首页> 外国专利> DEVICE SYSTEM METHOD AND PROGRAM FOR PROVIDING ONE-STOP ONLINE REPAIR SERVICE TO SPECIFY THE CAUSE OF FAILURE USING THE INSPECTION RESULTS OF INSPECTORS

DEVICE SYSTEM METHOD AND PROGRAM FOR PROVIDING ONE-STOP ONLINE REPAIR SERVICE TO SPECIFY THE CAUSE OF FAILURE USING THE INSPECTION RESULTS OF INSPECTORS

机译:一种利用检验员的检查结果,提供一站式在线维修服务,明确故障原因的装置系统方法及程序

摘要

According to an embodiment of the present invention, an apparatus for providing an online repair service is provided. The apparatus includes: a repair target specifying unit that specifies a repair target based on image data received from a customer terminal using a first artificial neural network; a preliminary failure type specifying unit that specifies at least one preliminary failure type matched with the repair target based on text data received from the customer terminal using a second artificial neural network; a required data request unit for specifying required data that is at least one of image data, video data, and sound data matched with the at least one preliminary failure type, and requesting the required data from a customer terminal; a failure type specifying unit for specifying at least one failure type based on the necessary data received from the customer terminal using a third artificial neural network; and requesting an inspection of the at least one failure type specific to a plurality of repair terminals, and determining a final failure type from the at least one failure type based on a plurality of inspection results received from the plurality of repair terminals. The first artificial neural network is a classification model machine-learned using a learning dataset generated by labeling a repair target with respect to image data for learning, and the second artificial neural network includes text data for learning. It is a classification model machine-learned using a training dataset generated by labeling a preliminary failure type for, and the third artificial neural network is a machine-learned classification model using a learning dataset generated by labeling necessary data for training and a failure type model, the inspection result includes information labeling correctness or incorrectness of the specified failure type, and the failure type inspection unit determines the ratio of the number of inspection results labeled as correct to the total number of the plurality of inspection results. If the accuracy of t is equal to or greater than the preset first reference value, the corresponding failure type is determined as the final failure type, and verification points are provided to the repair terminals that provide verification results labeled as correct for the final failure type.
机译:根据本发明的实施例,提供一种用于提供在线维修服务的装置。该装置包括:修复目标指定单元,该单元基于使用第一人工神经网络从客户终端接收的图像数据指定修复目标;一种初步故障类型指定单元,该单元根据使用第二人工神经网络从客户终端接收的文本数据指定至少一种与修复目标匹配的初步故障类型;所需数据请求单元,用于指定所需数据,该数据至少为与至少一种初步故障类型匹配的图像数据、视频数据和声音数据之一,并从客户终端请求所需数据;故障类型指定单元,用于根据使用第三人工神经网络从客户终端接收的必要数据指定至少一种故障类型;并请求对特定于多个维修终端的至少一种故障类型进行检查,并基于从多个维修终端接收的多个检查结果,从至少一种故障类型中确定最终故障类型。第一个人工神经网络是一个分类模型,使用通过标记修复目标生成的学习数据集进行学习,用于学习,第二个人工神经网络包括用于学习的文本数据。它是一种使用通过标记初始故障类型生成的训练数据集而生成的分类模型,第三人工神经网络是利用通过标记训练所需的数据生成的学习数据集和故障类型模型的机器学习分类模型,检查结果包括标记指定故障类型的正确性或不正确性信息, 故障型检测单元确定标记为正确的检测结果数量与多个检测结果总数的比值。如果t的精度等于或大于预设的第一参考值,则将相应的故障类型确定为最终故障类型,并向提供最终故障类型正确验证结果的维修终端提供验证点。

著录项

  • 公开/公告号KR1020230055337A;KR2023100055337A;KR20230055337A;KR20230055337

    专利类型

  • 公开/公告日2023-04-25

    原文格式PDF

  • 申请/专利权人 로봇앤컴주식회사;

    申请/专利号KR1020220038294;KR202200100038294;KR20220038294A;KR20220038294

  • 发明设计人

    申请日2022-03-28

  • 分类号G06Q10;G06N20;G06N3/08;G06Q10/06;G06Q10/10;

  • 国家

  • 入库时间 2024-06-14 23:56:13

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号