基于多时间尺度特征的语音识别模型

提出了基于多时间尺度特征的语音识别模型.该模型采用描述谱参数轨迹的段特征,在段尺度上实现了对语音信号帧间相关性的显式建模;采用段特征依赖的非平稳时间序列产生模型,实现了不同尺度特征间的相关性建模,并在帧尺度上通过参数化的均值轨迹函数,实现了对语音信号帧间相关性的隐式建模.给出了基于多时间尺度特征联合统计距离优化的分段算法及基于最大似然准则的模型参数估计算法.识别实验表明,该模型的识别性能优于标准HMM及趋势HMM.     

Charge-product control approach to electrostatic leader-follower formations in LEO plasma wakes

The feasibility of using electrostatic forces to stabilize a close-proximity leader-follower formation is investigated. The leader craft is equipped with a set of affixed spheres whose charge is modulated to hold the charged follower craft along a proscribed trajectory to its nominal leader-relative position. This charge structure and the follower craft are constrained to remain in the plasma wake generated behind all LEO craft because the more-dense ambient plasma outside the wake prevents object charging and electric field propagation. Once the formation is achieved, a controlled electric field is generated by the leader to counter relative accelerations from perturbations like differential drag and solar radiation pressure, holding the follow near its nominal position. Two controllers are derived for the system described, incorporating Coulomb accelerations and linearized gravity and drag accelerations. Simulations are run under unmodeled perturbations and sensor noise for different scenarios, demonstrating the challenges and benefits associated with electrostatic actuation.     

Performance analysis of variable speed tail rotors with Gurney flaps

Gurney Flaps (GFs) are used for improving the performance of variable speed tail rotors. A validated analytical helicopter model able to predict the main and tail rotor power is utilized. The fixed height GF has substantially small influence on the tail rotor power in hover and low to medium speed forward flight, and can obtain significant power reduction in high speed flight. This ability can be enhanced by decreasing the tail rotor speed. With the deployment of GF, the collective pitch of the tail rotor decreases, and the maximum tail rotor thrust increases. The GF can compensate the reduction of the maximum thrust by the decrease in the tail rotor speed. The GF with a height of 5% of the chord length can almost remedy 50% of the thrust reduction introduced by decreasing 10% of the tail rotor speed. With the increase of GF height, the maximum thrust generated by the tail rotor increases. The GF with larger height can cause the increase in the tail rotor power in hover and low to medium speed flight. The retractable GF can obtain more power savings than the fixed height GF. However, the benefit is substantially small even in high speed flight. Considering the side effects introduced by the active GF, the fixed height GF may be more preferable. The mechanism for the retractable GF to generate more tail rotor thrust is to increase the lift in advancing side due to the higher dynamic pressure.     

基于碱金属辅助电离的等离子鞘套模拟方法

为解决再入过程中通信黑障与探测异常问题，对等离子体鞘套地面模拟方法进行了研究。以RAM C-II 飞行试验为依据，计算了等离子体鞘套的参数范围。分析现有模拟设备，概括提出了高焓低速风洞的模拟方案，选取相适应的放电方式，并对流量功率进行了计算。针对高温环境对电磁设备造成损坏以及所需功率难以实现的问题，提出添加碱金属的解决方案，并进行了试验验证。试验结果表明，添加碱金属可以使电子密度提升约一个数量级。对方案效果进行评估，预期功率需求降至兆瓦量级。     

A method for automatic detection and characterization of plasma bubbles using GPS and BDS data

