travel is self-driving cars, and if you really want to watch Netflix while on a road trip without worrying about the road, you gotta wait for a little!


Copyright © 2019 by Techvik.