基于人体体型差异的机器人运动系统设计  

Robot Motion System Design Based on Human Body Shape Difference

作　　者：郑煜涵 蒋婉玥 ZHENG Yu-han;JIANG Wan-yue(School of Automation,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学自动化学院,青岛266071

出　　处：《自动化与仪表》2023年第3期49-53,共5页Automation & Instrumentation

基　　金：国家重点研发项目(20YFB1313604)。

摘　　要：机器人与人类的交互在人们的日常生活中越来越普遍。该文针对目前机器人在人机交互过程中不能区分交互对象的缺陷,设计了一种基于人体体型差异的机器人运动系统。该系统的设计以YOLO-V5检测算法作为人体检测器,并配合RealSense深度摄头获取机器人与人之间的距离;以SimDR算法获取骨骼的关键点坐标,结合距离信息判断人体身高,并依据身高将人分为三种体型,使机器人根据不同体型的人做出三种不同的运动行为。该系统已在机器人平台上进行实验,实验结果满足机器人运动差异化表达的预期,并且系统具有良好的实时检测性能。The interaction between robots and humans is becoming more and more common in our daily life. This article has designed a robot motion system based on the human-computer interaction process that cannot distinguish between interaction objects in the process of human-computer interaction. The design of the system uses the YOLO-V5detection algorithm as the human detector and cooperates with the RealSense deep camera to obtain the distance between the robot and the person;obtain the key point coordinates of the bones with the SimDR algorithm,combine the distance information to determine the height of the human body,and depending on the height of the height divide people into three types,so that robots do three different sports behaviors according to people of different body types.The system has experimented on the robot platform. The experimental results meet the expectations of differentiated expression of robot movement,and the system has good real-time detection performance.

关 键 词：人机交互 人体体型 机器人运动系统 目标检测 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]