Inkjet printed surface cell light-emitting devices from a water-based polymer dispersion

We report on the fabrication of polymer light-emitting diodes (PLEDs) and light-emitting electrochemical cells (LECs) in planar surface cell geometry (anode as well as the cathode are made of gold; interelectrode spacing: 1 μm) by means of inkjet printing. The active material for PLEDs is an aqueous poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEH-PPV) dispersion, and for LECs blends thereof with poly(ethylene oxide) (two different molecular weights: 100,000 g/mol (PEO-100,000) and 30,000 g/mol (PEO-30,000)) and lithium-triflate, building the solid state electrolyte. The surface PLEDs reveal very poor device performance with extremely high current and light emission onset voltages. However, adding the solid state electrolyte to the luminescent material, leading to the device type of an LEC, distinctly improves the performance obtaining onset voltages slightly above 3 V and remarkable enhanced light output. Due to the exchange of the high molecular weighted PEO-100,000 by the PEO-30,000, which leads to an elimination of the undesired bead-on-a-string effect during the inkjet printing process, the reproducibility of the device fabrication can be conspicuously improved. Additionally, the location of the light emission zone of a surface LEC can be easily determined, since one has a direct view between the electrodes. For such a device the light generation occurs near the cathode.     

A neural network modeling approach to circuit optimization andstatistical design

The trend of using accurate models such as physics-based FET models, coupled with the demand for yield optimization results in a computationally challenging task. This paper presents a new approach to microwave circuit optimization and statistical design featuring neural network models at either device or circuit levels. At the device level, the neural network represents a physics-oriented FET model yet without the need to solve device physics equations repeatedly during optimization. At the circuit level, the neural network speeds up optimization by replacing repeated circuit simulations. This method is faster than direct optimization of original device and circuit models. Compared to existing polynomial or table look-up models used in analysis and optimization, the proposed approach has the capability to handle high-dimensional and highly nonlinear problems     

A modular neural network approach to microelectronic circuit yield optimization

The unpredictable variation in microelectronic circuits due to process tolerances increases significantly with increased levels of miniaturization. If ignored, the variation will result in poor manufacturing yield. If a worst-case approach is adopted, a loss of competitive edge results. This situation provides the motivation for efficient robust design of VLSI circuits, the subject of this paper. Given the need for efficiency of analysis without significant loss of accuracy, a method is proposed which generates a neural network for mapping process-level parameters to circuit performance. The approach uses a modular neural network—an adaptive mixture of local experts competing to learn different aspects of a problem. Once the neural network model is established and validated, it is employed in performing extremely efficient optimization of the circuit yield at minimal cost: the trained ANN acts as a cheap but accurate simulator which when supplied with a set of inputs which characterize transistors at the process and device level calculates the circuit performance with 97% accuracy at 1% of the cost of a full SPICE simulation. Even when the cost of ANN training is factored in, average cost savings of 80% are achieved during yield optimization. The neural net approach offers significant advantages including vastly reduced computational cost with little loss of accuracy and complete generality of application.     

A new approach to efficiency enhancement of polymer light-emitting diodes by deposition of anode buffer layers in the presence of additives

Here we report a new approach for the preparation of anode buffer layer for efficient polymer light-emitting devices (PLEDs) by using glycerol to modify relative low conductivity poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS). This new type of anode buffer layer allows for a 50–90% increase in device performance for green emitting phosphorescent PLEDs in terms of luminous efficiency, and external quantum efficiency, while 90–150% in power efficiency, as compared to devices fabricated using commercially available PEDOT:PSS. The green emitting phosphorescent PLEDs with this modified anode buffer exhibit very high efficiencies, representing a significant step forward to matching and exceeding the efficiencies reported to date with vacuum-deposited small molecular devices. We anticipate that these findings can provide a simple experimental procedure for improvement of PLEDs.     

CSM neural network for degraded printed character optical recognition

In pattern recognition applications, the classification power of a system can be improved by combining several classifiers. Obviously performance of the system cannot be improved if the individual classifiers make all the same mistakes, thus it is important to use different features and different structures in the individual classifiers. In this context, we propose a two subnets neural network called CSM net. The first subnet, or similarity layer, is operating as a similarity measure neural network; it is based on the complementary similarity measure method (CSM). The second subnet is a competitive neural network (CNN) based on the winner takes all algorithm (WTA) that is used for the classification. In the proposed neural architecture, the statistical CSM method is analyzed, and implemented in the form of a feed forward neural network, it is named “similarity measure neural network” (SMNN). We show that the resulting SMNN synaptic weights are modified versions of the model patterns used in the training set, and that they can be considered as a memory network. We introduce a relative distance data calculated from the SMNN output, and we use it as a quality measurement tool of the degraded characters, what makes the SMNN classifier very powerful, and very well-suited for features rejections. This relative distance is used by the SMNN and compared to a first rejection threshold to accept, or reject, the incoming characters. In order to guarantee a higher recognition and reliability rates for the cascaded method, the SMNN is combined with a second subnet based on the WTA for classification using a second specific rejection threshold. These two submits combination (CSM net) boost the performance of the SMNN classifier. This is resulting in a robust multiple classifiers that can be used for setting the entire rejection threshold. The experimental results that we introduce are related to the proposed method, but the tests are introduced with various impulse noise levels, as well as the tests with broken and manually corrupted characters, and characters with various levels of additive Gaussian noise. The experiments show the effective ability of the model to yield relevant and robust recognition on poor quality printed checks, and show that the CSM net outperforms the previous works, both in efficiency and accuracy.     

