Shulou(Shulou.com)12/24 Report--

CTOnews.com December 19 news, Baidu today officially launched a new upgraded version of the Apollo open platform-Apollo open platform 9.0.

Zhang Liang, General Manager of Ecology Department of Baidu self-driving platform, summed up the main highlights of this upgrade: "Apollo Open platform 9.0 has achieved a comprehensive upgrade in terms of engineering, algorithms and tools, and the general layer can enable a variety of application scenarios to land on a large scale, the overall operation is more flexible and easy to use, and the use scene is universal and easy to expand. Greatly improve the development efficiency at the same time, can help more developers quickly build their own autopilot system. "

In terms of engineering framework, in order to enable developers to assemble autopilot applications more flexibly and facilitate secondary development, Apollo Open platform 9.0 comprehensively upgrades package management, splitting modules into smaller software packages according to functional granularity, developers can more easily choose and use according to their own needs, while also providing rich functional components and plug-ins, and the functional extension is enhanced and optimized.

Based on this, after the unified scheduling interface, developers can complete the construction of scenario Demo within 1 day at the earliest. The simplified parameter adjustment method doubles the parameter adjustment efficiency, and the new plug-in mechanism reduces the code learning cost by 50%, which greatly improves the secondary development capability of Apollo. The Apollo open platform 9.0 also adapts to the ARM architecture for the first time.

In terms of algorithm, Apollo open platform 9.0 optimizes the perception algorithm on the basis of the previous, Lidar detection uses a relatively new CenterPoint model, visual use of Yolo X + Yolo 3D model, the two models into millions of data training, recall and accuracy are claimed to be "greatly improved".

Apollo open platform 9.0 also provides incremental training to support the independent completion of model training, which can improve the detection ability of special targets and special scenes while maintaining the original detection ability of the model, so as to easily improve the detection effect of customized scenes with lower cost. In addition, the new platform fully supports 4D millimeter wave radar, which enhances the security of obstacle detection and extreme weather scenes.

In terms of tools, in addition to the new functions on the basis of the original tools, Apollo Open platform 9.0 has also added tools such as high-precision mapping, sensor calibration and integration. The newly upgraded Dreamview + of Apollo open platform 9.0 has been comprehensively improved in many aspects, such as multi-scene use, free layout, data resources and so on. The debugging process is simpler, the window layout is more flexible, and the access to resources is more convenient.

In addition, the Apollo open platform 9.0 also reconstructs the documentation platform, which makes it easier to operate, read more smoothly, enrich the content, and effectively reduce the learning and use costs of developers.

At the press conference, Dr. Dong Zhaozhi, Executive Vice President of Kai Wo Group and President of Academia Sinica, Qiang Xiaowen, Party Committee Secretary, Deputy General Minister and President of Commodity Research Institute of Dongfeng Motor Co., Ltd., Dr. Liu Mingchun, Vice President of forward-looking Technology Research Institute of Jinlong United Automotive Industry (Suzhou) Co., Ltd., and Dr. Liu Junchuan, Vice President of Tianzhun Technology Co., Ltd. Also attended the scene.

Baidu said that the Apollo open platform 9.0 has a strong scenario versatility and is easier to use and expand. It reduces adaptation by 40 per cent, code reading by 90 per cent and code debugging by 80 per cent. It can be used "out of the box" and self-driving vehicles can be closed-loop within a week. At the same time, the scenarios covered by Apollo open platform 9.0 are more abundant, and the closed-loop jobs of scenario application business systems can be completed in one month. On this basis, the sensor calibration and map creation cycle are shortened to the hour level, which greatly shortens the landing time and is more efficient.

At the same time, the hardware cost of Apollo open platform 9.0 is lower, the selection is more abundant, the hardware selection supports 3Lidar+4Camera, the camera supports more than 4 manufacturers, upgrading from USB3.0 to GMSL, lidar adds 32-line, 64-line and other multi-brand and multi-model equipment, positioning equipment is supported by more than 3 mainstream brand devices, which can help the scale applications of different scenarios to land in all aspects.

CTOnews.com learned from the press conference that at present, the Apollo open platform brings together more than 160000 developers from more than 170 countries and regions around the world, refactoring 120,000 lines of code and adding 200,000 lines of code in the development process of the Apollo open platform.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.