Pluggable multitask diffractive neural networks based on cascaded metasurfaces

Optical neural networks have significant advantages in terms of power consumption,parallelism,and high computing speed,which has intrigued extensive attention in both academic and engineering communities.It has been considered as one of the powerful tools in promoting the fields of imaging processing and object recognition.However,the existing optical system architecture cannot be reconstructed to the realization of multi-functional artificial intelligence systems simultaneously.To push the development of this issue,we propose the pluggable diffractive neural networks(P-DNN),a general paradigm resorting to the cascaded metasurfaces,which can be applied to recognize various tasks by switching internal plug-ins.As the proof-of-principle,the recognition functions of six types of handwritten digits and six types of fashions are numerical simulated and experimental demonstrated at near-infrared regimes.Encouragingly,the proposed paradigm not only improves the flexibility of the optical neural networks but paves the new route for achieving high-speed,low-power and versatile artificial intelligence systems.

Multi-dimensional multiplexing optical secret sharing framework with cascaded liquid crystal holograms

Secret sharing is a promising technology for information encryption by splitting the secret information into different shares.However,the traditional scheme suffers from information leakage in decryption process since the amount of available information channels is limited.Herein,we propose and demonstrate an optical secret sharing framework based on the multi-dimensional multiplexing liquid crystal(LC)holograms.The LC holograms are used as spatially separated shares to carry secret images.The polarization of the incident light and the distance between different shares are served as secret keys,which can significantly improve the information security and capacity.Besides,the decryption condition is also restricted by the applied external voltage due to the variant diffraction efficiency,which further increases the information security.In implementation,an artificial neural network(ANN)model is developed to carefully design the phase distribution of each LC hologram.With the advantage of high security,high capacity and simple configuration,our optical secret sharing framework has great potentials in optical encryption and dynamic holographic display.

Modular System of Cascaded Converters Based onModel Predictive Control

A modular system of cascaded converters based on model predictive control(MPC)is proposed to meet the application requirements ofmultiple voltage levels and electrical isolation in renewable energy generation systems.The system consists of a Buck/Boost+CLLLC cascaded converter as a submodule,which is combined in series and parallel on the input and output sides to achieve direct-current(DC)voltage transformation,bidirectional energy flow,and electrical isolation.The CLLLC converter operates in DC transformer mode in the submodule,while the Buck/Boost converter participates in voltage regulation.This article establishes a suitable mathematical model for the proposed system topology,and uses MPC to control the system based on this mathematical model.Module parameters are designed and calculated,and simulation is built in MATLAB/Simulink to complete the simulation comparison experiment between MPC and traditional proportional integral(PI)control.Finally,a physical experimental platform is built to complete the physical comparison experiment.The simulation and physical experimental results prove that the control accuracy and response speed ofMPC are better than traditional PI control strategy.

Coordinated Capacitor Voltage Balancing Method for Cascaded H-bridge Inverter with Supercapacitor and DC-DC Stage

Cascaded H-bridge inverter(CHBI) with supercapacitors(SCs) and dc-dc stage shows significant promise for medium to high voltage energy storage applications. This paper investigates the voltage balance of capacitors within the CHBI, including both the dc-link capacitors and SCs. Balance control over the dc-link capacitor voltages is realized by the dcdc stage in each submodule(SM), while a hybrid modulation strategy(HMS) is implemented in the H-bridge to balance the SC voltages among the SMs. Meanwhile, the dc-link voltage fluctuations are analyzed under the HMS. A virtual voltage variable is introduced to coordinate the balancing of dc-link capacitor voltages and SC voltages. Compared to the balancing method that solely considers the SC voltages, the presented method reduces the dc-link voltage fluctuations without affecting the voltage balance of SCs. Finally, both simulation and experimental results verify the effectiveness of the presented method.

Model Predictive Control for Cascaded H-Bridge PV Inverter with Capacitor Voltage Balance

We designed an improved direct-current capacitor voltage balancing control model predictive control(MPC)for single-phase cascaded H-bridge multilevel photovoltaic(PV)inverters.Compared with conventional voltage balanc-ing control methods,the method proposed could make the PV strings of each submodule operate at their maximum power point by independent capacitor voltage control.Besides,the predicted and reference value of the grid-connected current was obtained according to the maximum power output of the maximum power point tracking.A cost function was con-structed to achieve the high-precision grid-connected control of the CHB inverter.Finally,the effectiveness of the proposed control method was verified through a semi-physical simulation platform with three submodules.

