AutoX Founder and CEO Jianxiong Xiao talks about how AutoX is trying to make AVs affordable and accessible for everyone.
By Hamza ElMokhtar Shili, Keopagnapech Ngoun, and Saven Pat
Have you ever got into your car and felt drowsy on your commute home after a tiring day, or while driving during a long road trip? Feeling sleepy is natural, but nodding off while driving can have disastrous consequences for the driver and passengers alike.
According to data from the U.S. National Highway Traffic Safety Administration (NHTSA), drowsy driving is responsible for at least 100,000 car crashes and over 1,500 deaths per year. Self-driving cars are here to solve this problem.
“One of the biggest benefits of self-driving cars is that the car can drive safer than a human being and reduce accidents,” says Jianxiong Xiao, Founder and CEO of Chinese AV startup AutoX. Xiao is a scientist who is popularly referred to as Professor X by his students. While Marvel comics’ Professor X led an army of superheroes, Xiao leads an army of self-driving cars.
A self-driving car, also called Autonomous Vehicle (AV), is a technologically advanced car that can drive itself from point A to point B without human intervention. With Tesla, one of the pioneers of AV technology, becoming the world’s most valuable automaker, AVs are rapidly gaining popularity.
“At some point in the future, you can completely avoid most of the accidents. That’s the promise of autonomous driving we want to deliver – the promise to keep the roads safer,” says Xiao.
According to data from NHTSA, over 36,500 people died due to car accidents in 2018, with 94% of such accidents occurring due to human negligence. Since AVs have the potential to remove human error from the accident equation, automated driving systems (ADS) can possibly save well over 30,000 lives in the U.S.
A study was conducted to analyze AV accidents reported between September 2014 and January 2017 during a testing phase in California. The study found that the most frequent type of accidents involved a “fender bender” where a conventional vehicle collided with the AV in front of it.
The study found only one instance of an accident that occurred due to an AV colliding into a car in front. The cause of the accident was the error in judgement of the driver who was controlling the AV at that point.
This indicates that AVs can avoid accidents by themselves, but are as yet unable to avoid other cars crashing into them from behind. Moreover, as the researchers concluded, this may mean that most accidents involving AVs that were reported in the given timeline were caused due to human error and not self-driving car systems.
Since 2017, AVs have evolved further, and it is likely that they are safer now than they were before. Manufacturers are constantly striving to correct the vulnerability of AVs to rear accidents. Besides, with sustainability being increasingly important for governments and consumers, AVs have gained more relevance. According to a report by the Ohio University, AVs can reduce harmful emissions by 60% with optimized driving.
The U.S. is not the only country where startups in the driverless technology industry are gaining traction. China has emerged as a top competitor with startups like Nio, Pony.ai, AutoX, WeRide, Momenta, and more. While the startups are bagging enormous amounts of funding, mass adoption of AVs still has a long way to go, especially because of the high prices of AVs. That’s why Alibaba-backed AutoX is using robotaxis and robotrucks to commercialize AVs in China and the U.S.
China’s congested roads make AutoX’s AVs smarter than U.S. rivals
Founded in 2016, AutoX has recently deployed over 100 self-driving cars in China and California. In Shanghai, AutoX’s robotaxi fleet is fully open to the public where customers can book a ride using the Alibaba Amap App.
The startup’s global headquarters is located in Hong Kong. AutoX has raised over $160 million to date, according to Crunchbase data.
While the AV sector is quickly becoming crowded and increasingly competitive, Xiao believes AutoX is at an advantage. Since California has wider roads and a lower population, AVs operating only in California may not be able to handle the complex traffic conditions in densely populated cities like Shanghai and Shenzhen.
“A lot of self-driving car companies, if they only operate in the U.S., it’s hard [for them] to capture the compressed scenario in Asian cities,” says Xiao. “ So it’s very difficult for them to bring the technology to other Asian cities, because in all the Asian cities […] the city traffic is actually close to each other.”
AutoX’s vehicles “have been driving for the past two years in the most crowded and most populated part of Nanchang District, which is similar to Hong Kong,” says Xiao. Ideally, Xiao hopes that giving AutoX’s cars some practice in crowded Chinese cities will make them safer and better prepare them for Asian markets. AutoX currently has 100 robotaxis operating across China.
The robotaxi model allows AutoX to make its AVs smarter and safer every day
In a self-driving car, the driver is basically an artificial intelligence (AI) system that uses a wide range of technologies like light detection and ranging (LIDAR), computer vision, and more. Using machine learning, the driver AI in autonomous vehicles can also learn and become smarter with experience. Therefore, since AutoX is able to deploy its AVs on the streets using the robotaxi model, their driver systems are constantly improving.
According to Xiao, there are three aspects to an autonomous driving system (ADS): the algorithm, the computing platform or hardware, and the data platform. The computing hardware is a supercomputer that can process huge amounts of data at high speed to drive safely on the road. The algorithm is crucial: it optimizes the AI for speedily processing data and thereby driving more safely.
AutoX develops all three elements of ADS in-house. This is not an easy task, since computing hardware for AVs has to meet automobile industry requirements. For instance, computers not only have to be smaller in size than the supercomputers found in data centers, but they also have to be able to withstand high temperatures and vibrations, says Xiao.
Creating a data platform is a similarly Herculean task: it requires a group cloud computing platform that can process enormous amounts of data. Each self-driving car generates one terabyte of data every hour it drives – for context, that’s equivalent to roughly 500 hours of HD video or 6.5 million document pages. This data is what makes AVs smarter over time.
