Deep learning approaches for speech emotion recognition: state of the art and research challenges

Speech emotion recognition (SER) systems identify emotions from the human voice in the areas of smart healthcare, driving a vehicle, call centers, automatic translation systems, and human-machine interaction. In the classical SER process, discriminative acoustic feature extraction is the most important and challenging step because discriminative features influence the classifier performance and decrease the computational time. Nonetheless, current handcrafted acoustic features suffer from limited capability and accuracy in constructing a SER system for real-time implementation. Therefore, to overcome the limitations of handcrafted features, in recent years, variety of deep learning techniques have been proposed and employed for automatic feature extraction in the field of emotion prediction from speech signals. However, to the best of our knowledge, there is no in-depth review study is available that critically appraises and summarizes the existing deep learning techniques with their strengths and weaknesses for SER. Hence, this study aims to present a comprehensive review of deep learning techniques, uniqueness, benefits and their limitations for SER. Moreover, this review study also presents speech processing techniques, performance measures and publicly available emotional speech databases. Furthermore, this review also discusses the significance of the findings of the primary studies. Finally, it also presents open research issues and challenges that need significant research efforts and enhancements in the field of SER systems.

     

Investigation on the cleaning of KDP ultra-precision surface polished with micro water dissolution machining principle

A potassium dihydrogen phosphate (KDP) optical crystal was machined to an ultra-precision surface with water-in-oil (W/O) micro emulsion polishing fluid. The micro water dissolution principle utilized in the machining process is discussed, its planarization mechanism is illustrated, and an ultra-precision polished surface with 2.205 nm RMS roughness is obtained. However, a substantial quantity of residual contamination remained on the polished surface after machining. This can seriously impact the optical performance of the crystal, and so it must be removed. Fourier transform infrared (FTIR) spectroscopy was used to conduct an investigation into the composition of the surface residue, and the results showed that the residue was comprised of organic chemicals with hydrocarbon chains and aromatic ether, i.e., mostly the polishing fluid. The cleaning method and the principle on which the KDP ultra precision surface investigation is based are discussed in detail, and the cleaning experiments with selected KDP-compatible organic solvents were then performed. FTIR transmittance spectra measurement and microscopic observations were employed to assess the effects of the cleaning process on the surface of the KDP crystal. The results showed that toluene cleaning achieved the most desirable results. This cleaning method produced a surface roughness of 1.826 nm RMS, which allows the KDP crystal to be applied to subsequent engineering applications.     

Optimization-based cosmetic formulation: Integration of mechanistic model,surrogate model,and heuristics

Multiple functional and hard-to-quantify sensorial product attributes that can be satisfied by a large number of cosmetic ingredients are required in the design of cosmetics. To overcome this problem, a new optimization-based approach for expediting cosmetic formulation is presented. It exploits the use of a hierarchy of models in an iterative manner to refine the search for creating the highest-quality cosmetic product. First, a systematic procedure is proposed for optimization problem formulation, where the cosmetic formulation problem is defined, design variables are specified, and a set of models for sensorial perception and desired product properties are identified. Then, a solution strategy that involves iterative model adoption and two numerical techniques (i.e., generalized disjunctive programming reformulation and model substitution) is applied to improve the efficiency of solving the optimization problem and to find better solutions. The applicability of the proposed procedure and solution strategy is illustrated with a perfume formulation example.     

Structure,crystallization, and performances of alkaline-earth boroaluminosilicate sealing glasses for SOFCs

We study the structure, crystallization, and performances of the sealing glasses with the composition (mol.%) of 12Al₂O₃·8B₂O₃·40SiO₂·40RO (R = Mg, Ca, Sr) for solid oxide fuel cells (SOFCs) before and after isothermal treatment at 700°C, which is within the operation temperature range (600-800°C) of SOFCs. The crystallization behavior has been investigated by differential scanning calorimetry and X-ray diffraction under both dynamic and isothermal conditions. The structural evolution is probed using the Raman and nuclear magnetic resonance spectroscopies. The performances of the sealing glasses are characterized in terms of the coefficient of thermal expansion, the crystallization-induced stress at glass–steel interface. We find that strong crystallization occurs at the operation temperature (700°C) far below the crystallization onset temperature measured by DSC. The structure origin of this anomalous crystallization is discussed in terms of structural heterogeneity of the three studied glasses. We determine the residual stress at the interface between the Ca-containing glass and the steel after isothermal treatment at 700°C for 48 h, but this stress does not lead to falling off the glass layer from the steel. This indicates that this glass is a good candidate to be applied in SOFCs.     

