51WORLD Inks Cooperation Agreement with Suzhou Automotive Research Institute of Tsinghua University to Step up Building Autonomous Driving Scene Library with Chinese Characteristics
On March 18th, 51WORLD inked a cooperation agreement for the development and promotion of autonomous driving scene library, with Suzhou Automotive Research Institute, Tsinghua University and Jiangsu Intelligent and Connected Vehicles Innovation Center, a provincial ICV platform affiliated to the former, aiming to step up the construction of the self-driving scene library with Chinese characteristics.
A virtual scene database is a bridge connecting data and application. The scene library of autonomous driving tests is a limited reflection of the infinite driving environment, a test library for evaluating autonomous driving systems, a basis for judging the intelligent levels of autonomous cars, and a standard of verifying the safety of autonomous vehicles. The basic data service with scene library as the core and the complete toolchain for building the scene library is what the ICV industry is vigorously going after.
According to the agreement, 51WORLD provided autonomous driving simulation software and data processing toolchain, while Tsinghua Automotive Research Institute and the Innovation Center offered their rich data on autonomous driving scenes. This strategic cooperation would speed up the construction of the autonomous driving scene library with Chinese characteristics as China’s unique road traffic conditions and complex traffic behaviors were taken into consideration in building the scene library.
In 2018, 51WORLD debuted 51Sim-One, China’s first all-module autonomous driving simulation test platform. Incorporating a variety of simulations and testing technologies like road environment simulation, dynamic traffic flow and scene simulation, multi-sensor simulation, vehicle dynamics simulation, automated testing and cloud computing, the platform overcame multiple hurdles in establishing a super-large virtual scene library, and achieved large-scale simulation and testing in continuous traffic flow, blazing a trail of physical-level sensor simulation models in China.
Meanwhile, with its self-developed static scene data editing and automatic conversion tools, 51Sim-One was able to integrate various data of the existing scenes to convert a physical scene quickly into a high-fidelity virtual scene. In addition, it was adept at building a virtual scene from scratch according to varied needs of testing, to rapidly apply scene data to specific software, which provided a scene-based simulation test environment for developers and improved the efficiency of developing the autonomous driving algorithm.
It is foreseeable that this cooperation will further elevate the developing and testing of the ICVs in China and accelerate their applications supported by autonomous driving scene library.