Research On The Innovation Ecosystem Of Intelligent And Connected Automobile Industry

Self-driving technology as a set of multidisciplinary,interdisciplinary representatives of emerging technologies,has attracted researchers in the field of industry,enterprise and government department’s attention,but rarely in the present study deeply study the evolution law of technology integration and predict the future direction of integration,leading to the development direction of technology and industry is not clear,enterprise innovation ability can’t promote quickly.The prediction of technology fusion is a hot spot and difficult point in the current research.At present,it is mainly based on the traditional link prediction methods,which has the disadvantages of low accuracy and lack of foresight.Based on the existing research pain points,this paper studies the convergence and evolution of driverless vehicle technology through the graph convolution DGCNN link prediction method,and analyzes the evolution law of the co-occurrence network of TF(technical field)and IPC4(patent code)of driverless vehicle technology.The link prediction model based on graph convolutional neural network is established to predict the future potential fusion technology pairs,and the development prospect of the potential fusion technology pairs is analyzed in combination with the research status of driverless vehicles.The research contents and innovative achievements of this paper are as follows:(1)The measurement index system of unmanned driving technology fusion was constructed,and the co-occurrence network of TF and IPC4 was established.The topological attributes and evolution process of the network were analyzed through eight network indicators.The results show that the TF technology field has changed from the early large-scale and extensive fusion to the centralized fusion of the core technology field.IPC4 technology continues to converge to the core technology,and the fusion scope of the core technology also continues to expand,and presents a multi-leading fusion mode centered on multiple technologies.(2)The link prediction model based on graph convolution DGCNN is established,and the validity of the model is tested by combining with the previously established IPC4 patent co-occurrence network.At the same time,the AUC value of traditional link prediction is compared with that of calculation.The results show that the link prediction model based on graph convolution DGCNN has better performance.(3)Calculate the incidence of the prediction results of the three-stage IPC co-occurrence network in the subsequent stages,further verify the prediction performance and accuracy of the model,and analyze the prediction results of each stage with Louvain algorithm.The results show that in the future,there is a strong trend of integration in the field of vehicle driving control system,traffic control system,communication network and digital information transmission,as well as the field of visual image data recognition and processing.(4)Put forward feasible suggestions to research institutions,enterprises and governments according to the research status of the predicted potential fusion technology pair.In this paper,patent co-occurrence technology and graph convolutional neural network link prediction are used to make a quantitative analysis of the evolution characteristics of the convergence of unmanned driving technology and mining the potential convergence technology pairs,so as to provide a reference for the formulation of relevant policies and reasonable strategic layout.