Parallel global optimization with the particle swarm algorithm

Present day engineering optimization problems often impose large computational demands, resulting in long solution times even on a modern high-end processor. To obtain enhanced computational throughput and global search capability, we detail the coarse-grained parallelization of an increasingly popular global search method, the particle swarm optimization (PSO) algorithm. Parallel PSO performance was evaluated using two categories of optimization problems possessing multiple local minima-large-scale analytical test problems with computationally cheap function evaluations and medium-scale biomechanical system identification problems with computationally expensive function evaluations. For load-balanced analytical test problems formulated using 128 design variables, speedup was close to ideal and parallel efficiency above 95% for up to 32 nodes on a Beowulf cluster. In contrast, for load-imbalanced biomechanical system identification problems with 12 design variables, speedup plateaued and parallel efficiency decreased almost linearly with increasing number of nodes. The primary factor affecting parallel performance was the synchronization requirement of the parallel algorithm, which dictated that each iteration must wait for completion of the slowest fitness evaluation. When the analytical problems were solved using a fixed number of swarm iterations, a single population of 128 particles produced a better convergence rate than did multiple independent runs performed using sub-populations (8 runs with 16 particles, 4 runs with 32 particles, or 2 runs with 64 particles). These results suggest that (1) parallel PSO exhibits excellent parallel performance under load-balanced conditions, (2) an asynchronous implementation would be valuable for real-life problems subject to load imbalance, and (3) larger population sizes should be considered when multiple processors are available.     

Mode of action of a minor xylanase from Thermoascus aurantiacus on polysaccharides and model substrates

The mode of action of a minor xylanase on a variety of polysaccharides and model substrates was investigated. The enzyme was excreted by Thermoascus aurantiacus grown in solid state fermentation (SSF). The purified enzyme had a molecular mass of 33,000. Thin layer chromatography analysis showed that the endoxylanase liberated short fragments from polysaccharides. The enzyme hydrolysed aryl-beta-D-cellobioside and the chromogenic (fluorogenic) 4-methylumbelliferyl-beta-glycosides of xylobiose (MeUmbXyl2) and xylotriose (MeUmbXyl3) at the agluconic linkage. The results suggested that the endoxylanase belonged to family 10.     

Localized electrical nerve blocking

Localized electrical nerve blocking was investigated in computer simulation and in vivo trials for sinusoidal frequencies between 5 and 20 kHz. Computer simulations indicated that a localized transmission block of the axons could occur in each of the axon models. An approximation of nerve stimulation was derived from individual axon simulations conducted over axon diameters of 5-15 microm and electrode to axon distances of 0.25 to 2.0 mm. Examination of the membrane voltage and ionic gate potentials indicated that the block could be attributed to an elevated membrane voltage. The elevated membrane voltage could prevent conduction of action potentials through the region of the sinusoidal currents. At lower amplitudes, the sinusoidal current could stimulate the axon and generate a continuous series of action potentials. In vivo trials demonstrated that the sinusoidal frequencies of greater than 10 kHz would cause a localized block in rats. Sinusoidal frequencies below 5 kHz would lead to a reduction in muscle force that appeared to be caused by depletion of transmitter at the neuromuscular junction. As indicated by the computer models of rat nerves, the endplate depletion block occurred at a lower frequency (below 5 kHz) than the block (above 10 kHz). A partial block of the axon was demonstrated, suggesting that sinusoidal currents could be used to provide selective stimulation if they are combined with distal electrical stimulation.     

适用于机舱环境的抗噪语音指令识别方案

基于飞行器语音指令控制的应用环境,提出了一种高抗噪性的非特定人连接词语音识别方案。针对机舱噪声的频谱特性,采用了谱减法、倒谱均值归一化以及噪声模型适应技术。仿真测试表明,该方案的平均识别率在信噪比为0 dB的条件下达到92%以上。     

An S1 gradient of refractoriness is not essential for reentry induction by an S2 stimulus

This communication reports a numerical simulation of reentry induction by successive stimulation (S1-S2) that does not require an S1 gradient of refractoriness. The S1 action potential is uniform in space, so before the S2 stimulus there is no refractory gradient. Nevertheless, a unipolar S2 stimulus initiates quatrefoil reentry. The result supports the growing realization that virtual electrodes, hyperpolarization, de-excitation, and break excitation may be important during reentry induction.     

基于MSP430单片机水钻磨床控制器的设计

在掌握水钻磨床基本功能的基础上,使用MSP430单片机设计了水钻磨床控制器,使水钻磨床操作更方便,运行更稳定,从而使水钻磨床的整体性能显著提高。文章重点分析和研究了两部分,一是水钻磨床控制器电路的设计与连接;二是根据水钻磨床的运行状态来编写程序。控制器在模拟系统和现场运行情况中进行了实验验证,通过现场测试证明采用MSP430单片机控制的水钻磨床在操控实时性、产品精度等方面得到了大幅度的提高,并且还提高了电气的抗干扰等性能。     

解理面预处理方法对二次外延的影响

利用超晶格解理面方法制备定位生长的InAs量子线.首先以分子束外延技术在OaAs衬底上生长GaAs/AlGaAs超晶格,然后将样品取出外延系统进行解理,对解理面进行预处理之后在(110)解理面上进行二次外延.实验结果显示超晶格解理面的预处理方法对二次外延有重大影响,其中择优腐蚀比自然氧化更有利于量子线的定位生长,过高温度的脱氧除气会导致解理面的GaAs部分出现坑状结构,表明(110)面上的Ga原子容易脱附.同时,Ga原子在(110)面上的迁移长度比较大,原子的择优扩散方向为[001]方向.     

采用SiPM探测三硝基甲苯的时间分辨拉曼光谱

硅光电倍增器（SiPM）是近年来迅速发展起来的新型固态弱光探测器,有望取代传统的光电倍增器（PMT）用于拉曼检测中。为了解决拉曼检测中的荧光干扰以及SiPM的高暗计数率的缺陷,建立了基于SiPM为光探测器的时间分辨拉曼光谱测量系统。以三硝基甲苯（TNT）为样品,重点研究了其拉曼峰的峰本比（PBR）随选通时间的变化规律。结果表明:随着选通时间增大,拉曼峰的PBR呈现先增大后减小的趋势,最后变化缓慢。当选通时间为400 ps时,该方法所测得拉曼峰的PBR要优于商业拉曼谱仪和参考文献[12]采用的门控方法的结果,并且此时系统所记录的SiPM暗计数水平与PMT相当。该方法在很大程度上实现了对样品荧光和SiPM高暗计数率的抑制,显著提高了拉曼谱的PBR。     

Fuzzy logic control of FES rowing exercise in paraplegia

An indoor personal rowing machine (Concept 2 Inc., Morrisville, VT) has been modified for functional electrical stimulation assisted rowing exercise in paraplegia. To successfully perform the rowing maneuver, the voluntarily controlled upper body movements must be coordinated with the movements of the electrically stimulated paralyzed legs. To achieve such coordination, an automatic controller was developed that employs two levels of hierarchy. A high level finite state controller identifies the state or phase of the rowing motion and activates a low-level state-dedicated fuzzy logic controller (FLC) to deliver the electrical stimulation to the paralyzed leg muscles. A pilot study with participation of two paraplegic volunteers showed that FLC spent less muscle energy, and produced smoother rowing maneuvers than the existing On-Off constant-level stimulation controller.