Tracking the maximum power point is a critical issue with solar systems.The power output of the solar panel varies due to variations in irradiance and temperature.Nonuniform irradiation due to partial shading conditio...Tracking the maximum power point is a critical issue with solar systems.The power output of the solar panel varies due to variations in irradiance and temperature.Nonuniform irradiation due to partial shading conditions has a direct impact on the characteristics of photovoltaic(PV)systems.To build a diversity of maximum power point tracking algorithms in solar PV systems,this work focuses on perturb and observe,incremental conductance,and fuzzy logic control methodologies.The suggested fuzzy logic control method outperformed the conventional incremental conductance and perturb and observe algorithms with a collection of 49 rules.This paper presents a novel series-parallel-cross-tied PV array configuration with a developed fuzzy methodology.To comment on the performance of a proposed system under various partial shading conditions,a series-parallel PV array configuration has been considered.The simulation result demonstrates that the fuzzy method has a percentage improvement in the global maximum power point tracking efficiency of 24.85%when compared to the perturb and observe method and a 65.5%improvement when compared to the incremental conductance method under long wide partial shading conditions.In the case of the middle partial shading condition,the fuzzy method has a percentage improvement in the global maximum power point tracking efficiency of 12.4%compared to the perturb and observe method and a 60.7%improvement compared to the incremental conductance method.展开更多
针对当前多模态情感识别算法在模态特征提取、模态间信息融合等方面存在识别准确率偏低、泛化能力较差的问题,提出了一种基于语音、文本和表情的多模态情感识别算法。首先,设计了一种浅层特征提取网络(Sfen)和并行卷积模块(Pconv)提取...针对当前多模态情感识别算法在模态特征提取、模态间信息融合等方面存在识别准确率偏低、泛化能力较差的问题,提出了一种基于语音、文本和表情的多模态情感识别算法。首先,设计了一种浅层特征提取网络(Sfen)和并行卷积模块(Pconv)提取语音和文本中的情感特征,通过改进的Inception-ResnetV2模型提取视频序列中的表情情感特征;其次,为强化模态间的关联性,设计了一种用于优化语音和文本特征融合的交叉注意力模块;最后,利用基于注意力的双向长短期记忆(BiLSTM based on attention mechanism,BiLSTM-Attention)模块关注重点信息,保持模态信息之间的时序相关性。实验通过对比3种模态不同的组合方式,发现预先对语音和文本进行特征融合可以显著提高识别精度。在公开情感数据集CH-SIMS和CMU-MOSI上的实验结果表明,所提出的模型取得了比基线模型更高的识别准确率,三分类和二分类准确率分别达到97.82%和98.18%,证明了该模型的有效性。展开更多
瞄准制衡强敌“马赛克战”“决策中心战”等新技术驱动的作战概念及威胁挑战,聚焦未来跨域联合作战指挥控制(command and control,C2)全流程决策需求,遵循平行智能理论框架,提出了基于筹划-准备-执行-评估(planning-readiness-execution...瞄准制衡强敌“马赛克战”“决策中心战”等新技术驱动的作战概念及威胁挑战,聚焦未来跨域联合作战指挥控制(command and control,C2)全流程决策需求,遵循平行智能理论框架,提出了基于筹划-准备-执行-评估(planning-readiness-execution-assessment,PREA)环与观察-判断-决策-行动(observe-orient-decide-act,OODA)环的平行指挥控制与管理(command&control and management,C2M)新范式,以期实现智能机器辅助指挥员作战全流程的分层次、个性化决策支持,减轻指挥员认知负担、降低决策复杂度,实现机器对指挥员“人脑”的智能扩展与增强,为塑造全局决策优势提供牵引和支撑。展开更多
文摘Tracking the maximum power point is a critical issue with solar systems.The power output of the solar panel varies due to variations in irradiance and temperature.Nonuniform irradiation due to partial shading conditions has a direct impact on the characteristics of photovoltaic(PV)systems.To build a diversity of maximum power point tracking algorithms in solar PV systems,this work focuses on perturb and observe,incremental conductance,and fuzzy logic control methodologies.The suggested fuzzy logic control method outperformed the conventional incremental conductance and perturb and observe algorithms with a collection of 49 rules.This paper presents a novel series-parallel-cross-tied PV array configuration with a developed fuzzy methodology.To comment on the performance of a proposed system under various partial shading conditions,a series-parallel PV array configuration has been considered.The simulation result demonstrates that the fuzzy method has a percentage improvement in the global maximum power point tracking efficiency of 24.85%when compared to the perturb and observe method and a 65.5%improvement when compared to the incremental conductance method under long wide partial shading conditions.In the case of the middle partial shading condition,the fuzzy method has a percentage improvement in the global maximum power point tracking efficiency of 12.4%compared to the perturb and observe method and a 60.7%improvement compared to the incremental conductance method.
文摘针对当前多模态情感识别算法在模态特征提取、模态间信息融合等方面存在识别准确率偏低、泛化能力较差的问题,提出了一种基于语音、文本和表情的多模态情感识别算法。首先,设计了一种浅层特征提取网络(Sfen)和并行卷积模块(Pconv)提取语音和文本中的情感特征,通过改进的Inception-ResnetV2模型提取视频序列中的表情情感特征;其次,为强化模态间的关联性,设计了一种用于优化语音和文本特征融合的交叉注意力模块;最后,利用基于注意力的双向长短期记忆(BiLSTM based on attention mechanism,BiLSTM-Attention)模块关注重点信息,保持模态信息之间的时序相关性。实验通过对比3种模态不同的组合方式,发现预先对语音和文本进行特征融合可以显著提高识别精度。在公开情感数据集CH-SIMS和CMU-MOSI上的实验结果表明,所提出的模型取得了比基线模型更高的识别准确率,三分类和二分类准确率分别达到97.82%和98.18%,证明了该模型的有效性。
文摘瞄准制衡强敌“马赛克战”“决策中心战”等新技术驱动的作战概念及威胁挑战,聚焦未来跨域联合作战指挥控制(command and control,C2)全流程决策需求,遵循平行智能理论框架,提出了基于筹划-准备-执行-评估(planning-readiness-execution-assessment,PREA)环与观察-判断-决策-行动(observe-orient-decide-act,OODA)环的平行指挥控制与管理(command&control and management,C2M)新范式,以期实现智能机器辅助指挥员作战全流程的分层次、个性化决策支持,减轻指挥员认知负担、降低决策复杂度,实现机器对指挥员“人脑”的智能扩展与增强,为塑造全局决策优势提供牵引和支撑。