Li Deyi received his Doctoral Degree from Heriot-Watt University in Edinburgh, UK, in 1983. Currently, he is an academician at both the Chinese Academy of Engineering and International Academy of Sciences for Europe and Asia, Dean of College of Robotics of Beijing Union University, President of Chinese Association for Artificial Intelligence, and Director of China Cloud Computing Expert Committee as well as being adjunct professor and doctoral supervisor at Tsinghua University and the National Defense University. As one of the first scholars studying in the UK after the reform and opening-up, he was honored as a returning overseas scholar and young/middle-aged expert making outstanding contributions. Supported by special government allowance, he also obtained the Ho Leung Ho Lee Foundation and Significant Contribution Award for Military Professional Technology. He has published three English-version monographs, two Chinese-version monographs, and seven books of a technology series as an editor-in-chief. He also obtained over ten national patents for inventions. He has presented over one hundred academic papers, with one paper cited over a thousand times among and accumulative ten thousand citations in total. Meanwhile, he has cultivated a hundred or more students with doctoral and master’s degrees.
Academician Li Deyi has been engaging in research on computer engineering, artificial intelligence with uncertainty, big data and intelligent driving for a long time. He proposed a new theory on graphic pairs of “control-flow and data-flow” that proves parity of relational database schema and predicate logic, which won the Premium Award by the IEE. He proposed theories of cognitive formalization, such as cloud model, cloud transform and data fields used for solving problems like qualitative concept generation, similarity calculation, uncertainty reasoning and intelligent control. He also succeeded in controlling various postures of dynamic balance of a triple inverted pendulum, which garnered the Outstanding Paper Prize from the International Federation of Automatic Control. His monograph titled Artificial Intelligence with Uncertainty was selected for inclusion in 100 original books of Chinese domestic natural sciences. He once proposed establishing a cognitive map of driving state based on right of way and carried out R & D work for driving brains. He also led China’s largest joint team of intelligent vehicles in accomplishing a whole-course unmanned drive on the expressway from Beijing to Tianjin for the first time. He also took the lead in enabling the world’s first intelligent bus to self-drive on open roads from Zhengzhou to Kaifeng. Moreover, he has won first prize in China's Intelligent Vehicle Future Challenge many times.
Abstract:
In recent years, auto-assisted driving has been flourishing, from driver-assistance level (L1) to partial autopilot (L2), to self-driving by the robotic vehicle (L3), from previous warning and reminding the driver to wholely depending on the intelligent vehicle. The driver's role no longer exists if we can achieve L3 perfectly. However, it is not easy to reach to L3. On the one hand, lots of social problems need us to solve, for example, the driving permission problem, the liability problem in the traffic accident and so on. On the other hand, the technologies are not mature enough by now. A most difficult problem is to understand the environment and make a correct decision rapidly like humans. Considering that, we develop a driving-brain toolkit, which could study how human drivers behave when driving and extract the skills to apply to the intelligent driving. At the end of the report, We also want to discuss the future large-scale productions of the L3 intelligent vehicles.
