基于压缩感知的水声通信网络流量预测.pdfVIP

fl!~，!! 中由于回 - 1. . … · 技术分析 · 基于压缩感知的水声通信网络 流量预测 * 4午两敏，关哲久，哈卡米 (浙江工业大学信息学院，浙江杭州 3 1 ∞23) [ 摘 要} 水声通信中过多的流量数据给采样和网络传输带来了困难，而压缩感知是一种可行的低速采样理论。 提 出基于贝叶斯压缩感知理论的水声通信网络流量预测方法。 将网络流量预测转化为贝叶斯压缩感知重构问题，为 了将需要笠构的向量稀疏化，将其设置为受超参数控制的后验概率密度函数。 通过该方法可以自适应地找出含有 重要信息的网络流量，并用回归算法来进行重构。 实验结果显示该方法具有较高的预测精度。 [ 关键词} 水声通信网络; 流量预测 ; 贝叶斯压缩感知; 相关向量机 { 中图分类号] TP393 ( 文献标志码] A [H)I : I0. 16311/j. au~oe.2015. 02. 12 Traf1ic Prediction ror Und erwater Acoustic Communicatioo Network Ba统d 00 Compr附ive Sensing XU Limin. WU Zhefu. HA Kami (Zhej iang University of Technology. 5chωI of Information. Hangzhou 3 1 ∞23. China ) ( Abstract] It is difficult to sample it directly and trans阳rt network as its t∞ much data in underwater acoustic communica- tion. however. comprωsed sensing ( C5) is a thωry that provides a feasib1e way with lower sampling speed . A traffic pre- diction method for underwater acoustic communication nework based on Bayesian compressed sensing ( BC5 ) is proposed in this paper. Traffic prediction is translated into BCS recon slruction problem. in order to make parse for the vector that demand reconstruction. it sets posterior probability density function that is controlled by hyperparameters to the vector. This method can find the critical network dala and reconslruct Ihem wilh regression algorilhm. The experimenl results show that this meth- od can 陀conslruct original underwater acouslic nelwork Iraffic errectively. ( Key words] underwater acouslic communication nelwork; traffic prediction; Bayesian comp