...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>IEEE sensors journal >Multi-Criteria Analysis of Sensor Reliability for Wearable Human Activity Recognition
【24h】

Multi-Criteria Analysis of Sensor Reliability for Wearable Human Activity Recognition

机译：可穿戴人类活动识别的传感器可靠性的多标准分析

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In Body Sensor Networks (BSNs), evaluating reliability of sensors is an important research topic which aims to optimize the overall performance of BSNs. Previous studies have often addressed this problem based only on a single criterion. However, it is often unreliable to rely on a single criterion to assess sensors in real situations. Accordingly, in this paper, we propose a novel multi-criteria approach for evaluating sensor reliability in activity recognition problem based on belief function theory. Specifically, in the theoretical part, we first describe the Multi-Criteria Analysis of Sensor Reliability (MCASR) using Belief Function based the Technique for Order Preference by Similarity to Ideal Solution (BF-TOPSIS). And in our proposed MCARS, two criteria are chosen in this work: 1) the conflict between sensor readings and, 2) the imprecision of sensor readings. In the application part, in order to prove the efficiency of MCASR, we propose a novel fused Long-Short Term Memory (LSTM) with MCASR to solve the problem of activity recognition. By using our proposed strategy, the final recognition accuracy has been significantly improved as compared with classical methods.
机译：在身体传感器网络（BSN）中，评估传感器的可靠性是一个重要的研究主题，旨在优化BSN的整体性能。以前的研究通常只基于单个标准解决了这个问题。但是，依赖于单个标准通常不可靠地评估实际情况中的传感器。因此，在本文中，我们提出了一种基于信念函数理论的活动识别问题中的传感器可靠性的新型多标准方法。具体地，在理论部分中，我们首先使用信仰功能基于通过相似性与理想解决方案（BF-Topsis）的顺序偏好的技术来描述传感器可靠性（MCASR）的多标准分析。在我们提出的MCAR中，在这项工作中选择了两个标准：1）传感器读数之间的冲突，2）传感器读数的不精确。在应用程序部分中，为了证明MCASR的效率，我们提出了一种新颖的融合长期记忆（LSTM）与MCASR以解决活动识别问题。通过使用我们提出的策略，与经典方法相比，最终识别准确性得到了显着提高的。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

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