采摘机器人对猕猴桃果实硬度触觉感知研究  

Research on Tactile Perception of Kiwifruit Hardness by Picking Robot

作　　者：严正红 陈琳 Yan Zhenghong;Chen Lin(School of Electrical Technology,Anhui National Defense Technology Vocational College,Lu'an,Anhui 237011,China;School of Electronic Science and Applied Physics,Hefei University of Technology,Hefei,Anhui 230002,China)

机构地区：[1]安徽国防科技职业学院电气技术学院,安徽六安237011 [2]合肥工业大学电子科学与应用物理学院,安徽合肥230002

出　　处：《黑龙江工业学院学报（综合版）》2022年第7期84-88,共5页Journal of Heilongjiang University of Technology(Comprehensive Edition)

基　　金：安徽省高校自然科学研究重点项目“农业机器人触觉感知与柔性抓取控制方法研究”(项目编号:KJ2019A1187);安徽省高校自然科学研究重点项目“智能家居环境监测系统的中间件研究”(项目编号:KJ2020A1089);安徽国防科技职业学院2021年校级质量工程项目“新工科背景下《工业机器人应用系统建模》课程教学‘双线’融合探索与实践”(项目编号:gf2021jxyjy09)。

摘　　要：针对采摘机器人很难对果实硬度属性进行触觉抓取感知的问题,提出了一种基于主成分分析和K近邻相结合算法的猕猴桃硬度识别方法。该方法通过机械手集成的触觉传感器获取抓取接触信息,将接触数据预处理选取出高维特征序列,然后利用PCA对特征序列进行降维处理,最后通过训练KNN分类器完成识别分类。以猕猴桃为例,进行抓取识别分类,实验结果表明,该方法对猕猴桃硬度感知识别最优准确率达90.03%,实现了采摘机器人对猕猴桃硬度属性的智能感知。Aiming at the problem that it is difficult for the picking robot to grasp and perceive hardness property of the fruit,a kiwifruit hardness recognition method based on the combination of principal component analysis and K-nearest neighbor algorithm is proposed.In this method,the tactile sensor integrated by the manipulator obtains the grasping contact information,preprocesses the contact data and selects the high-dimensional feature sequence;Then,PCA is used to reduce the dimension of the feature sequence.Finally,KNN classifier is trained to complete the recognition and classification.Taking kiwifruit as an example,grasping recognition and classification are carried out.The experimental results show that the optimal accuracy of this method for kiwifruit hardness perception recognition is 90.03%,and the intelligent perception of kiwifruit hardness attribute is realized by picking robot.

关 键 词：PCA KNN 触觉信息 猕猴桃 硬度 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置] TP391.4[自动化与计算机技术—控制科学与工程]