针对可收集能量传感器网络的能量预测方法

Energy Prediction Method for Harvested Energy Sensor Networks

为改善WCMA算法简单根据时间间隔和平均值计算GAP因子的不足,提出一种基于天气相似度的太阳能收集功率预测方法 D-WSMA。根据参考天和参考时刻对预测值的影响程度不同,采取相似度刻画数据间关系,动态调整参考天和参考时刻的权重以及历史参考时刻的加权平均值,从而得到可变化的DGAP因子。同时,根据数据波动性特征,改进原有算法中的固定权重α,得到动态变化权重因子dα。实验结果表明,D-WSMA预测精度相对WCMA算法提高了14.04%、28.30%、4.76%、12.58%,平均提高了15%。因此,D-WSMA预测方案具有良好性能,适合更加多样化的天气条件。

In order to improve the WCMA algorithm and calculate the GAP factor based on time interval and average value,this paper proposes a solar energy harvested power prediction method D-WSMA based on weather similarity.According to the influence of the reference day and the reference time on the predicted value,it adopts the similarity to characterize the relationship between the data, which can change the weight of the reference days and time slots.The weight for the average of the historical reference time slots contributed to introduce a changeable DGAP factor.At the same time,according to the volatility characteristics of the data,the fixed weights in the original algorithm are improved and can be a dynamic weighting factor.The experimental results show that the prediction accuracy of this scheme is 14.04%,28.30%,4.76%,12.58% higher than WCMA algorithm,and the average improvement is 15%. Hence,D-WSMA prediction method has good performance and can adapt to a wider range of weather conditions.