Menu
China races ahead in intelligent driving technology
By Wang Zheng, People's Daily
Visitors explore automotive chips at the 2024 World Intelligent Connected Vehicles Conference in Beijing, Oct. 17, 2024. (Photo by Chen Xiaogen/People's Daily Online)
On a fog-shrouded mountain road in southwest China's Guizhou province, a sedan equipped with cutting-edge intelligent driving technology detects a flock of sheep crossing the path. Without hesitation, the vehicle automatically swerves to avoid them, simultaneously activating its hazard lights to alert following drivers.
At a recent press conference, Wang Chuanfu, chairman and president of Chinese automaker BYD, showcased this real-world test as evidence that advanced intelligent driving is set to become as indispensable as seat belts and airbags within the next two or three years. His remarks come amid a broader push by China's automakers to integrate smart driving systems into everyday vehicles.
Chinese car manufacturers such as Changan, BYD, and Geely have announced plans to bring intelligent driving into the mainstream.
Changan's "Beidou Tian Shu 2.0" initiative, for instance, aims to unveil 35 new intelligent vehicles over the next three years. Geely, meanwhile, is rolling out its "Qianli Haohan" intelligent driving system with future models under its Galaxy and Star series designed to incorporate the technology from the outset.
BYD has pledged to outfit its entire lineup with the "Tianshen Zhi Yan" advanced intelligent driving system. The automaker's latest model, the BYD Seagull, is priced at a modest 69,800 yuan (about $9,609.3), signaling a dramatic shift in making such technology accessible.
"Safety is the cornerstone of intelligent driving," said Gan Jiayue, CEO of Geely Auto Group. "We are working toward a system where every vehicle benefits from integrated perception and comprehensive risk prediction."
Beyond collision avoidance, AI-powered intelligent driving technology is also redefining vehicle navigation and parking. Huawei's ADS 3.0 system, enhanced by the Pangu Large Model, has raised the success rate of navigating complex intersections from 85 percent to 98 percent. Likewise, NIO's NOP+ system now boasts a 95 percent success rate for automatic parking in home garages, even learning the specifics of individual parking spots.
The leap in performance can largely be attributed to the advent of end-to-end large models. Unlike traditional "perception-decision-execution" architectures that rely on tens of thousands of engineered rules, these new models use vast amounts of real-world data - collected over tens of millions of kilometers - to generate adaptive solutions for over 95 percent of driving scenarios.
Industry experts believe that this technological transformation, combined with a surge in Chinese domestically produced large AI models, is reshaping both the technological and economic landscape of intelligent driving.
Cost reduction has been a central theme in this revolution. In Shenzhen, at the factory of Robosense Technology Co., Ltd., an M Platform LiDAR sensor rolls off the production line every 12 seconds. Once a prohibitively expensive component, costing as much as $80,000 in 2016, LiDAR sensors now retail for around $200 thanks to advances in solid-state design and chip integration.
"More than 30 intelligent electric vehicle models already incorporate this sensor, and over 14 additional models are scheduled to follow in 2025," said Qiu Chunchao, CEO of Robosense Technology Co., Ltd.
Similarly, BYD has slashed the cost of its 4D millimeter-wave radar to one-third of the industry average by leveraging its self-developed "Xuanji" intelligent driving chip. Changan, through its in-house domain controller and collaboration with China's autonomous driving chipmaker Black Sesame Technologies, has reduced the hardware cost for its intelligent driving system by 58 percent compared with 2024 levels.
"China is setting the new cost benchmark for intelligent driving," said Zhang Yongwei, secretary-general of the China EV100, a new energy vehicle industry think tank.
According to Shi Shuai, a partner of the Automotive Team in Roland Berger Greater China, China's competitive advantage lies not in a single technological breakthrough but in a synergistic combination of localized chip production, rapid algorithm iteration, and robust manufacturing capabilities. This ecosystem, he noted, is allowing Chinese firms to scale innovations at an unprecedented pace.
With a new version released every two months, advanced intelligent driving is growing rapidly among Chinese automakers. While several companies plan to launch Level 3 autonomous driving models this year, experts warn that widespread adoption hinges on clarifying the legal responsibilities of drivers and automakers.
According to data from China's Ministry of Industry and Information Technology, 55.7 percent of new passenger vehicles in the country were equipped with Level 2 or higher autonomous driving systems from January to June in 2024. Zhang predicts this figure will rise to 65 percent by 2025.
