Scholarlink

Crop recommendation is one of the substantial research areas of Agriculture 5.0. In conventional IoT-based crop recommendation systems, the information storage and analysis happen inside the cloud that has several shortcomings such as connection interruption, high latency, huge overhead on the cloud, etc.. To overcome these issues, this paper proposes a crop yield prediction system FemCrop based on femtocell, edge computing, and deep learning. The femtocell is a low-power base station that serves an intermediary device in FemCrop for facilitating seamless data transmission between IoT devices and the edge server by enhancing signal strength. The edge server utilizes sophisticated deep-learning methods to analyze data and produce recommended outcomes that are sent to the cloud. Users can access these predicted outcomes from the cloud for selection regarding appropriate crop. The outcomes indicate that FemCrop achieves \(\sim \) 99% prediction accuracy, and reduces \(\sim \) 30% latency and \(\sim \) 10% energy consumption than the conventional edge-cloud system.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

FemCrop: A Femtocell-Based Edge-Cloud Frame-Work for Crop Yield Prediction Using Deep Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Tanushree Dey,
  • Somnath Bera,
  • Anwesha Mukherjee,
  • Samarjit Roy,
  • Debashis De

摘要

Crop recommendation is one of the substantial research areas of Agriculture 5.0. In conventional IoT-based crop recommendation systems, the information storage and analysis happen inside the cloud that has several shortcomings such as connection interruption, high latency, huge overhead on the cloud, etc.. To overcome these issues, this paper proposes a crop yield prediction system FemCrop based on femtocell, edge computing, and deep learning. The femtocell is a low-power base station that serves an intermediary device in FemCrop for facilitating seamless data transmission between IoT devices and the edge server by enhancing signal strength. The edge server utilizes sophisticated deep-learning methods to analyze data and produce recommended outcomes that are sent to the cloud. Users can access these predicted outcomes from the cloud for selection regarding appropriate crop. The outcomes indicate that FemCrop achieves \(\sim \) 99% prediction accuracy, and reduces \(\sim \) 30% latency and \(\sim \) 10% energy consumption than the conventional edge-cloud system.