摘要
为满足水下目标识别抓取的自主性、准确性需求,设计了基于PLCnext水下目标抓取机器人。基于云边端协调设计理念,PLCnext边缘控制器向上通过工业以太网上传Proficloud云端,监控显示水下机器人各项数据,向下控制机械臂、双目相机、推进器等前端设备执行定位识别、抓取回收等指令。该水下机器人综合了机械臂控制、PLCnext、目标识别、双目测距等多种技术,基于YOLOv8和半全局块(SGM)算法对水下目标实现识别及立体测距,通过运动学逆解实现水下目标自主抓取。实验结果表明各种类矿石及矿石的识别正确率都在90%以上,抓取准确率在80%以上,满足了水下目标识别抓取的自主性、准确性需求,为水下目标自动抓取提供了一种新思路。
To address the requirements for autonomy and accuracy in underwater target identification and grasping retrieval,a PLCnext-based underwater grasping robot is designed.This robot employs a cloud-edge-end coordination approach,utilizing a PLCnext edge controller to upload data to the Proficloud cloud via industrial Ethernet for monitoring and displaying various data of the underwater robot while controlling front-end devices such as the robotic arm,stereo cameras,and thrusters for positioning and grasping tasks.Integrating various technologies such as robotic arm control,PLCnext,target recognition,and stereo ranging,the robot uses YOLOv8 and semiglobal matching(SGM)algorithms for underwater target identification and depth measurement.Autonomous target grasping is achieved through inverse kinematic.Experimental results indicate that the identification accuracy for various types of ores exceeds 90%,and the grabbing accuracy is over 80%.This performance meets the necessary autonomy and accuracy for underwater target retrieval,presenting a novel solution for automated underwater target grasping.
作者
张衍
王显峰
李浩男
薛垚家
王梁
ZHANG Yan;WANG Xianfeng;LI Haonan;XUE Yaojia;WANG Liang(College of Intelligent Systems Science and Engineering,Harbin Engineering University,Harbin 150001,China;"Harbin Engineering University Phoenix Contact"Industry 4.0 Demonstration and Innovation Center,Harbin 150001,China)
出处
《传感器与微系统》
北大核心
2025年第S1期88-93,共6页
Transducer and Microsystem Technologies
基金
“两高两智”牵引驱动自动化类新工科人才培养实验教学体系2.0改革与实践(SYJX2024—063)
关键词
机械臂
立体匹配
PLCnext
人机交互
mechanical arm
stereo matching
PLCnext
human-computer interaction
