We need Self Driving cars NOW

An explanation on why we need self driving cars.

How we drive today

A street is only cement with printed colored lines, it is only a promise. Depending on the country, cars drive on different sides of the road, make turns on different signs and so on. Driving isn’t a simple drawing a straight line from point A to point B, it is more complicated, requiring understanding different kinds of roads, the optimal time, route with minimum number of stops, or different prices depending on toll booths.

Ever since cars were invented, roads became better, wider, and highways were built. Transportation is the essential building block in developing the modern world. Goods are moved around using trains and trucks, people are able to l commute to work far away. Now making a faster car is not important anymore, the maximum speed of cars haven’t seen much difference in the last few decades. Now it is time for something revolutionary, but flying cars isn’t something around the corner for everyone just yet. But self driving cars are.

Why do we need self driving vehicles?

The main benefits for drivers are not having to repeat repetitive task of driving the same route every day and having better safety. As of 2016, when I am writing this, safety of autonomous cars are not guaranteed. Uber has recently been banned from driving autonomous driving cars in California after driving through some red lights. Not having to drive will reshape what we think of cars. There will no longer be a driving wheels, the concept of drunk driving be gone. No police officers checking parking tickets, or checking for speeding cars.

On a larger scale, there are more reasons than just a better driving experience. First is energy consumption. This statement might confuse many readers, the reason is cars consume different amount of gas depending on driving habits. Going continuous on the highway, starting to accelerate in a continuous pace consumes the minimum amount of gas. The fact is, not everyone drives that way. Next, traffic reduction. This is a great video so please watch [link]. The bottom line of the video is human’s bad reflexes are the cause of traffic. When a traffic signal changes to green, there is a lag between the car leaves and the car behind starts to move. There is a chain reaction which accumulates and comes to the point where, cars in the back will move, and stop to a red light even before getting there.

Ok, so how is it done?

So how do you teach a car to drive by itself? Well, it isn’t to hard. It boils down to very simple tasks. First and foremost, lane keeping. Most of the time in the car is just going straight. How do you teach that? Just tell the car to go between two lines and keep some distances from the car in front. That simple! Then the next step is to change lanes. Last, make left and right turns. There are other logistics such as stopping and reading signs, but that happens mostly while going straight, so I will add that as part of going straight.

Autonomous vs Centralized

There are actually two different types of self driving cars.

Autonomous

You might have heard a lot about autonomous cars recently. Tesla is one popular example. They produce cars with software which makes the cars autonomous and drives on its own. Currently, the software is a bit limited, but it is very close to fully functioning. Uber is another company which is expanding on this platform. Simply put autonomous cars mean cars drive on their own, without any dependency of outside source.

Centralized

On the other hand, a centralized system, is where a central system controls the flow. An example would be the subway system. A subway operator does not decide on the schedule of the train, it is predefined, the operator only operates it according to the plan. In the same sense, in this ideal world, cars will be managed by a central system.

I personally believe this is the most ideal method of self driving cars. Having a central system, every drive can be determined, giving the best optimal route with a precision of 99.999% accuracy of estimate time of arrival.

However, this is something impossible, due to the lack of joint action among the automobile companies and the government. All the cars have to be on the same system. And there is a problem of who will maintain the joint system. It seems like to much hassle for any company to join this union. Also, some other issues with the centralized way is a single point of failure can be catrastpic. If millions of cars are managed by a few systems, a single bug can shut down a whole city at worst cases.

Hybrid System

I think there can be a compromise in the middle which has the benefits of autonomous vehicles but also inherit a bit of the centralized method’s benefits, which is a hybrid. Its car-to-car communication. Recently, I have been very interested in mesh networks, I believe mesh networks are the future as well. A mesh network, is a system where cars talk to each other directly, if there is a car accident car can spread the information to surrounding cars, letting them know to slow down beforehand.

This hybrid system is more realistic, for only car manufacturer can create a uniform protocol to communicate. If companies decide on going their separate ways in shaping the future of cars, it would only get slightly better than what it is today.