Content provided by WeRide.
Deloitte has launched a brand-new podcast series Future of Mobility in 2020, bringing us key insights from decision-makers, innovators and shapers pushing the envelope on future ideas for transportation and beyond.
On Sep 1st, 2020, in the latest edition themed “Is the Future of Autonomous Driving Here”, Tony Han, Chief Executive Officer and Founder of WeRide, China’s leading autonomous driving company, as well as Niels de Boer, program director at CETRAN Nanyang Technological University of Singapore, joined Deloitte to discuss the possibilities and application of self-driving technology.
The 40-minute interview generated quite a few insightful ideas, helping us better understand the present and future of self-driving in a broader way.
The following is the highlights and summary of Tony Han’s interview in the podcast.
Go Along with Unpredictability
Keep Safety as Top Priority
Tony Han: The essential problem we are trying to solve in autonomous driving is safety. Unpredictability is the nature of human behavior. There’s the so called skirt in the round. There are always some people who are totally out of the normal distribution curve, the outliers. Some of them are really genius. That’s the nature of the world. Even as human being ourselves, it is still difficult for us to predict or project what others will do. The essential question for self-driving is safety. However we don’t need to predict all the behaviors to make the self-driving travel safe. The first step for our real operation is to make sure our cars don’t hit anything.
Regarding this aspect, autonomous driving has very big advantages. Take WeRide’s sensor suite for example, it’s integrated with 1 HD LiDAR, 6 peripheral cameras, 3 long-range cameras and 5 electronic scanning Radars. With all of these, their sensor power is synchronized by a specific hardware. Those sensors are triggered to a precision of milliseconds. We can achieve centimeter level resolution within 200 meters. This is far beyond the capability of human eyes. Therefore, we can see all of the obstacles and even when some unpredicted behaviors happen, the vehicle can safely stop in time. Our self-driving car may not seem drive as skillfully as an experienced human driver would, but it should be safer than human drivers. That’s the beauty of autonomous driving.
5G Application Is Another Indispensable Layer of Redundancy
It’s Also an Economic Problem
Tony Han: First of all, 5G is very very expensive. Let me tell you some statistics. The base station range of 4G is roughly 1 kilometer. While the range for 5G base station is 300 meters. That means the radius of 5G base station is about 1/3 of that of 4G. We know that the area of circle is π *r2. That means any area you cover by one 4G base station probably need 9 or 10 times of 5G base stations.
For very populated and very advanced areas like Singapore, Beijing, Guangzhou, Chicago or San Francisco, that’s made doable. What if in the relatively deserted and underdeveloped regions, will we invest that amount of money to use 5G to cover the whole area? It’s very challenging.
I think we do need to cover the whole area with 5G network in big cities, because 5G and V2X (vehicle to everything) application could significantly improve efficiency and add another layer of redundancy for self-driving. But we need to balance between the benefits 5G network and the huge cost of building it. This is an economic problem. We need to take a suitable position among this whole curve, and find the most optimal operation point to do that. We will definitely benefit a lot from 5G network, but we need to find the most cost-effective way to do so.
Keep Pushing the Technology
AD Will Become an Affordable Service
Decreasing Hardware Price VS. Increasing Human Cost
Tony Han: I’m very optimistic about the price and cost of self-driving service. In 2017, the price of a 64-line LiDAR was roughly 750,000 yuan, and in 2018, it had dropped to roughly 250,000 or 300,000 yuan. The hardware price drops drastically every year.
On the other hand, the cost for human labor is increasing. In developed countries like Japan, Korea and Singapore, they are facing the same problem of aging societies. There are not enough young people to drive taxis or buses. Therefore, the human labor cost of driving service is growing every year. The decreasing of hardware cost and increasing of labor cost is a growing profit margin for self-driving service. We can foresee a nice business model out of it, so we need to focus on continuously improving our technology.
There Are More Problems Than Available Resources, So We Should Collaborate
Tony Han: No company can build out autonomous driving vehicles all on its own. There are more problems than available engineering resources or technological resources can solve. Therefore, we should collaborate.
Even for giants like Google and Apple, it is impossible for them to produce everything from LiDAR to camera, from CPU to GPU altogether by themselves. We have to collaborate, and that’s the beauty of industrialization.
But on the other hand, in autonomous driving industry, the competition is very intensive. For example, Waymo and Cruise, would they be able to collaborate? Currently the likelihood is very very low. But would Waymo be collaborating with Nissan? Very likely. So we need to promote collaboration in the value chain.
In my opinion, the influence of capital is the most essential element. The capital has to get involved to foster this kind of collaboration. Previously, we had big giants like Softbank Vision Fund. They actually helped integrate different kinds of companies together to develop autonomous driving technology.
Build Trust and Confidence through Safe and Publicly Accessible Operation
Tony Han: There are many factors affecting public views towards self-driving. It’s a very complicated social psychological problem. One accident may change the view of the whole society.
Every leading company in autonomous driving industry has responsibility to try their best to keep the safety not only during operation but during technology development.