Detecting and characterizing Total Electron Content (TEC) depletion is important for studying the ionospheric threat due to the Equatorial Plasma Bubble (EPB) when applying the Ground-Based Augmentation System (GBAS) at low latitudes. This paper develops a robust method to automatically identify TEC depletion and derive its parameters. The rolling barrel algorithm is used to automatically identify the TEC depletion candidate and its parameters. Then, the depletion candidates are screened by several improved techniques to distinguish actual depletions from other phenomena such as Traveling Ionospheric Disturbance (TID) or abnormal data. Next, based on the depletion signals from three triangular receivers, the method derives EPB parameters such as velocity, width and gradient. The time lag and front velocity are calculated based on cross-correlation using TEC depletions and the geometrical distribution of three triangular receivers. The width and gradient of slope are then determined by using TEC depletion from a single receiver. By comparison, both the station-pair method and proposed method depend on the assumption that the EPB morphology is frozen during the short time when the plasma bubble moves between the receivers. However, our method relaxes the restriction that the baseline length should be shorter than the width of slope required by the station-pair. This relaxation is favorable for studying small-scale slope of depletions using stations of a longer baseline. In addition, the accuracy of the width and gradient is free of impact from hardware biases and small-scale disturbance, as it is based only on the relative TEC variation. The method is demonstrated by processing Global Positioning System (GPS) and BeiDou Navigation Satellite System (BDS) data on 15 August, 2018, in a solar minimum cycle.     

基于虚位移原理的宏细观统一本构模型

结合虚位移原理和宏细观统一本构模型的基本理论提出了一种宏细观统一本构模型.该模型将宏观位移表示为宏观应变与细观位移的线性组合,然后应用虚位移原理建立基本方程,采用有限元法进行求解,得到宏观应变增量和应力增量的关系式.通过算例分析对比了模型与FVDAM(finite volume directe average meth...     

Results of measurements of the response of a delta-doped CCD to neutral and charged particle beams

We report the measurements of the response of a delta-doped Charge Coupled Device (CCD) in imaging mode to beams of charged and neutral particles. That is, the detector imaged the incident beam over its 1024 × 1024 pixels, integrating the number of particles counted in each pixel during the exposure period. In order to count individual particles the exposure time would have had to be reduced considerably compared to the typical ?5 s used in these studies. Our CCD thus operated in a different manner than do conventional particle detectors such as the CEM and MCP that normally are used in a particle counting mode. The measurements were carried out over an energy range from 0.8 to 30 keV. The species investigated include H, H⁺, He⁺, N⁺, N₂⁺, and Ar⁺. The energy and ion mass covered wider ranges than previous measurements for the CCD. The results of these measurements show, as in the case of the previous measurement, for a given ion the CCD response increases with energy and for a given particle energy the response decreases with increasing mass of the particle. These results are in agreement with predictions of the theory of the range of ions in solids. The results also show the possibility for the application of the delta doped CCD as a detector for low energy particle measurements for space plasma physics applications.     

碳当量在3.6～3.9范围内不同Si/C比对铸件组织和性能的影响

以陕西压延设备厂生产的HT2 5 0以上的铸铁件为依据 ,通过 1998年 10月份前后大量数据的类比分析方法 ,寻找不同的Si/C比与机械性能、金相组织之间的关系 ,并严格控制熔炼工艺 ,提出了一些解决的措施 ,并在实际生产中取得了明显的效果     

带非线性扰动的时滞大系统的分散输出反馈镇定

文章研究一类时滞大系统 ,内部具有非线性扰动 ,同时带有时滞关联和非线性关联项。考虑到大系统的状态部分可测量 ,设计了输出反馈控制器 ,来使系统分散输出鲁棒镇定 ,仿真的结果证明了结论的有效性。     

全罗盘画面反走样算法的研究和实现

在一定分辨率的屏幕上，用常规画线函数生成的静态罗盘刻度线呈现出锯齿形失真，特别当罗盘作动态旋转时，刻度线扭曲，变形，为消除这一现象，笔者根据线条的反走样原理，提出了一种实用算法，消除了静态的锯齿形失真，并从CRT屏幕上仿真的动态效果来看，罗盘刻度线旋转时不扭曲，不变形，实验证明，在光栅图形显示器中，应用此算法可消除走样，大大增强罗盘刻度线动态显示效果。