论文部分内容阅读
针对如何兼顾WSN数据融合算法的高精度和低复杂度的要求,提出一种基于ARMA模型的低阶次WSN数据融合算法.首先介绍ARMA时间序列分析模型,利用节点采集的数据的强时空相关性预测未来数据;其次用BIC检验准则逐步升阶确定自回归部分阶次后,再用F-检验逐步降低滑动平均部分的阶次;最后利用节点的预测模型修正算法和簇头节点的预测融合算法确定适合传感器节点的ARMA模型.实验结果表明,该算法能够在保证融合高精确度的情况下,降低了预测模型的复杂度,降低了对处理能力和存储能力的需求,节约了能量.
Aiming at the high accuracy and low complexity requirements of WSN data fusion algorithm, a low-order WSN data fusion algorithm based on ARMA model is proposed.Firstly, the ARMA time series analysis model is introduced, and the strong spatiotemporal correlation of the data collected by nodes Predict the future data; secondly, using the BIC test criterion to ascertain the autoregressive partial order, and then use the F-test to gradually reduce the order of the moving average; Finally, using the node prediction model correction algorithm and cluster head node prediction fusion algorithm The ARMA model suitable for sensor nodes is determined.The experimental results show that the proposed algorithm can reduce the complexity of the prediction model and reduce the processing power and storage capacity while saving the energy with high accuracy.