“There’s no limitation to how smart an AI driver can be. The more data, the more practice the AI driver has,” says Xiao.
Last year, AutoX set up Asia’s largest robotaxi operations center in Shanghai to serve as a data hub for its robotaxi fleet in China. After each round of data collection, robotaxis come back to the operations center to upload their data on the cloud. This data is then analyzed and used to improve the robotaxis’ AI drivers. The new version of the AI is then deployed the next time the vehicles return.
“So next time when the car goes out into the street, they can use the latest AI to drive better than before,” says Xiao.
The robotaxi model can bring down the cost of AVs
AutoX aims to, “Not just build self-driving cars, [but] to make many, many of them universally accessible to people for transportation, for mobility, and for logistics,” says Xiao.
One of the biggest barriers to the adoption of AVs is the higher cost. Even though the prices are constantly falling, they are still considered high for middle and low-income households.
According to Xiao, the best way to bring down the cost of AVs is mass production. This would enable companies like AutoX to reap the benefits of economies of scale. And, with the robotaxi model, mass production would become possible much faster than if the startup was focusing solely on private ownership sales.
Besides, the robotaxi model ensures maximum utilization of AVs, says Xiao.
“When you buy a car and go to work, or go to school, the car is sitting for the whole day until you go back home,” says Xiao.
“The taxi scenario is much more suitable because you can significantly increase the utilization rate of the vehicle so that basically everyone is paying for the power of the vehicle instead of [one person] paying for the whole vehicle,” he adds.
In purely business terms, the rollout of robotaxis is likely to generate significant cost savings for transport and mobility companies. In 2019, China registered 1.4 million taxi drivers with an annual average salary of approximately $17,600 (114,000 CNY). For transportation and mobility companies, this cost can be reduced with the deployment of driverless AVs, allowing them to allocate funds towards research and development. Much like similar shifts to autonomous technologies in other industries, taxi drivers will need to be re-skilled in order to make this transition equitable.
However, eliminating taxi drivers altogether may not be an option unless AV companies adopt accessible designs. That’s because for older generations and people with disabilities, taxi drivers are the bridge between personal capability and vehicle accessibility. They are more than drivers –they are a helping hand for the infirm and disabled, or even for able-bodied individuals in case of emergencies.
Moreover, for AV companies like AutoX, the ability to deploy self-driving vehicles without safety drivers will pave the way towards profitability. As the regulatory environment for AVs is becoming more relaxed and governments are granting permits to deploy AVs without safety drivers, AutoX is looking to start recovering the unit cost of its vehicles.
In fact, as soon as safety drivers are no longer required by law to be present inside AVs, AutoX will start generating positive revenue, Xiao predicts. In December last year, AutoX became the first company to deploy a fully driverless fleet of RoboTaxis in Shenzhen, without accompanying safety drivers or remote operators. The startup also obtained a permit to operate AVs without safety drivers in California last year.
Apart from AutoX’s robotaxi rollout, the company is using robotrucks to deploy AVs in logistics. According to Xiao, robotrucks can not only reduce accidents, but can also reduce traffic congestion by driving at night when traffic is minimal. Moreover, because night driving can also reduce the time taken for deliveries from an hour to less than 20 minutes. This in turn results in fuel and battery power conservation, making robotrucks more energy-efficient.
The road ahead
An increase in AV adoption is not only dependent on the supply of safe AVs, but also regulatory changes, government investments, and most importantly, public sentiment.
“We need to educate the general public about the benefits of self-driving cars, and let people embrace this technology,” says Xiao. This is no easy feat, since three out of four Americans are scared of fully automated self-driving cars, according to a 2019 study.
“We also need to talk to the government and regulators to get legal approval and support to deploy this technology on the road,” says Xiao. Besides, well-constructed roads and highways are crucial for successful deployment of AVs, which requires large-scale government investment.
In China, the government is already looking at easing regulations to support the growth of AVs. Included in its new infrastructure plan is an expressway linking Beijing to the Xiongan New Area, with two inside lanes especially designed for AVs. Moreover, the Chinese government’s development of the 5G network will further help deploy AVs.
The trade-war between China and the U.S. has already been impacting Chinese AV startups. In China, AV startups depend on the U.S. for GPU and CPU chips. Intel CPUs and Nvidia GPUs are the most powerful semiconductors that currently exist in the market. The cost of procurement of these chips has increased following the sanctions imposed due to the ongoing trade-war. Moreover, as the relationship between the two countries hangs by a thread, with no end in sight, access to these chips may become near impossible in the future.
As far as commercialization of AVs is concerned, Xiao believes that it would take two to three years for the first stage of large-scale commercialization.
“If we’re talking about replacing the whole country’s taxis or maybe 90% of all taxis in China with self-driving cars, it’s going to take some time, probably up to 10 to 15 years,” says Xiao. “But if we’re talking about the first city, it could be a matter of two to three years.”
In the near future the startup is trying to expand its presence to Wuhan and other cities in China. It is also looking to expand to other Southeast Asian countries, says Xiao.
AVs were a distant dream even near the end of the last century. Fast forward 20 years, and AV startups are poised to rule the car market. According to data from Tracxn, globally, the AV sector has received total funding of around US$23.1 billion, with more than half of it raised in 2018 and 2019.
With the global demand for AVs expected to reach 4.2 million units by 2030, expanding at a compound annual growth rate of 63.1%, it is likely that the competition in the AV landscape will increase. Startups like AutoX will continue to compete to get a bigger slice of the pie.
Images courtesy of AutoX. Monika Ghosh contributing reporting and editing to this story.
