Sparse vector coding(SVC)is emerging as a potential technology for short packet communications.To further improve the block error rate(BLER)performance,a uniquely decomposable constellation group-based SVC(UDCG-SVC)is...Sparse vector coding(SVC)is emerging as a potential technology for short packet communications.To further improve the block error rate(BLER)performance,a uniquely decomposable constellation group-based SVC(UDCG-SVC)is proposed in this article.Additionally,in order to achieve an optimal BLER performance of UDCG-SVC,a problem to optimize the coding gain of UDCG-based superimposed constellation is formulated.Given the energy of rotation constellations in UDCG,this problem is solved by converting it into finding the maximized minimum Euclidean distance of the superimposed constellation.Simulation results demonstrate the validness of our derivation.We also find that the proposed UDCGSVC has better BLER performance compared to other SVC schemes,especially under the high order modulation scenarios.展开更多
极端天气事件的发生会导致电力负荷产生突增或突降,对电网的稳定性和供电能力带来挑战。然而,现有的超短期负荷预测方法对极端天气下非线性和动态变化的负荷特征预测能力有限。为应对极端天气下负荷突变性强及波动剧烈导致的预测精度降...极端天气事件的发生会导致电力负荷产生突增或突降,对电网的稳定性和供电能力带来挑战。然而,现有的超短期负荷预测方法对极端天气下非线性和动态变化的负荷特征预测能力有限。为应对极端天气下负荷突变性强及波动剧烈导致的预测精度降低的问题,提出了一种考虑极端天气的二次重构分解去噪和双向长短时记忆网络(bidirectional long short-term memory,BiLSTM)的超短期电力负荷预测方法。首先,利用最大信息系数选取出能够最大程度反映对负荷影响的气候特征。然后,通过二次重构分解去噪方法提取到负荷多个频段的特征,降低数据复杂性,为BiLSTM模型提供更干净和信息量更清晰的输入序列,从而改善模型的训练效果和预测能力。最后基于比利时、福建省某区域以及得土安市的历史数据集进行算例分析,不同算例中平均绝对百分比误差分别下降到1.024%、0.875%、1.270%和1.009%,实验结果验证了所提方法在极端天气发生时的电力负荷超短期预测方面具有较好的预测性能和广阔的应用前景。展开更多
基金supported by the National Science Fundation of China(NSFC)under grant 62001423the Henan Provincial Key Research,Development and Promotion Project under grant 212102210175the Henan Provincial Key Scientific Research Project for College and University under grant 21A510011.
文摘Sparse vector coding(SVC)is emerging as a potential technology for short packet communications.To further improve the block error rate(BLER)performance,a uniquely decomposable constellation group-based SVC(UDCG-SVC)is proposed in this article.Additionally,in order to achieve an optimal BLER performance of UDCG-SVC,a problem to optimize the coding gain of UDCG-based superimposed constellation is formulated.Given the energy of rotation constellations in UDCG,this problem is solved by converting it into finding the maximized minimum Euclidean distance of the superimposed constellation.Simulation results demonstrate the validness of our derivation.We also find that the proposed UDCGSVC has better BLER performance compared to other SVC schemes,especially under the high order modulation scenarios.
文摘极端天气事件的发生会导致电力负荷产生突增或突降,对电网的稳定性和供电能力带来挑战。然而,现有的超短期负荷预测方法对极端天气下非线性和动态变化的负荷特征预测能力有限。为应对极端天气下负荷突变性强及波动剧烈导致的预测精度降低的问题,提出了一种考虑极端天气的二次重构分解去噪和双向长短时记忆网络(bidirectional long short-term memory,BiLSTM)的超短期电力负荷预测方法。首先,利用最大信息系数选取出能够最大程度反映对负荷影响的气候特征。然后,通过二次重构分解去噪方法提取到负荷多个频段的特征,降低数据复杂性,为BiLSTM模型提供更干净和信息量更清晰的输入序列,从而改善模型的训练效果和预测能力。最后基于比利时、福建省某区域以及得土安市的历史数据集进行算例分析,不同算例中平均绝对百分比误差分别下降到1.024%、0.875%、1.270%和1.009%,实验结果验证了所提方法在极端天气发生时的电力负荷超短期预测方面具有较好的预测性能和广阔的应用前景。
