语音通信降噪研究

语音通信系统中，语音通过信道传输将不可避免地引入码间串扰和信号畸变，同时受到噪声污染。本文在分析自适应盲均衡算法CMA(constant modulus algorithm）和改进盲均衡算法的基础上，考虑到自适应盲均衡技术在语音噪声控制方面能力有限，将自适应盲均衡技术与小波包掩蔽阈值降噪算法联合使用，形成一种基带语音增强新方法。仿真试验结果显示自适应盲均衡技术可以使星座图变得清晰而紧凑，有效减小误码率。研究证实该方法在语音信号ISI和畸变严重情况下，在白噪及有色噪声不同的噪声环境中都具有稳定的降噪能力，消噪同时可获得汉语普通话良好的听觉效果。     

A Ru(II)-Based Nanoassembly Exhibiting Theranostic PACT Activity in NIR Region

Photoactivated chemotherapy (PACT) has appealing merits over traditional chemotherapy as well as photodynamic therapy (PDT) by virtue of its spatial and temporal control on drug activity and oxygen-independent mechanisms of action. However, the short photoactivation wavelengths, e.g., visible light–activated Ru(II)-based PACT agents, limit the clinical application severely. In this work, a facile construction of supramolecular nanoparticles from a poly(ethylene glycol) (PEG)-modified [Ru(dip)₂(py-SO₃)]⁺ (abbreviated as Ru-PEG, dip = 4,7-diphenyl-1,10-phenanthroline, py-SO₃ = pyridine-2-sulfonate) and 1,3-phenylenebis(pyren-1-ylmethanone) (BP) is shown. While Ru-PEG may undergo photoinduced ligand dissociation and release anticancer species of [Ru(dip)₂(H₂O)₂]²⁺, BP has extremely large two-photon absorption cross sections (δ₂) in the NIR region and intense fluorescence over the wavelengths where Ru-PEG has strong absorption. Thus, two-photon excitation of BP followed by an efficient Förster resonance energy transfer (FRET) from BP to Ru-PEG may lead to a potent inactivation against cisplatin-resistant cancer cells and 3D multicellular tumor spheroids (MCTSs). The residue fluorescence of BP also allows the cellular uptake of the particles to be visualized. This work provides a universal and convenient strategy to realize theranostic PACT in the ideal phototherapeutic window of 650–900 nm.     

三维位势问题的梯度边界积分方程的新解法

三维位势问题的边界元分析中,关于坐标变量的边界位势梯度的计算是一个困难的问题. 已有一些方法着手解决这个问题,然而,这些方法需要复杂的理论推导和大量的数值计算. 本文提出求解一般边界位势梯度边界积分方程的辅助边值问题法. 该方法构造了与原边界值问题具有相同解域的辅助边值问题,该辅助边值问题具有已知解,因此通过求解此辅助边值问题,可获得梯度边界积分方程对应的系统矩阵,然后将此系统矩阵应用于求解原边值问题,求解过程非常简单,只需求解一个线性系统即可获得原边值问题的解. 值得注意的是,在求解原边值问题时,不再需要重新计算系统矩阵,因此辅助边值问题法的效率并不很差. 辅助边值问题法避免了强奇异积分的计算,具有数学理论简单、程序设计容易、计算精度高等优点,为坐标变量梯度边界积分方程的求解提供了一个新的途径. 3个标准的数值算例验证了方法的有效性.      

Unsymmetrical Pyrene‐Fused Phthalocyanine Derivatives: Synthesis,Structure, and Properties

Novel pyrene‐fused unsymmetrical phthalocyanine derivatives 2,3,9,10,16,17‐hexakis(2,6‐dimethylphenoxy)‐22,25‐diaza(2,7‐di‐tert‐butylpyrene)[4,5]phthalocyaninato zinc complex Zn[Pc(Pz‐pyrene)(OC₈H₉)₆] ( 1 ) and 2,3,9,10‐tra(2,6‐dimethylphenoxy)‐15,18,22,25‐traza(2,7‐di‐tert‐butylpyrene)[4,5]phthalocyaninato zinc compound Zn[Pc(Pz‐pyrene)₂(OC₈H₉)₄] ( 2 ) were isolated for the first time. These unsymmetrical pyrene‐fused phthalocyanine derivatives have been characterized by a wide range of spectroscopic and electrochemical methods. In particular, the pyrene‐fused phthalocyanine structure was unambiguously revealed on the basis of single crystal X‐ray diffraction analysis of 1 , representing the first structurally characterized phthalocyanine derivative fused with an aromatic moiety larger than benzene.     

