摘要
本文回顾分析了驾驶意图识别在汽车换挡(制动)操作策略、汽车安全预警及智能驾驶技术等3个方面的研究成果。研究表明:车辆自身信息(制动、换挡、方向盘转角、车速及加速度)、交通环境信息以及驾驶人头部及注视信息被广泛用于驾驶意图识别,识别效果较好,同时对比分析了4种驾驶意图识别方法的优缺点,为驾驶意图被应用于智能汽车交通研究领域提供支持。
This paper reviews and analyses the research results of driving intention recognition in three aspects: shift (braking) operation strategy, vehicle safety early warning and intelligent driving technology. The study shows that vehicle information (braking, gear shifting, steering wheel angle, speed and acceleration), traffic environment information and driver's head and gaze information are widely used in driving intention recognition, and the recognition effect is good. Then, four main machine learning algorithms for driving intention recognition and their advantages and disadvantages are analyzed comprehensively. Driving intention is regarded as a key and difficult problem in the field of intelligent vehicle traffic research.
作者
潘鑫
成英
郭渊
PAN Xin;CHENG Ying;GUO Yuan(School of Automotive & Transportation,Tianjin University of Technology & Education,Tianjin 300222,China;Tianjin YunKong Technology Co.,Ltd.,Tianjin 300040,China)
出处
《内燃机与配件》
2019年第7期144-146,共3页
Internal Combustion Engine & Parts
基金
天津市科技创新平台专项--智能交通协同控制技术服务平台(16PTGCCX00150)
天津职业技术师范大学校级科研项目(KJ17-08)
天津市大学生创新训练计划项目(201710066107)
关键词
驾驶意图
汽车安全预警
智能驾驶技术
机器学习
driving intention
vehicle safety warning
intelligent driving technology
machine learning