WeRide has been publishing Robotaxi operation reports to the public, and our statistics directly showcase that with so many miles our fleet have covered, so many rides we have accomplished, there’s no accident involved. That’s an excellent record. What we need to do next is to convince more people to take a ride.
Let me explain it with the example of a human commercial plane. An airplane is just like a barrel of steel. You put 200 people in it, and there are some combustion engines that keep on doing controlled explosions so as to change the energy from gasoline to push power. You put people in the air and then land them on the ground. That’s scary. But do you feel extremely scary when you board a plane at the airport? Probably not.
It is the same with advocating the commercialization of self-driving service. First of all, we have to prove safety from basic science and engineering perspective. Secondly let people try it gradually. The first time, you would be nervous, but if we give you 1,000 trips safely, probably you would feel it is like just every other daily trip.
We have to do this gradually and very patiently. Starting from low speed and gradually to higher speed. Starting from very amendable regions and gradually to more challenging regions. That’s our way to expand the service and convince the public step by step.
Covid-19 Pushes the Engineering Power to Accelerate AD Technology
Tony Han: When Covid-19 first came to our attention, I asked our engineers a question: are we capable of delivering goods with self-driving cars? Sadly our technology was not mature enough to deliver goods by then. Today we may be able to do so. So I asked them again. If such kind of crisis happens again, can we do something for the whole society, even the whole world? The answer is definitely yes. So I told my engineers to go develop our technology at full speed. In the future if this kind of crisis happens again, autonomous driving technology can definitely help.
During pandemic like this, taxi drivers are an important source for the spread of virus. A taxi driver could be affected and then infect passengers. The passengers could become a new source and transmit the virus to others. But if we use machines, between the trips, we finish one trip for passenger A and have the vehicle sterilized and then pick up passenger B. It could cut off this kind of spread.
Together, we should accelerate the development of autonomous driving technology.
What Space X is doing is trying to help us explore the universe. What we are trying to do is making transportation safer and more efficient on this beautiful blue planet. If this kind of crisis happens again, we are able to do something to help stop the spread of virus.
Develop AD Technology
Share It to All Humankind
Tony Han: Thomas Adison invented the lightbulbs. Not only Americans but all the humankind use lightbulbs. We develop autonomous driving technology with the hope that all humankind, every single person in every country can use our technology. I believe that WeRide has the chance to expand our service to South Asian countries in the future.
Importing autonomous driving technology requires standards at international level. All the parties involved need to reach a consensus beforehand on the following issues. What’s the impact to be made on the society? What kind of regulations and standards should we have? To what extend should we have the autonomous driving mechanism accessible to the government?
All these questions need to be solved before we apply the technology.
If You Have a Chance to Decide the Future Once in Your Lifetime for the Good of All Human and the Society, Why Not Try Your Best?
Tony Han: What is the one thing that gets me up in the morning and excites me about the future of mobility? For this kind of question, my answer is always the same. If you have, in your life, just one chance to define the future, won’t you try all your best, use all of your energy, all your resources to do so? The answer is absolutely yes to me. So every morning when I wake up, what I want to do is turning autonomous driving technology into reality for people to travel more safely. That’s what I need to do.
Introduction to WeRide Robotaxi
WeRide is the frontrunner in China’s Robotaxi industry. WeRide launched China’s first publicly accessible Robotaxi service in Guangzhou in November 2019. Just download WeRide Go app and people can hail a WeRide Robotaxi in an area of 144 square kilometers in Huangpu District and Guangzhou Development District, with no application needed.
WeRide published China’s first Robotaxi operation report in February 2020, showing that a total of 8,396 Robotaxi rides were accomplished in the first month of Robotaxi operation in Guangzhou, China in December 2019. The average daily rides were 270.8, and the fleet accumulated 41,140km of mileage for the finished rides and provided service for 4,683 users.
In June 2020, WeRide’s Robotaxi service became available to the public in Guangzhou through Amap app, one of the most widely used ride hailing apps in China under the partnership between Alibaba’s Amap and WeRide. The cooperation pairs China’s first publicly available Robotaxi service with an innovative mobility aggregator.
In July 2020, WeRide became the first in China and second in the world to test fully driverless cars on open roads, after the company had obtained China’s first remote-control permit for road tests of Intelligent and Connected Vehicles (ICV).
So far, WeRide has accomplished over 900 days of road tests and operation, having served more than 90,000 passengers, and accumulated more than 2,700,000 km of self-driving mileage with zero accident. WeRide is planning to realize fully autonomous driving service to the public in the next 3 to 5 years.
This report offers a look at how China has leveraged artificial intelligence technologies in the battle against COVID-19. It is also available on Amazon Kindle. Along with this report, we also introduced a database covering additional 1428 artificial intelligence solutions from 12 pandemic scenarios.
Click here to find more reports from us.
We know you don’t want to miss any story. Subscribe to our popular Synced Global AI Weekly to get weekly AI updates.