压缩感知在合成发射孔径医学超声成像中的应用

为了解决合成发射孔径技术在医学超声成像实现中面临的数据量大及接收通道多的问题,提出一种超声成像系统频率域稀疏性模型的压缩感知成像算法。首先对超声系统频率域稀疏性模型进行了验证;然后根据稀疏性模型利用压缩感知理论对回波信号进行压缩采样,并使用最优化方法完成回波信号重建;最终通过合成发射孔径技术完成超声成像。针对医学成像中常用的点目标及模拟胎儿目标进行成像仿真实验,对重建图像在均方误差、分辨率及成像质量等方面与常规成像结果对比分析。实验结果表明在保证成像质量的同时,仅使用30%原始数据量及50%总接收通道数目可完成成像;频率域稀疏性模型的压缩感知成像算法可以大幅度减少合成发射孔径成像所需数据量及接收通道数,极大地降低了系统复杂度。      

Quantum metrology

The statistical error is ineluctable in any measurement. Quantum techniques, especially with the development of quantum information, can help us squeeze the statistical error and enhance the precision of measurement. In a quantum system, there are some quantum parameters, such as the quantum state, quantum operator, and quantum dimension, which have no classical counterparts. So quantum metrology deals with not only the traditional parameters, but also the quantum parameters. Quantum metrology includes two important parts： measuring the physical parameters with a precision beating the classical physics limit and measuring the quantum parameters precisely. In this review, we will introduce how quantum characters （e.g., squeezed state and quantum entanglement） yield a higher precision, what the research areas are scientists most interesting in, and what the development status of quantum metrology and its perspectives are.     

Auditory filter based broadband MUSIC algorithm for sound source localization

Based on the analysis of the shortcomings of broadband MUSIC algorithm with short-time Fourier transform （SF-MUSIC） for sound source localization, a broadband MUSIC algorithm with auditory filter （AF-MUSIC） was proposed. The proposed algorithm first em- ploys auditory filter bank to decompose the signals received on the microphone array, and then locates the sound source with MUSIC algorithm over every frequency channel. At last, by combining with the subinterval frequency estimation, the final localization result is gained. Evaluations on the proposed algorithm prove that comparing with the SF-MUSIC algorithm, the AF-MUSIC algorithm decreases the average error of the estimation results with 2.5479 de- gree in different source conditions. The accuracy of sound source DOA estimation is enhanced effectively.     

大气湍流三维波前探测模式层析算法分析

大气湍流三维波前探测是实现多层共轭自适应光学技术的关键和前提. 对湍流三维波前探测中最常用的模式法层析技术进行理论研究与分析, 提出该算法存在原理性限制, 并基于此对模式法层析技术产生误差的原因展开分析, 最后针对不同类型的模式层析重构误差给出数值仿真实验结果.分析表明, 模式层析重构中使用了Zernike分解基的一部分作为新的分解基进行波面拟合, 从而引入模式混淆和模式耦合两个方面的误差; 部分Zernike分解基不相关是避免模式混淆误差的必要条件, 模式耦合误差则无法避免. 最后结合仿真结果提出大视场探测、小区域重构的方法, 很好地抑制了模式耦合误差. 关键词： 三维波前探测 大气层析 模式法 误差分析     

Theoretical analysis of the performance of a single side band multicarrier source based on a re-circulating frequency shifter influenced by an optical bandpass filter

An optical bandpass filter （OBF） is an important element whose properties considerably influence the output performance of a single side band multicarrier source based on a re-circulating frequency shifter. The influence of the OBF in the loop on the output spectrum is theoretically analyzed. Numerical simulations and experiments are Mso carried out. Results show that the steepness and deepness of a nonflat-top filter influence the stability of the spectrum of an output multicarrier. To obtain multicarrier output with tone to noise ratio （TNR） 〉26 dB and error vector magnitude （EVM） 〈0.25, the steepness ratio of the filter should be greater than 0.75 and deepness should be larger than 0.99.     

Method to accurate calibrate deposition rates of EUV multilayer coatings

Random thickness error is an important factor which effects the calibration accuracies of deposition rates for extreme ultraviolet （EUV） multilayer coatings fabricated by sputter deposition techniques. A least square fitting method is proposed to determine deposition rates and extract random thickness errors accurately. The validity of this method is shown by evaluating two deposition systems with control abilities of -0.1 nm and better than 0.01 nm respectively.     

稳健估计下的声源定位最小二乘算法

采用声波技术测量落点坐标时,由于多种原因不可避免会产生粗差,严重影响定位结果。针对上述问题,提出了一种粗差探测方法和稳健估计下的声源定位最小二乘算法。该算法按照逐次迭代平差的方法,不断地改变观测值的权,使粗差观测值的权趋于零,从而降低粗差的影响。通过算例对文章算法进行了验证,结果表明:新算法的解算精度明显优于传统最小二乘算法。     

基于屋顶二值红外传感器网络的人体定位和行为识别系统设计

为了解决个人隐私的保护、受光照条件和障碍物等因素的影响,设计了一种基于屋顶二值红外传感器网络的人体行为识别系统。系统采用STM32芯片和ZigBee协调器建立ZigBee网络。当实验者在安装在屋顶的20个互连的AMN31111红外传感器下面做出一系列行为时,传感器对其进行二值数据采集,ZigBee协调器将数据通过串口发送到PC机上实时动态显示,通过Keil 5系统软件对采集的二值数据以txt文本格式存储。提出了像素值法对人体进行定位和BP神经网络算法在模拟的居家环境中对7种不同人物行为进行识别。实验结果表明：该系统实现了人体多种行为的识别,其识别率为84.7%,4名实验者得到平均识别精度相比固定在居家电器传感器设备要高4.7%左右,并且该系统采集精度高、性能稳定、可靠性高、成本低、功耗低,解决了一些目前人体行为识别监测系统存在的问题。     

基于半偏法探究测量中的电阻条件

半偏法是中学里直接利用待测电表及其他的电学原件组合以达到测量其阻值的方法. 而关于半偏法 实验中,各电学器件电阻间应满足怎样的大小关系,方能减小实验误差,以达到提高实验的准确性,是一个值得深 思的问题. 立足于常见电流表及电压表内阻的实验测量,通过半偏法实验误差分析,结合高考实例,定量探析其测 量原理背后各电阻需要满足的定量条件,以便读者在后期实验方案中能及时作出评估或改良     

模拟退火遗传算法在多声源定位中的应用

利用声源定位技术测量弹丸落点坐标时,针对多发炮弹短间隔连续射击试验中,传感器接收信号的时序关系复杂,严重影响声源定位结果准确性的问题,提出了采用模拟退火遗传算法实现短间隔连续多声源目标的精确定位。试验测试了发射间隔为0.15 s的7发炮弹在600 m×600 m区域内的定位情况,结果表明:文章的算法能快速有效的对短间隔连续多声源目标进行定位解算,定位精度小于5 m。