一种新的非线性光学现象——Cascaded二级非线性光学效应

从Maxwell方程组出发 ,研究了一种新的非线性光学效应———Cascaded二级非线性现象 ,推导了依赖于光强的折射率变化———类克尔效应 ,分析了Cascaded非线性系数依赖于基波波长及波矢失配的变化规律

基于改进的Cascade RCNN铸管字符检测算法

由于工业现场采集的铸管字符图像存在背景模糊、字符区域占比小、刻字位置不固定、油漆遮挡等问题,导致现有模型的检测精度难以满足工业现场的需求。针对上述问题,提出改进的Cascade RCNN铸管字符检测算法。首先对特征金字塔进行改进,提出融合小目标增强的特征金字塔(STE-FPN),利用多尺度特征融合的特征增强能力丰富铸管小目标字符的特征信息。其次引入自矫正/池化的ResNeSt(SCP-ResNeSt)作为特征提取网络,利用自矫正卷积和池化操作以提升背景复杂的铸管字符特征提取效率。最后对级联结构进行改进,引进Mask分支结构,可以自适应地检测字符区域并去除干扰区域,优化了检测结果。将改进后的算法在铸管数据集上进行测试,其平均检测精度mAP为99.1%,比原Cascade RCNN算法提高了2.3%,得到的精度表明改进后的性能优于原算法。

Three-wavelength generation from cascaded wavelength conversion in monolithic periodically poled lithium niobate

Tunable coherent emission is generated in a single-pass, cascaded wavelength conversion process from mode-locked laser-pumped monolithic periodically poled lithium niobate（PPLN）. Three ranges of wavelength, including visible output from 628 nm to 639 nm, near-infrared output from 797 nm to 816 nm, and mid-infrared output from 3167 nm to 3459 nm,were obtained from the monolithic PPLN, which consists of a 10-mm section for 532-nm-pumped optical parametric generation（OPG） and a 7-mm section for 1064-nm-pumped sum frequency generation（SFG）. A pump-to-signal conversion efficiency of 23.4% for OPG at 50°C and a quantum efficiency of 26.2% for SFG at 200°C were obtained.

Characteristics of Single Cathode Cascaded Bias Voltage Arc Plasma

A single cathode with a cascaded bias voltage arc plasma source has been developed with a new quartz cathode chamber,instead of the previous copper chambers,to provide better diagnostic observation and access to the plasma optical emission.The cathode chamber cooling scheme is also modified to be naturally cooled only by light emission without cooling water to improve the optical thin performance in the optical path.A single-parameter physical model has been developed to describe the power dissipated in the cascaded bias voltage arc discharge argon plasmas,which have been investigated by utilizing optical emission spectroscopy（OES） and Langmuir probe.In the experiments,discharge currents from 50 A to 100 A,argon flow rates from 800 sccm to 2000 sccm and magnetic fields of 0.1 T and 0.2 T were chosen.The results show：（a） the relationship between the averaged resistivity and the averaged current density exhibits an empirical scaling law as η∝ j^（-0.63369） and the power dissipated in the arc has a strong relation with the filling factor;（b） through the quartz,the argon ions optical emission lines have been easily observed and are dominating with wavelengths between 340 nm and 520 nm,which are the emissions of Ar^＋-434.81 nm and Ar^＋-442.60 nm line,and theintensities are increasing with the arc current and decreasing with the inlet argon flow rate;and（c） the electron density and temperature can reach 2.0 × 10^19 m^-3 and 0.48 eV,respectively,under the conditions of an arc current of 90 A and a magnetic field of 0.2 T.The half-width of the ne radial profile is approximatively equal to a few Larmor radii of electrons and can be regarded as the diameter of the plasma jet in the experiments.

A Novel Cascaded PID Controller for Automatic Generation Control Analysis With Renewable Sources

Present day power scenarios demand a high quality uninterrupted power supply and needs environmental issues to be addressed. Both concerns can be dealt with by the introduction of the renewable sources to the existing power system. Thus, automatic generation control(AGC) with diverse renewable sources and a modified-cascaded controller are presented in the paper.Also, a new hybrid scheme of the improved teaching learning based optimization-differential evolution(hITLBO-DE) algorithm is applied for providing optimization of controller parameters. A study of the system with a technique such as TLBO applied to a proportional integral derivative(PID), integral double derivative(IDD) and PIDD is compared to hITLBO-DE tuned cascaded controller with dynamic load change.The suggested methodology has been extensively applied to a 2-area system with a diverse source power system with various operation time non-linearities such as dead-band of, generation rate constraint and reheat thermal units. The multi-area system with reheat thermal plants, hydel plants and a unit of a wind-diesel combination is tested with the cascaded controller scheme with a different controller setting for each area. The variation of the load is taken within 1% to 5% of the connected load and robustness analysis is shown by modifying essential factors simultaneously by± 30%. Finally, the proposed scheme of controller and optimization technique is also tested with a 5-equal area thermal system with non-linearities. The simulation results demonstrate the superiority of the proposed controller and algorithm under a dynamically changing load.

