面向无人驾驶的数据采集与分析系统研究综述  

Review on Data Acquisition and Analysis System for Autonomous Vehicles

作　　者：高振海[1,2] 于桐 孙天骏 王雨蒙[2] Gao Zhenhai;Yu Tong;Sun Tianjun;Wang Yumeng(State Key Laboratory of Automobile Simulation and Control,Jilin University,Changchun 130022;College of Automotive Engineering.Jjlin University,Changchun 130022)

机构地区：[1]吉林大学汽车仿真与控制国家重点实验室,长春130022 [2]吉林大学汽车工程学院,长春130022

出　　处：《汽车技术》2021年第6期1-11,共11页Automobile Technology

摘　　要：车辆在行驶时会产生大量复杂的数据,对这些数据的采集和分析会促进无人驾驶技术的发展。通过研究,将无人驾驶所需数据分为交通场景数据和驾驶行为数据两类,综述了两类数据的采集和分析技术的发展历程及研究现状,并综合考虑企业、高校和相关研发机构的需求,提出了交通场景和驾驶行为数据采集系统需要向着建立本土化大规模的采集系统和建立基于智能手机的分布式采集系统的方向发展的趋势,以及交通场景和驾驶行为数据分析系统需要着力攻关场景理解技术并统一驾驶风格算法评价标准的发展趋势。Huge amounts of complex data will be generated when the vehicle is in motion,acquisition and analysis of these data will promote the development of autonomous vehicles.Through research,the data required by autonomous vehicles is divided into traffic scene data and driving behavior data.In this paper,the development history and research status of technologies for data acquisition and data analysis are summarized.Comprehensively considering the demand of enterprises,universities and relevant research and development institutions,it is pointed out that the traffic scene and driving behavior data acquisition system will develop towards the direction of establishing a system for collecting largescale local data and a distributed acquisition system based on smart phones,and the traffic scene and driving behavior data analysis system will focus on tackling the key problems of scene understanding technology and unifying the driving style algorithm evaluation standards.

关 键 词：无人驾驶 数据采集与分析 交通场景 驾驶行为 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TP392[自动化与计算机技术—计算机科学与技术]