IMPLICIT REPRESENTATION FOR THE MODELLING OF HYBRID DYNAMIC SYSTEMS

Differentmodels are proposed for the study ofhybrid systems that are composed of interactingcontinuous and discrete parts as,for instance,inRef.[1 ].Generally,the discrete part is modelledby an automation,and the continuous part by dif-ferent sets of state equations.Jump functions canbe added to represent the discontinuities.But noindication is given about the relations between thedifferent models and the jump functions.　　This paper isconcerned with the modelling forphysical systems with sw…     

电液力伺服系统的神经网络并行自适应预测 PI 控制

考虑电液伺服系统的复杂非线性和不确定性特性,提出一类基于神经网络的并行自适应预测PI控制结构,该结构使控制参数的调整和系统的实时控制操作可并行进行,不仅做到了神经模型和控制器的在线辨识和设计,而且避免了神经网络方法通常存在的实时控制的困难,使复杂系统的在线学习控制成为可能。仿真结果表明该控制器具有良好的适应性和鲁棒性。      

Moment equation description of interstellar hydrogen

The flow of interstellar hydrogen in the heliosphere can be studied using the moment equation approach. The Boltzmann equation is integrated over the velocity space to obtain the moment equations, the moment equations are then solved directly for the flow conditions. We present a closed system of moment equations. This approach can include anisotropic pressure when the distribution function is distorted into skewed ellipsoid.     

A unified approach to the design of decentralized detection systems

A unified approach to the design of decentralized detection networks with arbitrary topologies is proposed and analyzed. In this approach, a decentralized detection system of arbitrary topology is represented by a communication matrix that specifies the interconnection structure of the detection network. This matrix is used to derive the general decision rule using the person-by-person-optimal (PBPO) solution methodology. It is shown that the PBPO decision rule is a likelihood ratio test for statistically independent observations. The threshold of the test is shown to be a function of the decision input vector of the detector under consideration. This unified approach is used to obtain the PBPO decision rules of decentralized detection systems with various topologies. Various results in the literature are verified. In addition, the PBPO decision rules for a decentralized detection system with peer communication are presented. Numerical examples are presented for illustration     

Modeling quaternion errors in SDINS: computer frame approach

We propose new equivalent tilt error models which are applicable to the analysis of the terrestrial strapdown inertial navigation systems (SDINS), based on the quaternions. The currently available equivalent tilt error models, like the conventional Φ model of the gimbaled inertial navigation systems (GINS), are derived only by the true frame approach. The true frame approach has a computational disadvantage that it produces an error model where the attitude error equation is coupled with its position and velocity error equations. The motivation of this work is to solve this problem. As a result, two kinds of error models are derived. Among them, one is derived by the computer frame approach for practical onboard implementations. Thus, like the conventional GINS Ψ model, its attitude error equation is decoupled from the position and velocity error equations. The other is derived in order to show the relationship between the true frame approach and the computer frame approach which are applied to the quaternion-based SDINS. Thus, like the GINS δΘ model, it can be used to transform the error variables into each other which are calculated by the two different approaches     

航空机电系统测试性建模与分析新方法

针对航空机电系统测试性设计(DFT)需求,提出基于面向对象的贝叶斯网络(OOBN)与状态-测试关联灵敏度指标的系统测试性建模与分析的新方法。该建模方法能清晰地刻画系统故障与测试间的关联程度,反映复杂系统的层次结构关系。基于信息论的交叉熵原理提出状态-测试关联灵敏度指标,并给出计算方法。该指标反映复杂机电系统测试中的不确定性影响,克服了基于香农熵的测点评判分析方法的缺点,结合测试性建模获得的模型信息进行推理计算,可用于测试性的定量分析。运用该综合分析方法对飞机燃油系统进行测试性建模与分析,结果表明所提出的方法与指标在航空机电系统DFT中具有实用性。     

