基于支持向量机的物流传送带场景下RFID识读性能研究  

Study on SVM⁃based RFID reading performance in logistics conveyor belt scenario

作　　者：王耀伟 刘原麟 潘若禹 陈宝文 罗方欣 WANG Yaowei;LIU Yuanlin;PAN Ruoyu;CHEN Baowen;LUO Fangxin(School of Communications and Information Engineering,Xi’an University of Posts and Telecommunications,Xi’an 710121,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121

出　　处：《现代电子技术》2024年第9期150-156,共7页Modern Electronics Technique

摘　　要：为避免调整参数布置导致多次验证实验,探寻各因素对识读效率的影响程度,提出一种基于支持向量机的动态RFID系统识读效率分类模型。建立支持向量机模型预动态场景下的RFID识读效率,并使用网格搜索法寻优模型。结果表明,基于支持向量机的动态RFID系统识读效率分类准确率可达92.3%。使用此模型对各影响因素进行分析,可以得到动态场景下对RFID识读效率影响最大的因素,并在动态场景中能够依此进行最佳参数设置。A dynamic radio frequency identification(RFID)system reading efficiency classification model based on support vector machine(SVM)is proposed to avoid multiple verification experiments caused by parameter arrangement adjustment and explore the influence of various factors on reading efficiency.The SVM model is established to predict the reading efficiency of RFID in pre⁃dynamic scenarios,and the grid search method is used to optimize this model.The results show that the classification accuracy of the dynamic SVM⁃based RFID system's reading efficiency can reach 92.3%.This model is used to analyze various influencing factors,and the factors with the greatest impact on RFID reading efficiency in dynamic scenarios can be obtained,based on which the best parameters can be selected in dynamic scenes.

关 键 词：物流传送带场景 射频识别 高斯滤波 支持向量机 动态RFID系统 朴素贝叶斯 盘点 RFID识读率 

分 类 号：TN99-34[电子电信—信号与信息处理]