Chaos synchronization and transmission of information in coupled SC-CNN-based canonical Chua’s circuit

In this paper, an experimental implementation of State Controlled Cellular Neural Network (SC-CNN)-based canonical Chua’s circuit is presented. Two such electronic circuits are realized. One of the circuits is configured as driver or transmitter, and other circuit is configured as response or receiver system. Synchronization of chaos between the driver and the driven system is demonstrated. The conditions under which complete synchronization is ensured are derived by applying a strategy based on the Master Stability Function and Lyapunov function. We show that transmitter and receiver systems consisting of cascaded SC-CNN-based canonical Chua’s oscillator can synchronize when they are coupled using direct information from the transmitter into the receiver. The communication performances of different information signals such as sinusoidal, square, and speech signals are demonstrated by real-time hardware experiments. The experimentally observed results are in agreement with the numerically obtained results, and they closely match with each other.     

陀螺仪在线性能及故障AR模型和神经网络方法的研究

本着眼于惯导系统中在线运行的陀螺仪性能和故障诊断的研究。针对无故障和两种故障下采集到的陀螺仪前置放大器在线输出信号,分别采用AR模型法和神经网络法进行分析,并绘出相应的模型功率谱图。通过比较和分析之后,得出信号模型功率谱图可作为判断陀螺仪性能和故障的一种依据的结论。     

Design of neural network-based control systems for active steering system

Nowadays, safety of road vehicles is an important issue due to the increasing road vehicle accidents. Passive safety system of the passenger vehicle is to minimize the damage to the driver and passenger of a road vehicle during an accident. Whereas an active steering system is to improve the response of the vehicle to the driver inputs even in adverse situations and thus avoid accidents. This paper presents a neural network-based robust control system design for the active steering system. Primarily, double-pinion steering system used modeling of the active steering system. Then four control structures are used to control prescribed random trajectories of the active steering system. These control structures are as classical PID Controller, Model-Based Neural Network Controller, Neural Network Predictive Controller and Robust Neural Network Predictive Control System. The results of the simulation showed that the proposed neural network-based robust control system had superior performance in adapting to large random disturbances.     

基于物理信息的神经网络求解曲面上对流扩散方程

引入基于物理信息的神经网络PINNs(Physics-informed Neural Networks)并将其用于求解曲面对流扩散方程。区别于传统的神经网络模型,PINNs在建立模型过程中引入了自动微分技术,并将物理信息即偏微分方程信息编译其中,通过定义损失函数得到关于该模型中神经网络参数即权重和偏置的优化目标,随后利用已有的优化算法进行求解。显而易见,PINNs通过添加额外的物理信息约束放宽了对于数据量的要求,对于一个确定性模型显示出更好的鲁棒性。本文基于曲面微分算子与欧氏空间下标准微分算子的解析关系,引入两种曲面微分算子处理技术,即非本征技术和嵌入技术,并结合PINNs针对定义在高维复杂曲面上的对流扩散方程进行求解,多个数值算例证明了该方法的有效性、鲁棒性以及其在求解此类问题的潜力。     

卷积神经网络在装备磨损颗粒识别中的研究综述

铁谱法是用于装备故障诊断的1种重要方法,其中铁谱法的重点是铁谱图像的分析,即磨损磨粒分析. 卷积神经网络是当下最流行的深度学习算法之一,其广泛应用于图像识别领域,使得图像识别领域得到突破. 随着卷积神经网络的快速发展,磨损颗粒在智能识别方面的技术取得了重大的突破. 本文中首先简述了卷积神经网络与磨粒智能识别的发展历史,针对基于卷积神经网络的磨粒识别方法进行了从图像数据集处理到模型优化技术方面的介绍,并详细说明了这些技术在磨粒识别中的具体应用实例. 然后从现有网络和自设计网络两方面分类,整理了近年来卷积神经网络应用于磨粒智能识别的代表性文献,综述了这些工作所提出的模型结构和特点,分析并阐述了各个模型主要的识别原理,各个网络结构存在的优缺点,以及它们的数据采用情况等,并对未来磨粒智能识别的主要研究方向进行了展望. 最后肯定了卷积神经网络方法在磨粒智能识别方面的重要性,同时指出了基于此方法的磨粒识别模型的缺点,并提出了应紧跟图像识别领域的最新技术以促进磨粒智能识别水平提高等建议,对磨粒智能识别的发展具有一定的意义.      

材料非线性粘弹性本构关系的神经网络模拟

对Green-Rivlin非线性粘弹本构方程给出了用一个三层时延前馈神经网络表示的等价关系,从而利用该网络可以求解G-R方程中的各阶核函数,并给出了根据有机玻璃的一组实验数据计算核函数的应用实例.由此指出神经网络方法可广泛应用于材料本构关系的研究.     

