Research on Training Optimization of Artificial Intelligence Autonomous Driving Model Based on Big Data Analysis

With the development of artificial intelligence and world automobile transportation, intelligent transportation technology represented by automatic driving has become an important technical field. Modern autonomous driving technology is usually based on various predictive models established by data analysis, including but not limited to the deep learning models represented by convolutional neural network (CNN) and recurrent neural network (RNN). The operational safety and efficiency of these models are key indicators to measure the excellence of these models, so how to make these autonomous driving models play a more safe and accurate role is very important. This article combines literature review, experimental research and case analysis. This article summarizes the infrastructure and research status of autonomous driving model, the status of deep learning model and the shortcomings still faced. At the same time, this article also provides strategies for optimizing autonomous driving models based on key processing improvements of big data, including data acquisition optimization, preprocessing, and optimizing autonomous driving models from three systems: environment awareness, autonomous decision making, and control supervisor. It also provides realistic solutions for optimizing autonomous driving models in modern industry and research results in academia, and provides some future prospects.