神经网络结构优化设计方法

神经网络的结构设计、优化方法是神经网络研究的一个方向,至今仍没一个系统有效的解决方法。主要列举了多种神经网络结构的设计方法,包括凑试方法、增长方法、修剪方法、进化方法、自适应方法和正交多项式误差逐次逼近法,并对这些神经网络优化设计的方法进行了评述和比较。     

A novel neural network combined with FDTD for the synthesis of a printed dipole antenna

A novel synthesis artificial neural network (SYNTHESIS-ANN) is combined with the finite-difference time-domain method. Practical applications are illustrated through the optimization of a dipole antenna input impedance. The ANN architecture utilizes a hetero-associative memory, which exploits a fault tolerant number representation of a neural network for input and output data. In addition, the number representation reveals significant insight into a new method of fault tolerant computing. A new randomization process for the synthesis of antenna geometrical parameters is presented. Additional work is required to investigate the potential of this new paradigm.     

一种基于高维粒子群算法的神经网络结构优化研究

为克服传统BP神经网络在运算过程的不足,提出一种基于高维粒子群算法的神经网络优化方法。通过在高维PSO算法中引入随机变化的加速常数来获得最优权值,对BP神经网络进行优化和训练,再将优化好的高维BP神经网络运用到交通事件自动检测中,通过检测训练算法,并对训练后的数据进行分类测试,把分类测试的结果与传统BP神经网络和经典事件检测算法比较。结果显示,经过优化后的高维粒子群BP神经网络的检测率、算法性能均优于BP神经网络算法和经典算法,其中97,50个测试样本中仅有2个测试样本与应该达到的数值不一致,其他样本都满足测试要求,并且平均优化测试时间是传统BP神经网络检测时间的一半,因此,优化后的BP神经网络算法的性能十分优越。     

Graph neural network-based virtual network function deployment optimization

Software-defined networking (SDN) and network function virtualization (NFV) help reduce the operating expenditure (OPEX) and capital expenditure (CAPEX) as well as increase the network flexibility and agility. However, since the network is more dynamic and heterogeneous than before, operators have problems to cope with the increased complexity of managing virtual networks and machines. This complexity is paired with strict time requirements for making management decisions; traditional mechanisms that rely on, for example, integer linear programming (ILP) models are no longer feasible. Machine learning has emerged as one of the possible solution to address network management problems to get near-optimal solutions in a short time. However, applying machine learning to network management is also not simple and has many challenges. Especially, understanding the network environment is an important problem for designing a machine learning model. In this paper, we proposed to use graph neural network (GNN) for virtual network function (VNF) management. The proposed model solves the complex VNF management problem in a short time and gets near-optimal solutions. We developed a model by taking into account various network environment conditions so that it can be applied in the actual network environment. Also, through in-depth experiments, we suggested the direction of the machine learning-based network management method.     

Stability of polymer light-emitting diodes

In this paper, we report on the lifetime of polymer LEDs fabricated at Philips Research. For single-layer LEDS, we find that the operational lifetime in nitrogen gas is limited by the stability of the indium-tin-oxide (ITO) anode. By using a polymeric capping layer for the ITO, we obtain more stable devices. In air, the lifetime is limited by black spot formation. Small pinholes in the cathode layer are the origins of the black spots. Water or oxygen may diffuse through these pinholes and react with the cathode, causing degradation. By encapsulating the devices we can prevent black spot formation. Our present 8 cm² devices have lifetimes of many thousands of hours at daylight visibility under ambient conditions.     

Voltage independent white emission from all solution processed polymer light-emitting diode with dual emitting layers spaced by an alcohol soluble conjugated polymer as interlayer

Spatial control of recombination zone in multilayer white polymer light emitting diode (WPLED) is highly desirable for stable white-light emission. In this work, the utilization of 18-crown-6 (Cn6)-grafted polyfluorene (PFCn6) as an interlayer in between two emitting layers is demonstrated to control the recombination zones for the multilayer WPLED with β-phase and rubrene doped poly(9,9-di-n-octylfluorene) (PFO) as blue- and yellow-emitting layers, respectively. The device gives the maximum brightness of 15,695 cd/m² and maximum efficiency 5.43 cd/A, accompanying with voltage-independent electroluminescence spectrum having invariant Commission Internationale de L’Eclairage (CIE) coordinates of (0.32, 0.36). The performance with the luminance efficiency 5.43 cd/A and voltage independent white emission is the highest record among the reported multilayer WPLED.     

