嵌入式网络通信远程数据传输功耗准确预测  

Embedded Network Communication Accurate Prediction of Power Consumption for Remote Data Transmission

作　　者：董旭斌 何世彪[2] 李成勇 陆鹏 DONG Xu-bin;HE Shi-biao;LI Cheng-yong;LU Peng(School of Electronic Information,Chongqing University of Engineering,Chongqing 400000,China;School of Resources and Security,Chongqing University,Chongqin 400000,China)

机构地区：[1]重庆工程学院电子信息学院,重庆400000 [2]重庆大学资源与安全学院,重庆400000

出　　处：《计算机仿真》2024年第11期545-549,共5页Computer Simulation

基　　金：重庆市教委科学技术研究计划项目(KJQN202301913);重庆市自然科学基金项目(cstc2021jcyj-msxmX0921)。

摘　　要：随着互联网应用逐渐覆盖到各个领域,为了使嵌入式网络通信系统拥有更高的性能和可靠性,在数据传输过程中对内存和功耗的要求不断提升。为了满足上述需求,以卷积神经网络作为数据预测方法,并结合GINI指数分类来实现对精准远程数据传输功耗预测。上述算法首先采用离散小波变换将连续传输信号进行离散化表示,以提升;然后采用功耗分析的精确度GINI指数分类算法,将离散据进行类别分割;最后利用CNN网络学习功耗数据的复杂模式和特征能力,引入卷积神经网络(CNN)功耗进行建模预测,提升预测远程数据传输功耗的精确度。提出的远程数据传输功耗预测算法仿真结果表明,上述算法在传播速率方面有明显提升,达到118.72Mbps的传输速率;且以上算法有效地增加了数据传输的容积,容量增加率高达96.79%。以上研究不仅提高了预测的准确性,还显著的提升了网络性能和数据传输效率,在功耗预测领域具有重要的研究价值。上述研究对于满足不断增长的通信需求和提高网络可靠性具有关键意义。As Internet applications gradually cover various fields,in order to make embedded network communica-tion systems have higher performance and reliability.Memory and power requirements are increasing during data transfer.In order to meet the above requirements,this paper uses convolutional neural network as a data prediction method,and combines GINI index classification to predict the power consumption of accurate long-distance data transmission.Firstly,the discrete wavelet transform is used to discretize the continuously transmitted signal to improve it.Then,the accurate GINI index classification algorithm of power consumption analysis is used to divide the discrete data into categories.Finally,the CNN network is used to learn the complex mode and feature capabilities of power consumption data,and the power consumption of the convolutional neural network(CNN)is introduced for modeling and prediction,which improves the accuracy of predicting the power consumption of remote data transmission.The simulation results of the long-range data transmission power prediction algorithm proposed in this paper show that the propagation rate of the algorithm is significantly improved,reaching a transmission rate of 118.72 Mbps,and the algo-rithm effectively increases the volume of data transmission,and the capacity increase rate is as high as 96.79%.This study not only improves the accuracy of prediction,but also significantly improves network performance and data transmission efficiency,which has important research value in the field of power consumption prediction.The research is key to meeting growing communication demands and improving network reliability.

关 键 词：卷积神经网络 无线远程 数据传输 功耗预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]