• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Sensing and imaging >Hybrid Neural Network Based Wideband Spectrum Behavior Sensing Predictor for Cognitive Radio Application
【24h】

Hybrid Neural Network Based Wideband Spectrum Behavior Sensing Predictor for Cognitive Radio Application

机译：基于混合神经网络的认知无线电应用的宽带频谱行为传感预测因子

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

The work behind research is energy and time efficient spectrum occupancy analysis for the deployment of cognitive radio (CR) in India. The measurement campaign is initial step, which provides data for analysis. The campaigns conducted worldwide lead to a complex measurement set up. So, first aim was to provide solution for simple and compact measurement set up. A new wideband circularly polarized microstrip antenna is proposed instead of existing commercially available antennas; which show better spectrum sensing ability. The conventional spectrum occupancy analysis consumes more sensing time as well as network resources because it senses whole spectrum every time. This problem can be overcome by adopting predictive method to predict the behavior of spectrum. A new hybrid neural network (HNN) model is proposed as a predictor for spectrum behavior sensing through which the status of different channels can be predicted by proper learning. The required database for training and validation of HNN predictor was collected from the measurement campaign conducted first time for seven days at Solapur city, India. The HNN model performance was examined for popular bands, and different days (weekdays and weekend) using root mean square error (RMSE) performance metric. The result shows, it has greater ability as spectrum behavior sensing predictor, and would be better choice for energy as well as time efficient spectrum occupancy analysis for cognitive radio application.
机译：研究背后的工作是在印度部署认知无线电（CR）的能量和时间高效频谱占用分析。测量活动是初始步骤，提供分析数据。全球活动的活动导致复杂的测量设定。因此，首先是为简单而紧凑的测量设置提供解决方案。提出了一种新的宽带圆极化微带天线，而不是现有的市售天线;显示出更好的光谱感测能力。传统的频谱占用率分析消耗更多的感测时间以及网络资源，因为它每次都感应整个光谱。通过采用预测方法来预测频谱的行为，可以克服这个问题。提出了一种新的混合神经网络（HNN）模型作为频谱行为感测的预测器，通过该频谱行为感测，可以通过适当的学习来预测不同信道的状态。来自印度索尔卡尔市第一次在第一次进行的测量活动中收集了用于培训和验证HNN预测的数据库。使用root均方误差（RMSE）性能度量，对流行乐队和不同的日子（平日和周末）检查了HNN模型性能。结果表明，它具有更大的能力作为频谱行为传感预测因子，并且可以更好地选择能源以及认知无线电应用的时间有效的频谱占用分析。

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