Driven by advancements in large AI models and collaborative innovation across the industry, intelligent driving in China is steadily bridging the gap to full autonomy, redefining the global automotive landscape along the way.
At a recent press conference, Wang Chuanfu, chairman and president of Chinese automaker BYD, showcased this real-world test as evidence that advanced intelligent driving is set to become as indispensable as seat belts and airbags within the next two or three years. His remarks come amid a broader push by China's automakers to integrate smart driving systems into everyday vehicles.
Chinese car manufacturers such as Changan, BYD, and Geely have announced plans to bring intelligent driving into the mainstream.
Changan's "Beidou Tian Shu 2.0" initiative, for instance, aims to unveil 35 new intelligent vehicles over the next three years. Geely, meanwhile, is rolling out its "Qianli Haohan" intelligent driving system with future models under its Galaxy and Star series designed to incorporate the technology from the outset.
BYD has pledged to outfit its entire lineup with the "Tianshen Zhi Yan" advanced intelligent driving system. The automaker's latest model, the BYD Seagull, is priced at a modest 69,800 yuan (about $9,609.3), signaling a dramatic shift in making such technology accessible.
"Safety is the cornerstone of intelligent driving," said Gan Jiayue, CEO of Geely Auto Group. "We are working toward a system where every vehicle benefits from integrated perception and comprehensive risk prediction."
Beyond collision avoidance, AI-powered intelligent driving technology is also redefining vehicle navigation and parking. Huawei's ADS 3.0 system, enhanced by the Pangu Large Model, has raised the success rate of navigating complex intersections from 85 percent to 98 percent. Likewise, NIO's NOP+ system now boasts a 95 percent success rate for automatic parking in home garages, even learning the specifics of individual parking spots.
The leap in performance can largely be attributed to the advent of end-to-end large models. Unlike traditional "perception-decision-execution" architectures that rely on tens of thousands of engineered rules, these new models use vast amounts of real-world data - collected over tens of millions of kilometers - to generate adaptive solutions for over 95 percent of driving scenarios.
Industry experts believe that this technological transformation, combined with a surge in Chinese domestically produced large AI models, is reshaping both the technological and economic landscape of intelligent driving.
Cost reduction has been a central theme in this revolution. In Shenzhen, at the factory of Robosense Technology Co., Ltd., an M Platform LiDAR sensor rolls off the production line every 12 seconds. Once a prohibitively expensive component, costing as much as $80,000 in 2016, LiDAR sensors now retail for around $200 thanks to advances in solid-state design and chip integration.
"More than 30 intelligent electric vehicle models already incorporate this sensor, and over 14 additional models are scheduled to follow in 2025," said Qiu Chunchao, CEO of Robosense Technology Co., Ltd.
Similarly, BYD has slashed the cost of its 4D millimeter-wave radar to one-third of the industry average by leveraging its self-developed "Xuanji" intelligent driving chip. Changan, through its in-house domain controller and collaboration with China's autonomous driving chipmaker Black Sesame Technologies, has reduced the hardware cost for its intelligent driving system by 58 percent compared with 2024 levels.
"China is setting the new cost benchmark for intelligent driving," said Zhang Yongwei, secretary-general of the China EV100, a new energy vehicle industry think tank.
According to Shi Shuai, a partner of the Automotive Team in Roland Berger Greater China, China's competitive advantage lies not in a single technological breakthrough but in a synergistic combination of localized chip production, rapid algorithm iteration, and robust manufacturing capabilities. This ecosystem, he noted, is allowing Chinese firms to scale innovations at an unprecedented pace.
With a new version released every two months, advanced intelligent driving is growing rapidly among Chinese automakers. While several companies plan to launch Level 3 autonomous driving models this year, experts warn that widespread adoption hinges on clarifying the legal responsibilities of drivers and automakers.
According to data from China's Ministry of Industry and Information Technology, 55.7 percent of new passenger vehicles in the country were equipped with Level 2 or higher autonomous driving systems from January to June in 2024. Zhang predicts this figure will rise to 65 percent by 2025.
Driven by advancements in large AI models and collaborative innovation across the industry, intelligent driving in China is steadily bridging the gap to full autonomy, redefining the global automotive landscape along the way.
