COO of DiDi Autonomous Driving: DiDi is Building Full-Stack L4 Framework
DiDi considers Data, Networks and AI the components of the 'DNA' of autonomous driving technology.
Xing Meng, DiDi Autonomous Driving's COO, revealed DiDi's development in autonomous driving technology during the 2020 China Auto Forum on August 15th.
According to Xing, DiDi has a clear vision of how the DNA of autonomous driving will help to generate its future form. Based on the three keys of Data, Networks and AI, DiDi is building full-stack autonomous driving systems that will develop to full maturity to match rising infrastructural support – these include in-vehicle online systems, sensing and predicting systems, high-definition map navigation systems, etc.
With 8 years of accumulated experience in the vehicle-for-hire industry, DiDi has a massive database on the taxi business, traffic conditions and traffic safety. DiDi is now focusing on AI technologies to mine the data and build L4 autonomous driving.
Xing emphasized the importance of data processing during his speech. DiDi pays extra attention to the extreme cases that are rare in scenario testing data but must be taken into account by autonomous driving systems. DiDi collects data of vehicles from its tens of millions of drivers and applies simulation models to expand its data to billion-level, in order to cover the most extreme cases.