弹塑性接触问题神经网络计算模型

利用神经网络优化计算原理,引入LPNN(Lagrange Programming Neural Network)模型求解弹塑性摩擦接触问题,提高了神经网络计算精度。通过采用非线性函数作为神经元的传递函数,使神经网络的非线性问题同力学的非线性问题得到统一。最后对两个简单的弹塑性接触问题进行了数值仿真,验证了算法可行有效。     

基于神经网络的界面裂纹尖端场分析

通过构造反向传播神经网络,对裂纹尖端的应力场进行模拟,进而实现对裂纹尖端应力场甬数的逼近。得到的网络具有较高的联想、记忆能力和相当的稳定性,并且可以快速、准确地得到带裂纹构件的裂纹尖端应力场,从而确定裂纹尖端的塑性区和分析裂纹的扩展。数值计算给出了LY12-CZ材料裂纹扩展方向的计算结果,与实验结果吻合较好,还给出了两相材料含界面裂纹在复合型载荷作用下的塑性区形状的变化情况,并对两相材料含界面裂纹在复合型载荷作用下裂纹的扩展方向进行了预测。     

The application of Shuffled Frog Leaping Algorithm to Wavelet Neural Networks for acoustic emission source location

When using acoustic emission to locate the friction fault source of rotating machinery, the effects of strong noise and waveform distortion make accurate locating difficult. Applying neural network for acoustic emission source location could be helpful. In the BP Wavelet Neural Network, BP is a local search algorithm, which falls into local minimum easily. The probability of successful search is low. We used Shuffled Frog Leaping Algorithm (SFLA) to optimize the parameters of the Wavelet Neural Network, and the optimized Wavelet Neural Network to locate the source. After having performed the experiments of friction acoustic emission's source location on the rotor friction test machine, the results show that the calculation of SFLA is simple and effective, and that locating is accurate with proper structure of the network and input parameters.     

自适应联邦滤波器及其在组合导航系统中的应用

针对现有的信息分配方法在滤波精度和自适应性方面存在不足的问题,设计了一种在模糊神经网络基础上改进的自适应联邦滤波器,利用模糊神经网络自适应地调整信息分配系数,从而有效抑制干扰对子滤波器工作状态的影响,使主滤波器能够充分、有效地利用子滤波器的信息。将这种自适应联邦滤波器应用到MINS/GPS组合导航系统设计中,并对系统进行了仿真。通过与基于常规联邦滤波器的组合系统的仿真结果比较可知,自适应联邦滤波器的组合系统在性能上有了显著的提高,且速度误差标准差控制在0.06m/s以内,位置误差标准差控制在3.3m以内,具有良好的导航精度。     

基于神经网络和遗传算法的光塑性图像处理

本文论述了神经网络、遗传算法结合起来在光塑性图像处理中的应用 ,以楔横轧工序的光塑性图像处理为例 ,三次投射时光塑性图像中各点都有对应的等差线条纹级数和等倾线角度 ,首先利用神经网络建立起图像参数与等效应变之间关系的数学模型 ,然后用遗传算法对图像参数进行优化 ,优化结果与实际情况吻合良好。由此证明神经网络、遗传算法完全可以较好地优化光塑性图像参数 ,并应用于光塑性图像处理领域     

Exploring cold regions autonomous operations

The US Army must update its vehicle fleet to be better equipped for potential future military conflicts in northern climates (US Army, 2017). This process involves considering manned, optionally manned, and unmanned vehicles as viable options in the future. Optionally manned and unmanned vehicles in the armed forces have substantial benefits because they can operate without direct driver input or are able to perform missions deemed too dangerous for troops. Optionally manned vehicles allow the driver to shift some, or all, focus away from the task of driving the vehicle. In some cases, these autonomous vehicles may perform better than a human driver by rapidly sensing and reacting to terrain changes. Onboard sensing and decision making are equally applicable to both fully autonomous and teleoperated vehicles. This work will focus on the terrain sensing, waypoint navigation, and teleoperation potential of an optionally manned or unmanned vehicle. Results from a vehicle demonstration on two different terrain conditions will provide the basis for additional terrain sensing and autonomous vehicle development work in the coming year.     

Performance comparison of artificial neural networks and expert systems applied to flow pattern identification in vertical ascendant gas–liquid flows

Instantaneous readouts of an electrical resistivity probe are taken in an upward vertical air–water mixture. The signals are further processed to render the statistical moments and the probability density functions here used as objective flow pattern indicators. A series of 73 experimental runs have its flow pattern identified by visual inspection assisted by the analyses of the void fraction’s trace and associated probability density function. The flow patterns are classified into six groups and labeled as: bubbly, spherical cap, slug, unstable slug, semi-annular and annular. This work compares and analyzes the performance of artificial neural networks, ANN, and expert systems to flow pattern identification. The employed ANNs are Multiple Layer Perceptrons, Radial Basis Functions and Probabilistic Neural Network, with single and multiple outputs. The performance is gauged by the percentage of right identifications based on experimental observation. The analysis is extended to clustering algorithms to assist the formation of knowledge base employed during the learning stages of the ANNs and expert systems. The performance of the following clustering algorithms: self organized maps, K-means and Fuzzy C-means are also tested against experimental data.     