Engineering all-aromatic polyamide surface from hydrophilic to superhydrophobic and the accelerated strategy

Being a new kind of nanomaterials, aromatic polyamide nanofibers (ANF) have been much highlighted in recent studies. We here demonstrate an isopropyl alcohol (IPA) accelerated chemical cleavage on poly (p-phenylene terephthalamide) chopped fibers, which provides an efficient preparation method of ANF. The comprehensive study on the processes accelerated by different alcohols revealed that the preparation time of ANF in the mixed medium of dimethyl sulfoxide (DMSO)-alcohol (20:1 in volume) was shorten to 45 min and 75 min for methanol (ethanol) and isopropanol, respectively. However, the nanofibers prepared in DMSO-IPA exhibited the minimum in axial and radial dimensions, providing the finest and most uniform diameter of 16 nm. The corresponding ANF films through vacuum assisted filtration also showed the highest tensile strength of 150 MPa, in comparison with those of the ANF films prepared using other alcohols, which were about 110 MPa. Furthermore, ANF/silicon hybrid films were prepared by the ionic ring-opening reaction followed by the alkoxysilane condensation and nanoparticle fabrication. By changing the organo functional groups in the alkoxysilane, the surface of the films were adjustable in a wide contact angle range from 56° (hydrophilic) to 150° (superhydrophobic), suggesting the amendable interfacial properties potential applicable to composite fabrication with most of the resin matrix.     

负载换流逆变器驱动电励磁同步电机无速度传感器模型预测控制方法

负载换流逆变器(LCI)驱动电励磁同步电机控制系统广泛应用于中高压场合,在应对干扰和负载变化上,较大的系统惯性系数和较低的开关频率导致其动态响应性能欠佳.该文提出一种基于LCI驱动电励磁同步电机的无速度传感器模型预测控制方法.首先根据LCI的数学模型建立预测状态方程,同时给出合适的代价函数和约束条件;然后分析负载换相模式下的控制矢量模型,给出速度观测器的设计方法和参数选择依据,并针对负载变化对速度观测产生的影响进行分析和补偿.同时该观测器还能够同时实现逆变桥输入电压和端电压相位的观测;最后,通过3.75kW的LCI驱动电励磁同步电机实验平台对提出的无速度传感器模型预测控制方法进行实验验证.结果表明所提出的控制方法能够有效提升系统的动态响应性能,所设计的观测器具有可行性,速度观测结果在额定阶跃负载下误差小于13°.     

一种制备纳米复合镀层的新方法

研究了一种制备纳米复合镀层的新方法--在直流电镀过程中对电镀液施加强磁场以代替机械搅拌来制备Ni/Al2O3纳米复合镀层,考察了镀层的结构,并对镀层性能进行了研究.结果表明,在强磁场下可以制得纳米复合镀层;与未加磁场的纯镍镀层相比,施加磁场后镀层的晶体结构发生了择优取向;当磁场强度为8T时,复合镀层的磨损率仅为纯镍镀层的25%,其耐蚀性与Q235钢相比提高56.3%,与纯镍镀层相比提高46.8%.强磁场对电镀液起到较好的搅拌作用,通过影响镀层中纳米微粒的含量,使镀层结构发生变化,进而影响镀层的性能.     

机械合金化Al-Cu-Fe准晶相形成的研究进展

综述了机械合金化A1-Cu-Fe准晶相的形成机理及热力学稳定性方面的研究进展.     

车顶蒙皮焊接工艺改进

西安某汽车厂客车蒙皮组焊件焊接采用CO2焊,生产效率低,焊接变形大,表面处理难度大,质量难以保证,使产品的质量和产量受到严重的制约,为此必须进行工艺改进.通过采用全自动缝焊焊接工艺及选择合理的工艺参数,在保证结构和强度的前提下,使蒙皮组焊件焊接表面质量大幅度提高,劳动强度大幅度下降,生产效率提高数倍.     

基于TMS320C6678的多核程序加载研究与实现

针对多核D SP系统程序加载复杂的问题,基于T MS320C6678对多核程序加载进行了研究与设计。从一级引导程序出发,设计并优化了多核程序内容存储格式。设计了简洁的二级引导程序,以修正一级引导程序只识别主核程序入口地址,而从核入口地址缺失的现象。为了快速生成特定格式的多核程序内容,设计了多个工具用于添加S PI启动参数表、DDR3启动表、从核程序入口地址以及完成程序内容格式的转换。实现了S PI Flash多核程序加载以及基于I2C主模式的Nand Flash多核程序加载。