Tesla’s Autonomous Driving
Tesla has been the center of attraction in the automobile market for the past decade. Tesla started with a roadster which was nothing but a rebranded electric version of a Lotus car. But everything changed for Tesla when they launched the model S in the market. Tesla’s Model S was an immediate success. It came with features that were very advanced at that time. One of the main selling points of the then-model S was the connected car technology. Many other automobile manufacturers were also impressed by what Tesla had to offer, and they also try to implement their version of those features in their cars. One of the Industry changing, and industry’s first features were autonomous driving. Autonomous driving is something that everyone has been dreaming of for a while now. I remember watching an Audi drive on its own in the movie I-Robot. Ever since the three-year-old me saw that car, I always dreamed of owning a car that would drive on its own, while the driver would not have to worry about the surroundings or anything.
Before diving deep into how Tesla implemented autonomous driving let’s first talk about what autonomous driving is. So autonomous driving can be referred to as a car driving on its own without the assistance of the driver. At present, there are three stages or three types of autonomous driving. The first stage or the first type of autonomous driving only focuses on keeping the car in a straight line or in between the lane. The second stage or the second type of autonomous, driving focuses on keeping the car in a straight line or between the lane as well as taking turns and doing proper cornering. The third stage of the final type of autonomous driving focuses on all the other aspects mentioned above as well as also can learn from the surroundings and using this summon feature that will allow the car to come to its owner comes to us pinpoint the location specified. You will be surprised to know that the very first autonomous car was introduced back in 1925. “While the rise of autonomous cars has gained prominence over the past two decades, it is true that its beginnings date back to the 1990s. In 1925, Francis Houdina, an electrical engineer from New York, was the first to implement the concept of an autonomous vehicle, although the car was remotely controlled.” (Source: BBVA)
Autonomous Driving uses a very common technique to identify objects and make judgments based on the resulting outcome that is being fed to it. It uses various kinds of object recognition technology that uses the data already being fed to it, while training that model, and collects live data while you are using them. It then processes those data using Machine Learning models and then processes the data to do the follow-up command. Having worked on a similar model, I can tell the complexity of these models is very complex and this must be complex just to make sure that the user can use it effectively without any problem. These object recognitions are divided into various sub-categories so that the processor can handle the processes in a very efficient way. To keep it simple, let’s assume that there is a processor with 4 cores, One core is responsible for the recognition of all the car activities around, the other is responsible for the activities related to humans, and so on.
Tesla started the self-driving car or autonomous driving in 2015 but came up with a “Full Self Driving car” in 2020 (Source: Euro News). The early adaptation of Tesla’s autonomous driving was limited to Lane-assist, cruise control (the most basic or the initial form of autonomous driving), and automatic parking. At present, it does all these but better and can drive and steer on its own. It can also be called at your doorstep with a summon feature, which allows the car to take full control, without any driver inside of it. With the introduction of a Yoke steering wheel, Tesla is pushing more people to opt for the Full Self Driving mode. The company’s motive behind this move is clear and shows the confidence that they have but enforcing a feature on users may be wrong to some extent.
Autonomous driving in Tesla does not just provide ease and comfort for the user but also makes it safer because of the emergency braking, driver drowsiness alert, and less margin of error when compared to a human, because human eyes may not see everything in certain cases like a night or fog, but the sensor can sense and make more clear judgments based on these calculations. These features can be life-changing and are no doubt a very human-centric approach. But there is no way to verify the functionality of these features because these technologies are still in their early stages. Almost a decade back, companies were testing out autonomous driving with computers filled all over the car and big sensors popping outside of the car. But now these things are so refined and compact that it is not even visible to us. In fact, a processing unit for an ADAS chip is the same size as a mobile chip and you can build your own version of ADAS, which although will not work on an actual car, the processing unit can be made and tested out on your phone if you happen to have a phone with depth sensing (LiDAR in newer iPhones)
It will be interesting to see how Tesla will move forward with these technologies and must make sure that they continue to be safer for human use. There have been reports of this Autonomous driving not working properly on a few models, which include over-speeding, which was reported recently in China, where 2 people lost their lives. These technologies are evolving and will get cheaper with time as more people start using them. Although the cheaper version will be limited to lesser options but hopefully with time, Stage 3 will become Stage 1, and Stage 3 will have more advanced features of Autonomous Driving.
