Bye Bye, Driving License
By Aditi Bhat
Take a step back into the memory lane and remember your driving lessons. Remember how often you used to forget to switch the gear before accelerating your car? Experienced drivers get so used to this that their minds go on auto pilot mode. If you were to pay attention, you would be amazed at how many things you’re taking into account when you hit the brakes, change the gear, or go through the traffic, be it automobiles, pedestrians or cows for that matter.
Yet, even with a driving license (which you get after hours of driving lessons), it’s no secret that humans, collectively, aren’t great drivers. We break traffic rules, fly into road rage, crash into one another and even hit stationary objects.
Self-driving cars are here to change the world and maybe even make it a safer place
Driverless cars are not bound by human limitations. They don't yell at other drivers, don’t experience driver fatigue or get distracted by anything shiny. They are only concerned about getting safely from one point to the other.
Self-driving cars need to be programmed to sense everything a human can, and use that information to make decisions like we do. The car has to collect data about its surroundings while maintaining routing information. It should ‘experience’ how these surroundings may change. This is where data science comes in.
Data science as an idea is simpler than in execution. It aims to use information to make decisions, similar to how you do while you drive. The complexity comes in when trying to program a car to think like a human. It requires breaking down decision making into the smallest possible pieces of sensory input. Each of these inputs then needs to be wired to sensors, which specialize in gathering that type of data.
Once that data is gathered, it needs to be processed to create a map of the immediate surroundings. Once this map is created, the decision making begins. What is the long term goal? i.e. “Where do I need to get to?” What is the short term goal? i.e. “What immediate action do I need to take?”
After the decision is made, action needs to be taken. If the car needs to take a turn, it has to slow down, turn the steering wheel, and slowly speed up again. This entire series of algorithms has to run repeatedly at a lightning speed to factor in everything that happens on the road to react to it. That’s the power of data science.
Although prototypes for self-driving cars have been attempted since the 1920s, technology giants like Tesla Motors, Google, and Uber are the closest to creating commercially viable options. Uber has already changed the game for the taxi industry twice over, including the ability to provide cabs without drivers.
From a consumer perspective, aside from convenience factors like not having to worry about parking, or eating fries while driving and being able to spend your commute time doing something other than driving, autonomous cars have the potential to improve road safety. Driver distractions and drunk driving can be significantly reduced. People with disabilities would no longer be dependent on others or on public transport. Over time, we could eliminate road signs, speed limits and traffic flow could be improved, and we would no longer require a driving license!
Before we get ahead of ourselves, let’s not forget that these cars are built for driving conditions in the US. The real test of self-driving cars will be when they can drive on roads in developing countries like India. We’ve got a long way to go, but here’s hoping we won’t need driving licenses by 2050!