Baidu, Inc, leading AI company, announced plans to build the world’s largest autonomous ride-hailing service area in 2023. The announcement came during a celebration of Apollo Day, Baidu’s autonomous driving tech event (Click to replay 2022 Baidu Apollo Day). The plan outlined a goal to expand the operation area for Baidu’s fully driverless robotaxis, allowing Baidu to reach more potential customers. Additionally, Baidu revealed a series of new technology breakthroughs including an AI big model built for autonomous driving perception, high-definition autonomous driving maps, a closed-loop autonomous driving data system, and the successful end-to-end adaptation of AI chips for autonomous vehicles.
Since August 2022, Baidu has already rolled out fully driverless ride-hailing services (with no human drivers in the car) in the cities of Chongqing and Wuhan, with access to hundreds of square kilometers of operation area. Based on this leading position, Baidu will continue to expand its operation area next year to build the world’s largest service area for fully driverless robotaxi service.
Currently, Baidu’s autonomous ride-hailing platform Apollo Go covers more than 10 cities in China including all first-tier cities. In Q3 2022 alone, Apollo Go has completed more than 474,000 rides, up 311% year over year, and a 65% increase compared to last quarter. In first-tier cities like Beijing and Shanghai, each robotaxi on Apollo Go can provide 15 rides a day on average, nearly the same daily ride average of typical online ride-hailing car services. By the end of Q3 2022, the accumulated rides provided to the public by Apollo Go have reached 1.4 million. As Baidu continues to scale up its operation area of robotaxi service, it is one step closer to the goal of providing autonomous driving services to more people, while further strengthening its leading position in the global autonomous ride-hailing market.
“Backed by its solid AI technology, Baidu Apollo has created a safe, intelligent and efficient autonomous driving technology system, bringing robotaxi services from designated zones to open roads at scale,” said Jingkai Chen, Baidu’s autonomous driving technology expert at the event. The generalization ability of Baidu’s autonomous driving technology has progressed at a more advanced pace than expected. Now, the lead time to deploy autonomous driving technology in a new city is only 20 days.
Industry’s first AI big model for autonomous driving, addressing the “long tail” problem
The AV industry has long grappled with the “long tail” problem, in which an autonomous vehicle runs into a scenario it has not seen or experienced before. To address this problem, Baidu’s autonomous driving technology expert Jingdong Wang has announced the industry’s first AI big model for autonomous driving, a pre-trained visual-language model with weak supervision, backed by the Baidu WenXin Big Model, which recognizes thousands of objects, helping to enlarge the scope of semantic recognition. The model will enable autonomous vehicles to quickly make sense of an unseen object, such as special vehicle (fire truck, ambulance) recognition, plastic bag misdetection, and others. In addition, Baidu’s autonomous driving perception model—a sub-model of the WenXin Big Model—leveraging more than 1 billion parameters, is able to dramatically improve the generalization potential of autonomous driving perception.
High-definition autonomous driving map to safeguard a smarter and more efficient autonomous driving experience
“Baidu’s new generation autonomous driving map is equipped with comprehensive capabilities such as automatic production, real-time fusion, and knowledge enhancement.” according to Jizhou Huang, Baidu’s autonomous driving technology expert. It will be put into mass production to realize a “safe, reliable and efficient” autonomous driving experience.
- 96% automation rate for map production: With AI as the key driving force to increase efficiency and bring down the cost, the automation rate of Baidu’s high precision map production has now reached 96%.
- Real-time map update to ensure driving safety: The integration of vehicle-side perception data and multi-source maps to generate online maps in real time has helped to significantly ensure the safety of autonomous driving.
- Massive data plus human driver knowledge to improve driving reliability: With more than 12 million kilometers of road networks and data accumulated on Baidu Maps, along with hundreds of millions of human drive-hours, Baidu’s autonomous driving map effectively integrates these data points to improve the reliability of autonomous driving.
Closed-loop data system to further enhance the intelligence of autonomous driving
With autonomous vehicles being deployed at a larger scale, the scale of data received will also increase exponentially. This brings challenges to identifying valuable data and efficiently using data to continuously improve autonomous driving. Ang Li, Baidu’s technology expert, introduced the concept of “Fine Purification, Strong Ingestion” and Apollo Loop, a closed-loop data system, to effectively identify and utilize data. To purify the data, the system leverages both on-board small AI models and cloud-based big AI model to achieve high-efficiency data mining and automated labeling. The data ingestion architecture achieves automated training with its group-optimization ability and data distribution understanding to effectively utilize data and further enhance the overall intelligence of autonomous driving.
Developing a mutually reinforcing parallel use model for L4 and L2+ autonomous driving technologies
Baidu is also actively bringing autonomous driving technology to empower advanced assisted driving products. Baidu’s autonomous driving technology expert Liang Wang explained how Baidu is leveraging its decade-long experience in autonomous driving to explore a technical route through which L4 and L2+ autonomous driving can coexist. Currently, the technology stack level enables the unification of L4 and L2+ smart driving products in terms of visual perception scheme, technical architecture, map unification, data interconnection and infrastructure sharing. Baidu envisions a mutually beneficial relationship in which L4 will continue to provide advanced technology migration for L2+ smart driving products in urban use cases, while L2 data feedback will also help to improve L4 generalization ability.
Baidu’s Kunlun AI chip completed end-to-end autonomous driving real-world testing
Jian Ouyang, CEO of Kunlun Chip, also revealed at today’s event that Baidu’s 2nd-gen Kunlun AI chip has completed an end-to-end performance adaptation for autonomous driving. This marks a major milestone further solidifying and integrating Baidu’s advantages in both autonomous driving software and hardware.