基于卷积神经网络的煤矿支架搬运车轮胎助力机械手抓取姿态规划方法  

Grasping Attitude Planning Method of Tire Assisted Manipulator of Support Truck in Coal Mine Based on Convolutional Neural Network

作　　者：卢志琦 Lu Zhiqi(Shanxi Tiandi Coal Mining Machinery Co.,Ltd.,Taiyuan 030006,China;CCTEG Taiyuan Research Institute Co.,Ltd.,Taiyuan 030006,China;China National Engineering Laboratory for Coal Mining Machinery,Taiyuan 030006,China)

机构地区：[1]山西天地煤机装备有限公司,太原030006 [2]中国煤炭科工集团太原研究院有限公司,太原030006 [3]煤矿采掘机械装备国家工程实验室,太原030006

出　　处：《煤矿机械》2024年第2期179-182,共4页Coal Mine Machinery

基　　金：山西天地煤机装备有限公司工艺创新项目(GYCX-2022-06)。

摘　　要：为解决当前煤矿支架搬运车轮胎助力机械手抓取姿态规划精度较低、无法保证抓取轮胎受力均匀、导致轮胎部分位置受力过大、对轮胎产生了不同程度损害、消耗大量人力与物力资源的问题,提出基于卷积神经网络的轮胎助力机械手抓取姿态规划方法。引入卷积神经网络,通过提取目标物体位姿变化特征,得出目标物体在坐标系中的位置信息与姿态信息;基于卷积神经网络建立抓取预测模型,获取机械手抓取物体成功的概率;设计轮胎助力机械手的抓取轨迹,并设计机械手抓取姿态规划算法。通过试验分析可知,该方法在抓取过程中最大耗时与平均耗时均较短,抓取质量与效率得到了显著提升。In order to solve the problem that the current grasping attitude planning accuracy of the tire assisted manipulator of the support truck in coal mine is low,and the force on the grasping tire cannot be guaranteed evenly,resulting in excessive force on some parts of the tire,causing different degrees of damage to the tire,consuming a lot of human and material resources,a grasping attitude planning method of tire assisted manipulator based on convolutional neural network was proposed.The convolutional neural network was introduced,and the position information and attitude information of the target object in the coordinate system were obtained by extracting the pose change characteristics of the target object.Established a grasping prediction model based on convolutional neural network to obtain the probability of successful grasping of objects by the manipulator.Designed the grasping trajectory of the tire assisted manipulator,and designed the grasping attitude planning algorithm of the manipulator.Through the experimental analysis,this method has shorter maximum time and average time in the grasping process,and the grasping quality and efficiency have been significantly improved.

关 键 词：卷积神经网络 煤矿 支架搬运车 轮胎 抓取姿态 助力机械手 规划方法 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置] TH122[自动化与计算机技术—控制科学与工程]