预测修正下的动态数据传送机制

Dynamic Data Transmission Mechanism Based on Prediction Revision

利用无线传感器网络中节点感知数据的相关性,提出了一种基于预测修正的动态数据传送机制.核心思想是:将数据预测与模型计算分离,接收节点先对数据进行平稳化处理,然后建立模型或动态更新模型,再把模型参数发送给采样节点.采样节点用精简的预测修正算法预测数据,通过比较法确定需要发送的采样数据,从而减少了数据传送的次数,延长了网络的生命期.仿真结果表明,所提算法可以滤除真实采样观测序列中83%的冗余数据,预测精度较一般预测算法提高了22%,它适用于能量约束性较强的无线传感器网络.

A dynamic data transmission mechanism based on prediction revision is presented according to the inherent correlation of sampling data between nodes in WSNs. The core of the dynamic data transmission mechanism is to separate the data prediction and the model computing. The sampling data in sink node are firstly stabilized, the model is built or updated dynamically, and then the parameters are sent to sensor nodes by sink. The predictions in sensor nodes are made using a simplified prediction revision algorithm, and then, the sampling data that needs to be sent out is determined based on a comparison between the prediction data and the sampling data. The times of data transmission can be reduced and the lifetime of WSNs can be prolonged. Simulation studies show that the proposed algorithm can decrease times on sending data by more than 83% and the prediction precision can be increased by about 22% compared with the traditional prediction algorithms in real sampling data series. The algorithm is suitable for WSNs which are strictly constrained by energy.