On December 29, Baidu Create 2021 (Baidu AI Developer Conference) autonomous driving sub-forum "Co-creation of automotive robots connection technology and scenarios" was held in the "Xilang" APP, which is the first conference held in the metaverse in China, which can accommodate 100,000 people to interact with the same screen at the same time. At the forum, the 7.0 version of Baidu Apollo's self-driving open platform was officially released, and it showed the layout of Apollo X's plan in the diversified automotive robot ecology. Baidu Apollo is using technology to empower ecological partners through a variety of automotive robot connection technologies and scenarios, and accelerate the development and commercialization of China's autonomous driving industry.
As an important support for the landing of Baidu's multi-faceted automotive robots, the autonomous driving open platform has achieved a milestone complete evolution from code to tools, from open source platform to tool platform. On the basis of four open source platforms of cloud services, open source software, hardware development, and vehicle certification, Apollo 7.0 provides a series of upgrades including the one-stop practice cloud platform Apollo Studio, the industry's leading simulation service, and the efficient new model, which not only opens the full code capability, but also provides a full-stack tool chain for autonomous driving, which is easier to use, more advanced, and more efficient to help developers use the platform capabilities.
Baidu Apollo 7.0 architecture diagram
At the level of cloud service platform, Apollo 7.0 officially upgraded the "data pipeline" service that was popular among developers in version 6.0 to Apollo Studio, covering the whole process of cloud toolchain for developers from machine to car practice, providing developers with a one-stop practice platform experience. At the simulation platform level, Apollo 7.0 launched the industry's first PnC reinforcement learning model training and simulation evaluation platform, which has multiple advantages such as data reality, powerful functions, comprehensive evaluation standards, and scalable architecture, which is expected to provide a unified verification standard for reinforcement learning research. At the level of open source software platform, Apollo 7.0 upgrades the perception and prediction algorithm modules, and introduces three deep learning-based models of MaskPillars, SMOKE and Inter-TNT, effectively reducing missed detection, jitter and other issues.
At present, Baidu Apollo Autonomous Driving Open Platform has completed 11 iterations, with more than 80,000 developers in 135 countries, more than 210 partners, and an overall open source code volume of 700,000 lines, which has grown into the world's most active autonomous driving open platform.
Apollo X multi-vehicle robot
While empowering developers, Apollo 7.0 further expands the ability of automotive robots to connect technologies and scenarios. Through collaborative industry and empowerment industry, Baidu Apollo and a number of ecological partners jointly created "automotive robots" throughout the fields of passenger transport, freight, life services and other fields, all-round landing passenger cars, buses, trunk logistics, warehousing and distribution, mine port operations and other subdivision scenarios. In the "scene" of the sub-forum of automatic driving, the "Apollon" bus car robot, the DeepWay trunk logistics car robot, the Huaneng Yimin autonomous mining card car robot, and the Neolithic retail car robot in baidu Apollo ecology made a collective appearance.
As China's first intelligent solution to achieve all-round coverage of the vehicle road cloud in the mining field, Baidu Apollo Mine Special Operation Automobile Robot has fully automatic driving capabilities, and can achieve accurate location and operation status information through the vehicle-road collaboration program, and will continue to evolve in the future to support cross-surface full-process operations; Deepway, a trunk logistics automotive robot, is driven by data and self-learning, so that the automatic driving ability of high-speed heavy-duty truck scenarios can be rapidly iterated. Deepway expects to invest 1,000 heavy-duty trucks in 2023, making Deepway an old driver of AI in trunk logistics.
Apollo EDU Autonomous Driving Education Talent Program
In addition, Baidu Apollo also shared the rich resources and latest progress of the Apollo EDU Autonomous Driving Education Talent Program with developers, and announced that it plans to empower tens of millions of colleges and universities and enterprises in the next three years, cultivate millions of autonomous driving engineering talents, fully cover applied and research talents, so that autonomous driving can truly empower all walks of life and promote the sustainable prosperity and development of the autonomous driving industry.
Adhering to the core concept of "open ability, shared resources, accelerated innovation, and sustained win-win", Baidu Apollo continues to evolve with an open attitude. "Co-creating automotive robots, connecting technologies and scenarios" has become an important consensus of Baidu Apollo Autonomous Driving Ecological Alliance. In the future, Baidu Apollo will continue to accelerate the landing of subdivided automotive robot scenarios, work with partners to the road of autonomous driving commercialization, and continuously promote the development process of autonomous driving in China and even the world.