Tesla Rival XPeng and Alibaba Cloud Set Up China's Largest Cloud-Based Computing Center to Train Machine Learning Models for Autonomous Driving
【Summary】Electric vehicle manufacturer XPeng Inc, which is becoming one of Tesla’s biggest rivals in China with its high tech EVs, has set up a data center with Alibaba Cloud to train its autonomous driving systems. The cloud-based data center will be used to train XPeng’s machine learning models to improve the autonomous driving capabilities of its vehicles much more quickly.
Electric vehicle manufacturer XPeng Inc, which is becoming one of Tesla's biggest rivals in China with its high tech EVs, has set up a data center with Alibaba Cloud to train its autonomous driving systems, Chinese new outlet Gasgoo reports.
The cloud-based data center will be used to train XPeng's machine learning models to improve the autonomous driving capabilities of its vehicles much more quickly. For self-driving cars, machine learning algorithms are an important tool that can be used to predict the behavior of other road users, such as other human drivers, pedestrians and bicyclists.
XPeng and Alibaba Cloud's center offers 600 FLOPS (floating point operations per second) of compute power that can reduce training time of XPeng's autonomous driving algorithms from a week to just one hour, according to the company. With its powerful compute capabilities, the data center will also help to promote the development of advanced Level-3 autonomous driving capabilities in China.
The data center is located in Ulanqab City, which is located around 220 miles west of Beijing. Computer giant Apple Inc. also operates a data center in the city to support its iCloud services for customers in China.
Xpeng's autonomous driving system is called XPILOT and it's offered on the automaker's P7 sedan, which is a strong competitor to the Tesla Model S in China. The latest version of XPILOT brings several autonomous driving features to Chinese customers for the first time, including Level 3 autonomous driving functions for highways, urban roads and automated parking.
The P7 was China's first mass produced electric sedan to come with a 360-degree dual camera and a radar fusion perception system for added safety. The electric sedan is outfitted with 14 cameras, 5 millimeter-wave radars, 12 ultrasonic sensors, lidar, centimeter-level high-definition GPS positioning, decimeter-level high-definition mapping.
In Jan 2021, XPeng unveiled the beta version of its more advanced autonomous driving feature called Navigation Guided Pilot (NGP), which is similar to Tesla's Full Self-Driving (FSD) feature.
Xpeng's NGP is highly advanced and specifically designed for China's complex and crowded road conditions and driving style. The system will follow the route guidance while executing and selecting the optimal route in real-time as well as perform autonomous lane changes.
Autonomous driving systems like Xpeng's NGP and XPILOT rely heavily on machine learning algorithms to make better driving decisions over time.
Tesla, for example, continuously collects this type of data from its vehicles remotely, such as when drivers activate its Autopilot automated driving feature. The data collected by Tesla is crunched by machine learning algorithms which are used to make improvements to its autonomous driving software.
However, training these algorithms requires a massive amount of real world data, which is known as "training data". It also requires a tremendous amount of computing power to do these tasks more quickly and efficiently.
Other automakers are also planning using data centers to process data from their vehicles, including Sweden's Volvo Cars. A big part of Volvo's future plans include having ten of thousands of software-powered connected cars on the road in the next decade traveling millions of kilometers and continuously sharing data with the automaker.
All of these software-based vehicles will be like rolling smartphones that can be updated OTA.
To process the massive amount of real-time traffic data collected from Volvo vehicles, Volvo Cars and the automaker's software unit Zenseact are investing in a massive AI-powered data center. The data center can store over 200 Pebibytes (225 million gigabytes) of data.
The crowd-sourced data collected from Volvo vehicles will include continuous inputs from vehicle sensors that monitor the environment, including high-resolution lidar data used for autonomous driving.
By allowing customers to share their vehicle data anonymously, Volvo will be able to continuously make software improvements to its future software-based vehicles, including advanced safety systems and autonomous driving systems. Volvo uses AI and a NVIDIA DRIVE Orin processor installed in the vehicle.
XPeng also uses a Nvidia processor in its P7 sedan. The electric car features the NVIDIA DRIVE AGX Xavier platform to support XPILOT. Nvidia's Xavier SoC is a powerhouse, delivering 30 TOPS (trillions of operations per second) of performance while consuming just 30 watts of power.
Over the past several years, Xpeng has invested heavily in autonomous driving. Roughly half of the company's employees work in R&D of new technology for the electric automaker. Many of these employees work solely on the automaker's autonomous driving technology.
Originally hailing from New Jersey, Eric is a automotive & technology reporter covering the high-tech industry here in Silicon Valley. He has over 15 years of automotive experience and a bachelors degree in computer science. These skills, combined with technical writing and news reporting, allows him to fully understand and identify new and innovative technologies in the auto industry and beyond. He has worked at Uber on self-driving cars and as a technical writer, helping people to understand and work with technology.
Baidu CEO Believes That SAE Level-4 Autonomous Driving Systems Will the First to Enter Commercial Use After L2, Skipping Over L3
Volvo’s Electric Vehicle Brand Polestar Reports $1 Billion in Revenue in the First Half of 2022, Adds 6 New Global Markets
Toyota is Investing an Additional $2.5 Billion to Expand its North Carolina Factory to Boost EV Battery Production
Nexar Releases its ‘Driver Behavioral Map Data’ That Can Help Autonomous Vehicles Operate More Like Human Drivers
Honda and LG Energy Solutions to Build a $4.4 Billion Joint Venture EV Battery Plant in the U.S.
Mercedes-Benz Begins Production of the Highly Anticipated EQS Electric SUV in Alabama
Toyota is Working With the U.S. Dept of Energy to Advance ‘Megawatt-Scale’ Fuel Cell Powered Stationary Energy Generators
Mercedes-Benz Signs MoU With the Government of Canada to Source the Raw Materials for Electric Vehicle Batteries
- General Motors Launches ‘EV Live’ an Interactive Virtual Experience Where Participants Can Learn More About Electric Vehicles
- $12,500 Federal EV Tax Credit Proposal Reportedly Dropped
- EV Startup Fisker Announces ‘Fisker Finance’ a Digital Financing Platform for Purchasing the Electric Ocean SUV
- GM’s Plan to Sell More EVs Than Tesla Hinges on Offering More Affordable Electric Models
- Autonomous Truck Developer Embark Trucks Adds Freight Carrier U.S. Xpress to its Partner Development Program
- Audi Hits EV Startup NIO With a Trademark Lawsuit Over its Vehicle Naming
- Volvo’s EV Brand Polestar Reports Record Q1 Sales, its New SUV Will Be Officially Unveiled in October
- Faraday Future Delays Launch of FF91 Electric SUV Again
- General Motors is Doubling its Super Cruise Hands-Free Driving Network to 400,000 Miles of Roads in North America
- China’s Tesla Rival XPeng Unveils its New ‘S4’ 480kW Electric Vehicle Supercharger That Can Add 125 Miles of Range in Just 5 Minutes