Abstract:    Due to the great potential in technological, market and societal aspect, autonomous vehicle has become one of the best industrial choices for countries to develop digital economy and intelligent society. However, as autonomous vehicle technology integrated a variety of cutting-edge technologies, the technological trajectory of autonomous vehicles has a very high degree of complexity and uncertainty, which poses huge challenges to corporate innovation strategy planning and industrial policy formulation. The existing method of main path analysis identify the technological trajectory mainly based on existing patent citation relationship which is useful for technologies in stable or mature stage, but its accuracy will be greatly reduced when we try to use it for technology that is still in rapid development or turbulent stage, and it is also difficult to make accurate predictions about the future direction of the industrial technological trajectory. In recent years, more and more scholars tried to combine main path analysis with cluster analysis, text mining and other methods, or integrate multi-source data for more accurate identification and prediction of technological trajectory.
   Based on the recent development in the literature, this paper proposed a systematic analysis framework which integrated the development dynamics of industrial technology, the evolution of technology main path and the strategy of leading enterprises. Through the systematic analysis of different information, the process, paths and trends of autonomous vehicle technological trajectory were comprehensively analyzed. To be specific, this paper first searched and sorted out the autonomous vehicle patent data from 1995 to 2018 in Derwent Innovations Index, and got 42332 patents filed in 34 countries/regions including the U.S., Japan, Europe, China and South Korea etc. Then the technology community was explored through community analysis, word cloud analysis, social network analysis and other methods, and the development path of autonomous vehicle technology was obtained through key-route main path analysis method. At last, the path, direction and driving factors of technological trajectory evolution were further explored by combining the leading corporate strategy and industrial development practice with the above results.
    This study finds out that the development of autonomous vehicle technology has three different phases and ten main technological communities emerged gradually. Although the theme, scale, structure and leading firms of different communities are different, these communities have jointly promoted the continued development of autonomous vehicle technology. Second, the development of autonomous vehicle technology has formed a complex main path. This main path originates from different starting points and has undergone a process of fusion—separation—refusion—reseparation. With the advancement of the main path, the focus of technology has gradually evolved from assisted driving to partially automated driving and then to automated driving in complex scenarios. Third, as a whole, there are two routes for the technological evolution of autonomous vehicles: gradual innovation route and radical innovation route. The development of leading technologies has two different directions: self-intelligence mode depending heavily on sensors of the car and connected vehicle mode making use of V2X technology. The future evolution results will be determined by the joint influence of technology, market, and policy. 
    The development of autonomous vehicle technology started relative late in China, and generally speaking it is still in a fast-catch-up situation. Although the total number of patents of China is only less than the United States and Japan, the influence of our technology is still relatively weak, hence the future research and development work has a long way to go. The research results of this article have the following two enlightenments for the development of China′s autonomous vehicle industry: First, strengthen the research, monitoring and prediction of industrial technology development of autonomous vehicle, and accelerate the advancement of industrial technology innovation. Second, make action plans for key technology areas such as the connected car technology, and try to seize an advantageous position in the future development of autonomous vehicle technology.

Key words:  autonomous driving, technological trajectory, evolution, main path, technological community