基于改进KNN的离散型车间物流车定位系统  

Discrete Workshop Logistics Vehicle Positioning System Based on Improved KNN Algorithm

作　　者：朱萌 李敬兆[2] 李化顺 石晴 ZHU Meng;LI Jingzhao;LI Huashun;SHI Qing(School of Artificial Intelligence,Anhui University of Science and Technology,Huainan 232001,China;School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan 232001,China;Huaibei Hezhong Mechanical Equipment Co.,Ltd.,Huaibei 235000,China)

机构地区：[1]安徽理工大学人工智能学院,安徽淮南232001 [2]安徽理工大学电气与信息工程学院,安徽淮南232001 [3]淮北合众机械设备有限公司,安徽淮北235000

出　　处：《湖北民族大学学报（自然科学版）》2023年第3期368-374,共7页Journal of Hubei Minzu University:Natural Science Edition

基　　金：安徽理工大学横向科学研究项目(2021HT003)。

摘　　要：针对传统的K近邻法(K-nearest neighbors,KNN)室内定位算法精确度不高且无法解决车间内无线网络(wireless fidelity,Wi-Fi)特征维度不确定等问题,提出了基于余弦夹角的改进KNN定位算法(cosine angle based KNN,CA-KNN)的车间物流车定位系统。该系统首先利用余弦夹角将过滤后的样本空间进行分割,采用标签匹配机制克服Wi-Fi特征维度不确定问题,并判断待测点归属的样本空间区域,然后使用余弦夹角作为相似度选择K个近邻点,最后根据欧式距离判断待测点定位结果。实验结果表明,该系统在车间内具有良好的定位效果,相较于传统的KNN算法,在同等条件下,CA-KNN算法在待测点的最小误差为0.089 m,平均减少0.040 m,定位精确度最高可达92%。所研究的离散型车间物流车定位系统可满足数字化车间制造执行系统(manufacturing execution system,MES)研发、车间物流车定位等需求,为数字化车间提供更加智能化的服务。As the traditional K-nearest neighbors(KNN)indoor positioning algorithm was not accurate and can not solve the problems such as the uncertainty of the characteristic dimension of Wi-Fi in the workshop.This paper proposes an improved cosine angle based on KNN(CA-KNN)positioning algorithm is proposed to locate the logistics vehicle in the workshop.The system firstly uses cosine angle to segment the filtered sample space,and then uses the tag matching mechanism to overcome the uncertainty of the Wi-Fi characteristic dimension to judge the sample space region to which the measured points belong.Then,the cosine angle is used as the similarity to select K nearest neighbor points.Finally,the positioning results of the measured points are judged according to the Euclidean distance.The experimental results show that the system has a good positioning effect in the workshop.Compared with the traditional KNN algorithm,under the same conditions,the minimum error of CA-KNN algorithm at the point to be measured can reach 0.089 m,which reduces by 0.040 m compared with the similar algorithms,and the positioning accuracy can reach 92%.The discrete workshop logistics vehicle positioning system proposed in this paper can meet the requirements of digital workshop manufacturing execution system(MES)research and development,workshop logistics vehicle positioning and so on,and provide a more intelligent construction for the digital workshop.

关 键 词：车间物流车定位 KNN算法 特征维度 余弦夹角 欧氏距离 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]