Abstract: Autoregressive AR（p） model is an effective method to reduce the frequency of data communication and decrease the energy consumption in Wireless Sensor Network（WSN）. In view of the problem that AR（p） model ignores the different influence of historical data in different periods on the predictive value in the modeling process and affects the prediction accuracy of prediction model and the frequency of data communication in WSN, an improved AR（p） prediction model which is known as FAR（p） is proposed. By specifying a fuzzy membership value for each historical data with a new designed fuzzy membership function introduced in AR（p）, it can achieve the forecasting of historical data “farther is more weight”, in addition, with the data sequence processed by two weighted average algorithm a new prediction model is built. Simulation results show that the improved model can effectively improve the prediction accuracy, reduce data communication and decrease energy consumption in WSN.

Key words: AR（p） model, Wireless Sensor Network（WSN）, fuzzy membership function, prediction accuracy, data communication

摘要： 自回归AR（p）预测模型是无线传感网络（WSN）中一种减少数据传输次数和降低能量消耗的方法。针对AR（p）模型在建模过程中忽略了不同时期的历史数据对预测值的影响存在的差异，导致模型预测精度不高、网络通信频率受影响的问题，提出了一种改进的预测模型FAR（p）。在AR（p）模型中引入一种新的模糊隶属度函数，通过模糊隶属度函数对预测模型的每个历史数据赋予权值，实现历史数据“重近轻远”的预测效果，并经二次加权平均算法处理后重新构建预测模型。仿真结果表明，改进的预测模型有效地提高了模型预测精度，减少了传感网络中数据传输次数，降低了能量消耗。

关键词: AR（p）模型, 无线传感网络, 模糊隶属度函数, 预测精度, 数据通信