Abstract: In the context of software-defined vehicle, over-the-air (OTA) update has become an extremely important way for the intelligent connected vehicle technology improvement, performance services improvement and software sales ecosystem build. Based on mass vehicles over-the-air (OTA) update will generate tens of millions of vehicle data. With deep data mining of vehicle data generated by over-the-air update and big data governance and analysis, it will contribute to guide and optimize the design scheme and update strategy of vehicle over-the-air update. Meanwhile, the digital management level of enterprises can be improved. This paper constructed the data system of vehicle over-the-air update, realized the granular analysis and management of key technical parameters data and carries out the process risk warning. The development structure based on the big data development model of data mining technology and Hadoop open source platform.From the system application, the developed system can effectively improve the level of analysis, storage and intelligent deep analysis of massive data generated by OTA update. Meanwhile, it can contribute to monitor and trace the vehicle change during the whole over-the-air process. Furthermore, it will promote the closed-loop management of OTA update.

Key words: Vehicle over-the-air update; Big data; Deep mining; Data management system; Data mining

摘要： 在软件定义汽车的背景下，汽车远程无线升级（OTA）技术已经成为智能网联汽车技 术改进、性能服务提升、软件销售生态构建的重要方式。基于规模化的汽车远程升级将产生 千万级的海量车辆数据。通过对远程无线升级产生的车辆数据进行数据深度挖掘，同时通过 大数据治理与分析，可以指导并优化车辆远程升级的设计方案与升级策略，提升企业数字化 管理水平。文章基于数据挖掘技术的大数据开发模型，以 Hadoop 开源平台为支撑，构建汽车 远程升级数据系统，实现了对车辆远程升级关键技术参数数据的颗粒化分析与管理，并进行 过程风险预警。通过实例分析表明，该系统可以有效地提高车辆远程升级产生的海量数据的 分析、存储与智慧化深度分析水平，实现全流程的数据监控与追溯，完成汽车远程升级的闭 环管理。

关键词: 汽车远程升级；海量数据；深度挖掘；数据管理系统；数据挖掘