Training and optimization of an artificial neural network controlling a hybrid power filter

A hybrid power compensator (HPC) consisting of a static VAr compensator and a dynamic compensator needs to be optimally controlled during the compensation of nonlinear loads. The HPC must be controlled to meet minimum requirements in terms of power factor and harmonic distortion, while at the same time minimizing its total cost. An artificial neural network (ANN) is used to control the HPC amidst a very dynamic power system environment. The performance of a reference ANN is evaluated while controlling an HPC connected to a typical nonlinear industrial load. The training and performance of the ANN is then optimized in terms of training set size, training set packing and ANN topology and the performance compared to the reference ANN. This paper highlights the importance of optimising the mentioned ANN parameters to achieve optimum ANN training and modeling accuracy. The results obtained reveals that the application of an ANN in controlling an HPC is feasible given that the ANN parameters are chosen appropriately.     

Modeling of pneumatic artificial muscle using a hybrid artificial neural network approach

Pneumatic Artificial Muscle (PAM) actuator has been widely used in medical and rehabilitation robots, owing to its high power-to-weight ratio and inherent safety characteristics. However, the PAM exhibits highly non-linear and time variant behavior, due to compressibility of air, use of elastic-viscous material as core tube and pantographic motion of the PAM outer sheath. It is difficult to obtain a precise model using analytical modeling methods. This paper proposes a new Artificial Neural Network (ANN) based modeling approach for modeling PAM actuator. To obtain higher precision ANN model, three different approaches, namely, Back Propagation (BP) algorithm, Genetic Algorithm (GA) approach and hybrid approach combing BP algorithm with Modified Genetic Algorithm (MGA) are developed to optimize ANN parameters. Results show that the ANN model using the GA approach outperforms the BP algorithm, and the hybrid approach shows the best performance among the three approaches.     

基于遗传优化神经网络的边坡稳定性评价

由于引起滑坡的因素复杂,传统预测方法难以得到高精度的结果。文中利用遗传算法(GA)全局搜索能力强、不易陷入局部极小值的特点对样本的初始权值和阈值进行优化处理,使得前馈型神经网络(BP)在学习和预测时能够得到一个最佳的权值和阈值,从而探索出影响滑坡的因子与边坡稳定性之间潜在的关系。从仿真结果可知:优化权值后的BP神经网络得到边坡稳定性的判对率达到100%,而随机权值BP神经网络的判对率仅为54.5%,判对率提高了45.5%;安全系数较随机权值BP神经网络的平均误差提高了6.08%。因此,优化BP神经网络的预测精度得到明显提高,在今后边坡稳定性的实际应用评价中可作为一种有效的辅助手段。     

基于BP神经网络的RSSI测距优化算法

基于接收信号强度指示(received signal strength indication, RSSI)测距的研究和应用领域很广泛,一直是物联网研究的热点. 为降低传统基于反向传播(back propagation,BP)神经网络的RSSI测距误差,文中提出一种基于K-means聚类算法对样本数据进行预处理的BP神经网络测距算法,来解决由于RSSI值衰减程度不同引起的不同距离区间RSSI值和真实距离之间映射关系不均匀的问题. 将K-means聚类算法应用于BP神经网络模型中,对样本数据进行距离区间划分,然后将已经分类好的数据分别输入BP神经网络建立网络模型并进行实验仿真. 结果显示:传统基于BP神经网络的RSSI测距算法的均方根误差为1.425 7 m;而经过K-means算法改进后的BP神经网络测距算法的均方根误差为1.288 7 m,降低了测距误差,并优化了目标RSSI值与真实距离的映射关系.     

Development and analysis of a neural network approach toPisarenko's harmonic retrieval method

Pisarenko's harmonic retrieval (PHR) method is perhaps the first eigenstructure based spectral estimation technique. The basic step in this method is the computation of eigenvector corresponding to the minimum eigenvalue of the autocorrelation matrix of the underlying data. The authors recast a known constrained minimization formulation for obtaining this eigenvector into the neural network (NN) framework. Using the penalty function approach, they develop an appropriate energy function for the NN. This NN is of feedback type with the neurons having sigmoidal activation function. Analysis of the proposed approach shows that the required eigenvector is a minimizer (with a given norm) of this energy function. Further, all its minimizers are global minimizers. Bounds on the integration time step that is required to numerically solve the system of nonlinear differential equations, which define the network dynamics, have been derived. Results of computer simulations are presented to support their analysis     

Improved polymer light-emitting diodes by the tuning of charge balance

We have prepared polymer-small molecule hybrid electroluminescence devices with improved brightness and efficiency by a doping method. The doping effect on the hole transporter N,N′-diphenyl-N,N′-bis(4′-[N,N-bis(naphth-1-yl)-amino]-biphenyl-4-yl)-benzidine has been investigated in bilayer devices with structure of ITO/MEH-PPV: TPTE/PBD/Al. It was found that the effective hole mobility increases with increasing TPTE content in the blend, which is directly determined by space charge limited current at high voltage. The brightness and EL efficiencies of the doped OLEDs were increased significantly over those obtained for devices based only on MEH-PPV, which is attributed to more balanced electron and hole recombination due to improved hole mobility.