Scholarlink

Smart home devices and its applications have a great potential in successfully predicting the electricity consumption across households. The proposed framework comprises of a hardware setup for data acquisition and software setup for predicting the energy consumption. As an initial process, the energy consumption is measured and stored in a database with the help of PHP via the ESP 32 microcontroller. The AWS cloud service is used for users to view the consumption in real time. Data acquired is utilized by the server to run LSTM neural networks to predict possible future usage. The accuracy check is performed by comparing the predicted future value with the actual consumption. The proposed framework provides 98% accuracy using LSTM in training data and 99% in testing data.

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

LSTM-Based Deep Learning Model for Energy Management

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shridatha Hegde,
  • Suneeta V. Budihal,
  • Saroja V. Siddamal

摘要

Smart home devices and its applications have a great potential in successfully predicting the electricity consumption across households. The proposed framework comprises of a hardware setup for data acquisition and software setup for predicting the energy consumption. As an initial process, the energy consumption is measured and stored in a database with the help of PHP via the ESP 32 microcontroller. The AWS cloud service is used for users to view the consumption in real time. Data acquired is utilized by the server to run LSTM neural networks to predict possible future usage. The accuracy check is performed by comparing the predicted future value with the actual consumption. The proposed framework provides 98% accuracy using LSTM in training data and 99% in testing data.