The past, present and future of autonomous cars.
It’s truth that in the last few years, self-driving cars have been a huge innovation in the field of driving, many companies have started the race to built this dream, but is still really new, we haven’t found the right model to follow yet. There have been a lot issues related to the launch of cars which can really self drive, mostly in human safety. In this article I’ll talk about the past, the present and the future of self-driving cars from my point of view and how this innovation improve human transportation.
The beginning of autonomous car
According to Bel Geddes predictions of Futurama Expo in 1939, in the future, aside from the multiple innovations that today exists, there will be systems to prevent car crashes. This systems will enable communication V2V (vehicle to vehicle) to send and receive messages so it can coordinate the transit, and as it sounds, it’s amazing, also in this time.
We need to remember that this was before the WWII and it was a time when most of the people didn’t have a car, so they couldn’t imagine how it could be. We can see this video from the expo:
After this, many companies and universities tried to create the autonomous car, including Mercedes Benz with the collaboration of Ernst Dickmanns (this name is going to sound a lot) and his team from Múnich University. Worth to know that Ernst Dickmanns is one of the pioneers of the computer vision applied to dynamics and robotics for self-driving.
Thenceforth, many universities and companies in the world started a new study field: computer vision for self-driving, but as far as we know we have a couple of good examples of how this work in the real commutes for the normal people like Mercedes Benz, BMW and Tesla according to U.S. News. Because we have a big problem… humans.
The hardest problem: human
The system behind an autonomous car is really complex. We have many sensors which are getting data from outside and we need a powerful computer in order to process all the information in short time. The way the computer have control over the entire system is because the programming have been done in a totally different way: machine learning.
And if we type on Google, “machine learning” is probable to have million of results which may confuse us, so we can summarize it in the ability to detect patterns in the information, predict results and act as a human would do. But, the human thinking is complex and strange, isn’t it? We all have similar common sense, but different bias about the facts, so we can decide over thousands of millions of possibilities in the way we learned in our lifes. So impartiality is quite difficult.
The people behind this projects, train the cars computers with machine learning algorithms how to generalize decisions and how to act with the highest precision. Learning from the behavior of the pedestrians and drivers in order to reduce car crashes, to respect traffic lights, the road, and also with unforeseen scenarios, detecting people and everything around. I won’t deepen this topic, but we need to keep in mind the challenges with the machine learning because is related with artificial intelligence.
Now, why is the human the hardest problem? Let’s think about it, if all the cars were driverless and could communicate between them, there wouldn’t be any problem, they would anticipate problems and follow the rules, the thing is that many cars nowadays are still driven by people who has an unpredictable behavior so is really challenging how to get the fastest answer before any accident happen.
Moreover, the present reality of this technology is that is really expensive to get to the masses, so there is work and train in process to achieve the 1939 dream.
The Future
We made a tour by the history and facts for the self-driving cars, but what is next?.
Personally in matter of technology, I think we are not far from reaching the point of artificial intelligence that could deeply understand the interactions of driving, pedestrians and passengers at the same time and could take the hardest decisions before any accident.
In terms of purchasing power, I think we have a gap from people that could not afford this. If the tech evolves as good as it is going right now and the prices decreases, in a couple of years many of us could buy a full self-driving car, and then, we may reach the vision of the greatest minds that thought this was possible.
NOTE: This article is pure academic.
References
- Learn more from Erns Dickmanns at http://ieeetvdev.ieee.org/history/robotics-history-narratives-and-networks-oral-histories-ernst-dickmanns?
- US NEWS report about self driving cars: https://cars.usnews.com/cars-trucks/cars-that-are-almost-self-driving
- The toughest challenge for self-driving cars? Human drivers. https://www.foxnews.com/auto/the-toughest-challenge-for-self-driving-cars-human-drivers