基于惯性传感器的驾驶动作监测识别研究  

Research on driving motion monitoring and recognitionbased on inertial sensor

作　　者：侯宇晗 朱玉杰[1] 王景峰[1] 陈雨诗 HOU Yuhan;ZHU Yujie;WANG Jingfeng;CHEN Yushi(School of Engineering and Technology,Northeast Forestry University,Harbin 150040,China)

机构地区：[1]东北林业大学工程技术学院,黑龙江哈尔滨150040

出　　处：《传感器与微系统》2020年第3期53-56,共4页Transducer and Microsystem Technologies

摘　　要：基于惯性传感器构建人体动作闭环研究框架;利用模拟驾驶试验台和Captiv可穿戴动作捕捉系统测量不同驾驶动作的各类参数;利用分类器并使用K最近邻算法对测量数据进行分析、分类,可达到监测识别驾驶动作的目的,从而识别出驾驶员的不良驾驶动作,并可在后续研究中予以提示。此研究为基于惯性传感器的驾驶动作监测与反馈系统提供了一种技术方法,为对该技术与其他技术的交叉融合提供思路。Closed-loop research framework of human motion is constructed based on inertial sensors.The various parameters of different driving actions are measured by the simulated driving test bench and the wearable motion capture system of Captiv.Using classifiers and using K-nearest neighbor(KNN)algorithm to analyze and classify the measured data,to achieve the purpose of monitoring and recognizing the driving action,thereby identifying the driver’s bad driving action and suggesting it in subsequent research.This research provides a technical method for the driving motion monitoring and feedback system based on inertial sensors,which provides ideas for the cross-fusion of this technology with other technologies.

关 键 词：惯性传感器 驾驶动作监测 模式识别 特征选择 K最邻近算法 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]