朱里程, 李鹏斐, 李孟轩, 唐凯, 李静楠, 陈曦. 基于静电感应信号的路面识别方法[J]. 机器人, 2017, 39(1): 57-62. DOI: 10.13973/j.cnki.robot.2017.0057
引用本文: 朱里程, 李鹏斐, 李孟轩, 唐凯, 李静楠, 陈曦. 基于静电感应信号的路面识别方法[J]. 机器人, 2017, 39(1): 57-62. DOI: 10.13973/j.cnki.robot.2017.0057
ZHU Licheng, LI Pengfei, LI Mengxuan, TANG Kai, LI Jingnan, CHEN Xi. A Road Identification Method Based on Electrostatic Induction Signal[J]. ROBOT, 2017, 39(1): 57-62. DOI: 10.13973/j.cnki.robot.2017.0057
Citation: ZHU Licheng, LI Pengfei, LI Mengxuan, TANG Kai, LI Jingnan, CHEN Xi. A Road Identification Method Based on Electrostatic Induction Signal[J]. ROBOT, 2017, 39(1): 57-62. DOI: 10.13973/j.cnki.robot.2017.0057

基于静电感应信号的路面识别方法

A Road Identification Method Based on Electrostatic Induction Signal

  • 摘要: 提出了一种基于静电信号测量技术的路面识别新方法,该方法能够对机器人经常接触的4类典型室外路面环境——地砖、沙地、草地、沥青路面——进行有效识别.分析了机器人足部与路面接触/分离过程中的感应电荷及泄放情况,建立了感应电荷理论模型.通过仿真发现,不同路面材料的表面的电荷泄放特性存在明显差异.在此基础上,通过设计模拟测试系统,模拟金属电极与不同的路面的接触/分离过程.采集4类路面多组静电信号,并提取信号的特征值作为分类器参量.使用k最近邻分类算法对路面静电信号进行识别分类,识别结果显示平均正确识别率达83.3%.

     

    Abstract: A novel method for material identification based on electrostatic signal detection technology is presented. 4 kinds of typical roads, i.e. brick, sand, grass and asphalt, which can be often encountered in outdoor environment, are effectively identified using the proposed method. The induced charge change on robot foot is analyzed by establishing an equivalent model for the contact/separation process between the robot foot and the road surface. The simulation result shows that there are obvious differences in the discharge of surface charge of different pavement materials. Based on that, a special structure of measurement platform is proposed for simulation of contact and separation between robot foot and roads. 4 kinds of road surface electrostatic signals are collected, and the feature value of the signal is extracted as the classifier parameter. The k-nearest neighbor classifier is used to classify the road surface electrostatic signals. The result shows that the overall recognition rate is about 83.3%.

     

/

返回文章
返回