Effects of boundary conditions and anisotropy on elastically bent silicon

In four-point bending, the rollers that are used for load application impose additional constraints on the specimen that affect the anticlastic specimen curvature and cause the specimen displacement and stress profiles to deviate from the pure beam bending case. In this study, x-ray microdiffraction is used to map both the principal and anticlastic curvatures of elastically bent, rectangular (100)-type Si strips possessing width:thickness ratios of 40:1. We quantify the amount of roller constraint and show that the region over which the anticlastic specimen curvature is affected away from the roller is approximately five times the roller diameter. Consequently, for bending tests used to determine Poisson's ratio, if a region on the sample that is free from roller effects is not chosen, measurement errors as high as 46% can occur. Furthermore, we show that, due to the anisotropy of single crystal Si, this roller-constraining effect depends on crystallographic orientation and is more pronounced when the principal bending axis lies along the <100> direction as compared with the <100> direction.     

Slow flow of a non-Newtonian fluid past a micro-capsule

The streaming motion past a spherical microcapsule is studied. The particle consists of a thin elastic membrane enclosing an incompressible fluid. Since the problem is highly nonlinear, a perturbation solution is sought in the limiting case where the deviation from sphericity is small. Obviously, the capsule remains nearly spherical when λ, the ratio of viscous forces to elastic (shape-restoring) membrane forces is small. In this limit, the rheology of the inside fluid is immaterial and the problem is essentially characterized by three parameters: λ, the Reynolds number Re (interia effect), and the Weissenberg number We (non-newtonian effect). The deformation is obtained explicitly under the restriction We<1, Re<1. It is shown that to leading order, the capsule deforms exactly into a spheroid which can be either oblate or prolate, depending mainly upon the elasticity number We/Re: for We/Re<0.57 the spheroid is oblate, while for We/Re>0.81 a prolate spheroid results. For 0.57<We/Re<0.81 additional details of the rheology of the membrane and of the suspending fluid are needed. The degree of the deformation is governed by the parameters λ Re. All parameters of the problem enter into the expression of the drag force. On a qualitative basis, these results are similar to those for droplets although major differences exist quantitatively.     

基于PTCNN的结构布局优化问题研究

对结构优化中的难点布局优化问题,基于交叉学科的优势,运用脉冲暂态混沌神经网络（PTCNN）的方法,对一个新型布局优化模型进行了寻优计算。在PTCNN算法中应用脉冲混沌动力自然解决了杆件的恢复和删除问题;利用神经网络的多神经元并行构成特点,通过参数调整方法解决了不同数量级变量问的耦合问题。因此PTCNN不仅是模型的寻优算法,更成为布局优化的一部分。此外构造了一个应力关联系数σ_AI,使面积变量随杆中应力大小按比例自适应下降。算例结果表明,基于PTCNN方法解决结构布局优化问题,有效且具有自适应性,布局优化效果明显。     

Photometric measurement of terrain roughness

Information on the roughness of terrain is an important factor in designing the rolling gear of cross-country vehicles. In view of a number of advantages—especially the great information density—of the photometric measuring of terrain roughnesses, this study report discusses the applicability of this technique as a new method to complement those already in use. The report demonstrates that the resolution limit of approximately ± 5 cm in a scale 1:2000 reproduction is sufficient, for all practical purposes, because the resulting total measuring error will be distributed over the entire profile-frequency range.A comparison between an accurate ground-based technique and photometric measurement is presented to demonstrate that, in the range of the profile wavelength of 1 m, the roughness root-mean-square-values differ only by a few mm, so that these differences may be neglected.It is suggested that, for assessing terrain roughness, range-pair classification be used in addition to the usually applied power spectral density, because range-pair classification describes the terrain amplitudes occuring in terms of their absolute magnitude.The report indicates evaluation errors that may result from non-stationary roughness profiles, and it shows the possibility of conducting the photometric terrain measurement in wider and less expensive steps, and of complementing this measurement by superimposing a suitable random noise signal on the measurement curve.     

随机神经网络优化方法及其计算机仿真

研究了结构优化问题求解的随机神经网络方法,可较好地克服Hopfield网络方法容量陷入局部解的缺点,针对实例,对随机神经网络和Hopfield网络进行了动态仿真,直观地给出了两种网络的动态并行运行,能量函数的动态变化过程及两种网络的全局性的差异,同时也仿真出了网络并行运行至稳定优化解的时间。     

Calculating fractal parameters from low-resolution terrain profiles

Driver comfort on rough terrain is an important factor in the off-road performance of wheeled and tracked ground vehicles. The roughness of a terrain has typically been quantified by the U.S. Army as the root-mean-square elevation deviation (RMS) of the terrain profile. Although RMS is an important input into many mobility calculations, it is not scale invariant, making it difficult to estimate RMS from low resolution terrain profiles. Fractal parameters are another measure of roughness that are scale invariant, making them a convenient proxy for RMS. While previous work found an empirical relationship between fractal dimension and RMS, this work will show that, by including the cutoff length, an analytic relationship between fractal properties and RMS can be employed. The relationship has no free parameters and agrees very well with experimental data - thus providing a powerful predictive tool for future analyses and a reliable way to calculate surface roughness from low-resolution terrain data in a way that is scale invariant. In addition, we show that this method applies to both man-made ride courses and natural terrain profiles.