Can artificial intelligence make self-driving cars a reality? Things to know before learning AI

Suddenly, can you remember what Japanese companies looked like 30 years ago?

It was said that “Japan is number one”, and many Japanese products had a strong presence in the world.

Today, not only American companies but also Chinese companies have emerged and are leading the world. On the other hand, it is a sad sight that not a single Japanese company has been ranked in the top 10 by market capitalization in the world since 2018.

Furthermore, with economic development, the speed of change in the world is accelerating, and there are more opportunities to seize them, such as an increase in the number of companies that stand out in the world in just a few years.

For example, we have achieved great innovation by realizing a future that was a fantasy 20 years ago, such as “people all over the world use PCs as a matter of course.”

This kind of innovation cannot be realized without “magnificent delusion” and “understanding the current issues and accumulating solutions”.

In this article, we will capture the current AI challenges with the theme of self-driving cars, which many people fantasize about as the future.

I would be happy if the number of people involved in creating a new era would increase through this article.

When will self-driving cars come true?

Tesla’s Elon Musk predicts 2020

Do you know who Elon Musk is?

He is a “genius” who has continuously started and managed superhuman businesses.

His business scale and ideas attract many people, and his remarks always attract attention.

Acchan’s Elon Mask explanation video is very easy to understand, so please take a look.

I will introduce it briefly.

Mr. Elon Musk aims to use a rocket called SpaceX to move humans to Mars for the continuation of mankind.

Such a genius, Mr. Elon Musk, became a hot topic when he said, “Completely automated driving will be realized in 2020.”

Fully automated driving means that even if you fall asleep in the driver’s seat, you will be able to reach your destination (and even park!).

There is no doubt about Elon Musk’s achievements so far, but many of what he has said so far have been met with skepticism.

Is it possible to achieve this “completely self-driving car realization” in 2020?

SoftBank CEO Masayoshi Son predicts 2035

At the financial results briefing for the fiscal year ending March 31, 2019, SoftBank Corp. President Masayoshi Son predicted that AI-based self-driving cars would become widespread in 2035.

The SoftBank Group has invested in the world’s top companies such as Uber and Alibaba, and many IT companies that are at the core of autonomous driving technology such as Nvidia.

Therefore, the ability to see the future may be higher than that of ordinary companies.

In the future, if the spread of communication technology and IoT infrastructure, and the increase in computational resources, etc., the technology will continue to improve at an accelerated pace.

Can AI Realize Self-Driving Cars?

Many people are looking forward to a bright future in the self-driving car society, but is it even possible?

[Companies are also shrinking and withdrawing] Huge funds are needed to develop self-driving cars

Self-driving technology requires a huge amount of money to develop.

This is because it is extremely technically difficult to develop a car as hardware.

Until now, many companies that have been developing self-driving technology have downsized or withdrawn.

It can be said that it is a race against time to clear technical issues in autonomous driving or run out of funds.

95% of humans are unconscious! ? Difficulty in reproducing with AI

 Driving a car includes many unconscious elements, and it is not easy to reproduce them with AI.

It is said that more than 95% of human behavior is unconscious. A good example of unconsciousness is breathing.

We rarely pay attention to our breathing except when we meditate or do yoga.

Cognitive science shows that people who act unconsciously verbalize events afterward.

For example, “I saw a child riding a bicycle, so I put my foot on the brake and slowed down. ”

Humans make a ton of unverbalized decisions every day.

At present, it is extremely difficult to reproduce the unconscious processing speed of humans with deep learning.

AI programs output the closest optimal solution according to the given data, but humans sometimes do not make rational decisions.

For example, even if you usually take the shortest route to work, do you ever feel like taking a detour?

Such human decision-making is like “noise” for programs seeking optimization, and programs cannot easily cope with large amounts of noise in the real world.

Airplanes are already self-driving.

The big difference between cars and airplanes is the amount of noise and complexity of “sky” and “ground”, so it can be said that there is a difference in the practical application of autonomous driving.

The development of autonomous driving technology is not always superior to major companies

Japanese companies such as Toyota and Honda are also developing autonomous driving technology, and are collaborating with software companies while leveraging their knowledge of hardware.

However, Google’s Waymo and Tesla are leading the existing car companies.

We take it for granted that we charge our smartphones and computers with chargers.

A self-driving car is like a big computer that recognizes pedestrians in real-time, recognizes space, and parks.

In other words, I think that you can somehow imagine that it is very compatible with electric vehicles. And the design of electric vehicles is fundamentally different from that of gasoline-powered vehicles.

Therefore, the technical knowledge of Toyota, the world’s top automobile company, does not necessarily give it a competitive edge.

Moreover, Toyota Motor Corporation is a major supporter of Japan’s manufacturing industry and, by extension, the Japanese economy, making it difficult to shut down business with existing manufacturers.

In that respect, startups like Tesla, which have nothing to lose, are moving faster than incumbents toward self-driving technology and high-level electric vehicles.

In other words, even start-up companies can reverse the trend in autonomous driving.

The current state of startup Tesla

Tesla is a start-up company, but it is growing fast enough to catch up with Honda.

Despite Tesla’s car production volume in 2018 being one-fifth of Honda’s, its market capitalization is in line with Honda’s back (over 5 trillion yen).

Technological innovation is the reason for the established theory that restructuring of the banking industry, which was at the top of the employment rankings ten years ago, has expanded and will disappear from the employment rankings in the next ten years.

Of course, even big companies can innovate. However, it is also true that there are many times when there is too much to lose and it is difficult to move.

And startups see it as an opportunity and have a chance to turn it around.

We shouldn’t set our goals in life to be “surpassing big companies” or “getting rich”, but the real thrill of technological innovation is that it can have such an impact on society.

The next video is that Tesla car predicting an accident and giving a warning sound.

As far as I see this, I feel that it has already exceeded human ability.

In this way, autonomous driving technology has already come to the point where it supports our human capabilities.

As technology becomes more sophisticated, it will gradually move away from human hands.

Summary: Problem-solving is important for AI to realize self-driving cars

While self-driving technology has overcome some difficult challenges, the challenges remaining to its full implementation are not trivial.

Elon Musk is undoubtedly one of the world’s rare entrepreneurial geniuses.

However, it is very important for us not to just say “yes” to his remarks without thinking, but to think about the content of “Why is it feasible?” is.

At the same time, it is very exciting to be in the arena to create a company that will surpass a giant company that earns trillions of yen in a few years by investing in dream technology.

To think about the contents of new technology with your head, it is important to move your hands and experience it.

Elon Musk, who is making the world’s leading space business and electric vehicle business, seems to be based on what he “learned in books” rather than studying them at university.

Startups indeed have difficult aspects, such as the problem of data volume, to create a global company in the AI ​​industry.

On the other hand, there may be great opportunities hidden in those challenges.

In an era when IT and AI will become the foundation of our society, we recommend that those who want to create and create an exciting future should study programming.