Yolk‐Shell Porous Microspheres of Calcium Phosphate Prepared by Using Calcium L‐Lactate and Adenosine 5′‐Triphosphate Disodium Salt: Application in Protein/Drug Delivery

A facile and environmentally friendly approach has been developed to prepare yolk‐shell porous microspheres of calcium phosphate by using calcium L ‐lactate pentahydrate (CL) as the calcium source and adenosine 5′‐triphosphate disodium salt (ATP) as the phosphate source through the microwave‐assisted hydrothermal method. The effects of the concentration of CL, the microwave hydrothermal temperature, and the time on the morphology and crystal phase of the product are investigated. The possible formation mechanism of yolk‐shell porous microspheres of calcium phosphate is proposed. Hemoglobin from bovine red cells (Hb) and ibuprofen (IBU) are used to explore the application potential of yolk‐shell porous microspheres of calcium phosphate in protein/drug loading and delivery. The experimental results indicate that the as‐prepared yolk‐shell porous microspheres of calcium phosphate have relatively high protein/drug loading capacity, sustained protein/drug release, favorable pH‐responsive release behavior, and a high biocompatibility in the cytotoxicity test. Therefore, the yolk‐shell porous microspheres of calcium phosphate have promising applications in various biomedical fields such as protein/drug delivery.     

卷积神经网络用于近红外光谱古筝面板木材分级

目前,我国乐器制作行业在古筝面板用木材等级的筛选上主要依赖于技师主观评判,但此法缺少科学理论的依据,效率低,客观性及出材率的提高等方面受到限制,无法满足乐器市场的大量需求。实现古筝面板用木材快速、智能化的分级工作是一个急需解决的课题。近红外光谱非常适用于测量含氢的有机物质。古筝面板木材主要化学成分的化学键均由含氢基团组成,不同等级板材的化学成分存在差异,这些差异反映在近红外光谱中,为判断木材等级提供了可能。同时卷积神经网络对非线性数据具有较强的特征提取能力,所以提出一种应用卷积神经网络模型对光谱数据进行分析的方法,进而判别木材的等级。应用了Savitzky Golay一阶、二阶微分两种预处理方法和核主成分分析、连续投影算法两种数据压缩方法,通过所设计的卷积神经网络模型以样本识别准确率和模型构建过程中的损失值作为判定指标选出最佳预处理和数据压缩方法。为了提高模型提取分析光谱数据的能力和避免过拟合现象,应用了多通道卷积核、批量归一化和early stopping策略,将通过两层卷积层提取的特征信息送入全连接层,从而充分提取剩余信息,通过Softmax函数获得板材的最终预测等级,从而确定了最终模型。最终Savitzky Golay一阶微分和核主成分分析为最佳数据处理方法,同时得出用于区分不同等级的古筝面板用木材的主要关键谱带,分别为1 163~1 243, 1 346~1 375和1 525~1 584 nm。将该模型应用于测试集样本,古筝面板用木材的等级识别准确率为95.5%。实验结果表明所提出的方法可以高效地处理光谱数据,有效识别区分不同等级的古筝面板用木材的关键特征,从而为广阔的乐器市场提供一定的技术支持。     

结合VEC和信心水平分析Black-Litterman投资组合模型

在理论上通过推导得出了Black-Litterman模型(B-L模型)最优权重与信心水平的关系式.实证部分开创性地将多元时间序列VEC模型运用到B-L模型的观点收益预测中.结果表明:且通过向量误差修正模型,实证取得了较好的效果.在任意一种做空限制下,随着信心水平在0%~70%水平上升,收益率有连续上升的趋势,并逐渐趋稳;风险有并不明显的下降的趋势.对同一信心水平而言,随做空限制的放宽,收益率有上升趋势.     

多孔复合钙铝氧化物的制备、表征及在重整生成油中的脱氯应用

采用低廉的铝盐和钙盐,通过共沉淀法制备了一系列不同CaO含量的CaO-Al₂O₃复合氧化物,并用XRD、TG-DTG、N₂吸附-脱附、SEM等手段对其进行表征。结果表明,制备的复合氧化物两相分布均匀、孔隙率较高、存在较多的介孔和大孔,比表面积也较大。采用动态吸附法将制备出的复合氧化物应用于重整生成油中氯化氢的脱除反应中,结果表明,高比表面积和大孔容的CAO-1具有最佳的HCl脱除效果,当反应温度为55 ℃、液空速为3 h^-1,重整生成油中氯含量小于等于15 ng/μL时,其氯容可达到18%。