Measurement of electron density and electron temperature of a cascaded arc plasma using laser Thomson scattering compared to an optical emission spectroscopic approach

As advanced linear plasma sources, cascaded arc plasma devices have been used to generate steady plasma with high electron density, high particle flux and low electron temperature. To measure electron density and electron temperature of the plasma device accurately, a laser Thomson scattering（LTS） system, which is generally recognized as the most precise plasma diagnostic method, has been established in our lab in Dalian University of Technology. The electron density has been measured successfully in the region of 4.5？×10^19m^-3 to7.1？×10^20m^-3 and electron temperature in the region of 0.18 eV to 0.58 eV. For comparison,an optical emission spectroscopy（OES） system was established as well. The results showed that the electron excitation temperature（configuration temperature） measured by OES is significantly higher than the electron temperature（kinetic electron temperature） measured by LTS by up to 40% in the given discharge conditions. The results indicate that the cascaded arc plasma is recombining plasma and it is not in local thermodynamic equilibrium（LTE）. This leads to significant error using OES when characterizing the electron temperature in a non-LTE plasma.

基于改进Cascade R-CNN算法的船舶目标检测方法

为了解决实际场景下船舶目标检测精度低的问题,本文在Cascade R-CNN算法的基础上,提出一种船舶目标检测方法 Boat R-CNN。Boat R-CNN使用带自注意力机制的Swin-Transformer Tiny网络提取图像特征,使用Soft-NMS非极大值抑制方法提升候选框过滤精度,使用Smooth_L1损失函数加速模型收敛并减少梯度爆炸情况,使用CIOU边界框回归损失提高候选框回归质量,并针对船舶目标的形状特征优化锚框的长宽比,提高锚框的生成质量。实验结果表明,Boat R-CNN算法的精度相比原版Cascade R-CNN算法提高了21.8%,相比主流Faster R-CNN算法提高了30.3%,有效提升了实际场景下的船舶目标检测精度。

A distributed VSG control method for a battery energy storage system with a cascaded H-bridge in a grid-connected mode

With the high penetration of renewable energy,new challenges,such as power fluctuation suppression and inertial support capability,have arisen in the power sector.Battery energy storage systems play an essential role in renewable energy integration.In this paper,a distributed virtual synchronous generator(VSG)control method for a battery energy storage system(BESS)with a cascaded H-bridge converter in a grid-connected mode is proposed.The VSG is developed without communication dependence,and state-of-charge(SOC)balancing control is achieved using the distributed average algorithm.Owing to the low varying speed of SOC,the bandwidth of the distributed communication networks is extremely slow,which decreases the cost.Therefore,the proposed method can simultaneously provide inertial support and accurate SOC balancing.The stability is also proved using root locus analysis.Finally,simulations under different conditions are carried out to verify the effectiveness of the proposed method.

基于改进Cascade R-CNN的安全帽检测算法

针对安全帽检测中,目标形状、尺度变化大,易出现漏检、误检等问题,提出了一种基于改进级联基于区域的卷积神经网络(Cascade R-CNN)的安全帽检测算法。首先,对ResNet50进行改进形成D-ResNet50,利用可变形卷积仅增加少量参数就可增大感受野的特性,对特征提取网络的C2~C5卷积层进行重塑,提高网络对目标几何变换的适应能力和特征提取能力。其次,将D-ResNet50作为主干网络引入Cascade R-CNN,形成级联目标检测器,在每个阶段对正负样本重采样,抑制误检问题。再次,对递归特征金字塔进行改进,更高效地进行多尺度特征融合,并且基于反馈信息对特征进行二次处理,增强特征表达,提高网络的分类和定位能力。最后,使用Soft-非极大值抑制(Soft-NMS)进行后处理,进一步解决漏检问题。提出的方法在Hard hat workers数据集上的AP值相比检测基线提高了3.5%,与Sparse R-CNN、TridentNet、VFnet等先进算法相比分别提升了4.7%、5.9%、2.3%等。

Ultra-broadband absorber based on cascaded nanodisk arrays

An ultra-broadband perfect absorber consisting of cascaded nanodisk arrays is demonstrated by placing insulatormetal-insulator-metal nanodisks on insulator-metal film stacks.The absorber shows over 90% absorption in a wavelength range between 600 nm and 4000 nm under transverse magnetic(TM) polarization,with an average absorptivity of 91.5%and a relative absorption bandwidth of 147.8%.The analysis of the electric field and magnetic field show that the synergy of localized surface plasmons,propagating surface plasmons,and plasmonic resonant cavity modes leads to the ultrabroadband perfect absorption,which accords well with the results of impedance-matched analysis.The influences of structural parameters and different metal materials on absorption performance are discussed.Furthermore,the absorber is polarization-independent,and the absorption remains more than 90% at a wide incident angle up to 400 under TE polarization and TM polarization.The designed ultra-broadband absorber has promising prospects in photoelectric detection and imaging.

Adaptive Cascaded High-Resolution Source Localization Based on Collaboration of Multi-UAVs

Effective information fusion is very important in hybrid source localization. In this paper, the performance analysis of conventional joint direction of arrival(DOA) and time difference of arrival(TDOA) system is derived and it is shown that this hybrid system may inferior to the single system when the ratio of angular measurements error to distance measurements error exceeds a threshold. To avoid this problem, an effective DOA/TDOA adaptive cascaded(DTAC) technique is presented. The rotation feature of UAVs and spatial filtering technique are applied to gain the signal-to-noise ratio(SNR), which leads to more accurate estimation of time delay by using DOAs. Nevertheless, the time delay estimation precision is still limited by the sampling frequency, which is constrained by the finite load of UAV. To break through the limitation, an enhanced self-delay-compensation(SDC) method is proposed, which aims at detecting the overlooked time delay within the sampling interval by adding a tiny time delay. Finally, the position of the source is estimated by the Chan algorithm. Compared to DOA-only algorithm, TDOA-only algorithm and joint DOA/TDOA(JDT) algorithm, the proposed method shows better localization accuracy regardless of different SNRs and sampling frequencies. Numerical simulations are presented to validate the effectiveness and robustness of the proposed algorithm.

Cascaded dual-channel fiber SPR temperature sensor based on liquid and solid encapsulations

In order to control the working wavelength range of the fiber surface plasmon resonance(SPR)temperature sensor and realize the wavelength division multiplexing type multi-channel fiber SPR temperature sensor,by comprehensively investigating the influence of liquids with different thermal-optical coefficients and solid packaging materials on the performance of fiber SPR temperature sensor,a dual-channel fiber SPR temperature sensor based on liquid-solid cascade encapsulation was designed and fabricated.The liquid temperature sensing stage encapsulated in capillary worked in 616.03 nm-639.05 nm band,the solid sensing stage coated with pouring sealant worked in 719.37 nm-825.27 nm band,and the two stages were cascaded to form a fiber dual-channel temperature sensor.The testing results indicated that when the temperature range was 35℃-95℃,the sensitivity of two-stage temperature detection was−0.384 nm/℃and−1.765 nm/℃respectively.The proposed fiber sensor has simple fabrication and excellent performance which can be widely used in various fields of dual-channel temperature measurement and temperature compensation.

Optical encryption with cascaded fractional wavelet transforms

On the basis of fractional wavelet transform, we propose a new method called cascaded fractional wavelet transform to encrypt images. It has the virtues of fractional Fourier transform and wavelet transform. Fractional orders, standard focal lengths and scaling factors are its keys. Multistage fractional Fourier transforms can add the keys easily and strengthen information se-curity. This method can also realize partial encryption just as wavelet transform and fractional wavelet transform. Optical reali-zation of encryption and decryption is proposed. Computer simulations confirmed its possibility.

An on-line free handwritten Chinese character recognition method based on component cascaded HMMs

This paper presents a cascaded Hidden Markov Model （HMM）, which allows state＇s transition, skip and duration. The cascaded HMM extends the way of HMM pattern description of Handwritten Chinese Character （HCC） and depicts the behavior of handwritten curve more reliably in terms of the statistic probability. Hence character segmentation and labeling are unnecessary. Viterbi algorithm is integrated in the cascaded HMM after the whole sample sequence of a HCC is input. More than 26,000 component samples are used tor training 407 handwritten component HMMs. At the improved training stage 94 models of 94 Chinese characters are gained by 32,000 samples, Compared with the Segment HMMs approach, the recognition rate of this model tier the tirst candidate is 87.89% and the error rate could be reduced by 12.4%.

Design and offline processing of an ultrafast digitizer based on internal cascaded DRS4

In this paper, we present an ultrafast digitizer utilizing the DRS4 switched capacitor array applicationspecific integrated circuit to achieve an ultrafast sampling speed of at most 5 GS/s. We cascaded all eight channels(sub-channels) of a single DRS4 chip for increased storage depth. The digitizer contains four DRS4 chips, a quadchannel analog-to-digital converter,a controlling fieldprogrammable gate array, a PXI interface, and an SFP+connector. Consequently, each DRS4 channel has a depth of 8192 points and a vertical resolution of 14 bits. The readout sequences should be broken into several segments and then reordered to obtain the correct sequential data sets, and this offline procedure varies in different readout modes. This paper describes the design and implementation of the hardware;in particular, the respective processing procedures are described in detail. Furthermore, the offset error is calibrated and corrected to improve the precision of the captured waveform in both single-channel and highresolution modes.