神经网络最优闭环制导律设计

对于复杂的非线性导弹制导系统 ,很难求得其解析的最优制导律 ,只能求得开环的数字解 ,不能适用于具有时变不确定性的导弹制导系统。利用神经网络的学习和推广能力 ,对开环的数字最优制导律进行离线的学习 ,作为闭环的神经最优制导律在线应用。研究分别选择系统状态变量和视线角速率等不同的神经网络输入对制导系统性能的影响 ,以及各种制导律的鲁棒性问题 ,并采用模块化神经网络结构提高神经网络的学习和推广能力 ,仿真结果得到一些有益的结论。     

关于非线性大系统的局部状态反馈和稳定区域估计

本文研究了非线性定常大系统利用局部状态反馈的控制以及稳定区域估计问题。文中给出的方法可用来处理较为广泛的系统,特别适合于“最小强相联系统”或具有类似性质的系统。对大型望远镜,应用此方法得到了相当好的结果。     

基于反馈误差学习的神经网络控制

研究了应用神经网络和ＰＤ反馈控制实现非线性系统的自适应跟踪问题。ＰＤ反馈控制器不但保证闭环系统的稳定性,同时其输出又作为训练神网络的参考信号。证明了通过选择适当的初始加权和加权的调节速率可以实现非线性系统的固定点跟踪。     

误差对同轴六分支分扭人字齿轮传动系统静均载性能影响

以同轴六分支分扭人字齿轮传动系统为研究对象,依据各齿轮受力状态建立该系统的静力平衡方程。考虑到制造误差和安装误差及输入输出轮浮动导致的错位,基于当量啮合误差理论,分析误差的存在性,最后根据系统功率闭环特征建立系统变形协调方程,形成了同轴六分支人字齿轮传动系统静均载分析方法,并结合实例求出系统各齿轮之间静均载系数及分支静均载系数。研究结果表明:在无误差或各齿轮误差均相同为常值时,第Ⅰ级各齿轮静态啮合力为1.773×105 N,第Ⅱ级各齿轮静态啮合力为3.673×105 N,系统具有很好静均载性能,系统分支静均载系数为1,该系统构成功率闭环误差可相互抵消;制造和安装误差幅值同时作用为50 μm时,求得制造误差下分支静均载系数变化幅度比安装误差下分支静均载系数要大,可知制造误差对系统静均载性能影响程度要大;分扭和并车误差幅值同时作用为50 μm时,并车级比分扭级静均载性能更容易受误差的影响,因此输出构件应该有浮动量。综上所述,随制造或安装误差增大或减少,都会对系统静均载性能造成不良的影响,其研究成果可为同轴减速器传动系统制造误差和安装误差精度确定,均载系数确定提供科学依据。      

Wireless sensor networks: scheduling for measurement and data reporting

An optimal load allocation approach is presented for measurement and data reporting in wireless sensor networks with a single level tree network topology. The measurement problem investigated involves a measurement space, part of which can be sampled by each sensor. We seek to optimally assign sensors part of the measurement space to minimize reporting time and energy usage. Three representative measurement and reporting strategies are studied. This work is novel as it considers, for the first time, the measurement capacity of processors and assumes negligible computation time which is radically different from the traditional divisible load scheduling research to date. Aerospace applications include satellite remote sensing and monitoring and sensor networks deployed and monitored from the air.     

Sensitivity analysis for probabilistic anti-resonance design of aeronautical hydraulic pipelines

In an uncertainty scheme, reliability and global sensitivity analysis is studied in this work, to provide helpful information for probabilistic anti-resonance design of vibration systems.Discussions show that the resonance failure problem can be viewed as a series system, in which input uncertainties are modeled by random variables. In order to quantitatively measure the contributions of input variables to the system reliability, a global sensitivity index is proposed, the properties of which are also discussed. Then the proposed index is tested with an aeronautical hydraulic pipeline system, which is under the excitation of pump vibration and at a risk of resonance failure. Sensitivity results under different failure criteria and variation coefficients are obtained and studied, from which significant and insignificant input variables can be identified.The proposed method provides a relatively new insight for anti-resonance design of engineering structures.     

Landing system verification based on petri nets and a hybrid approach

One of the most important activities of control system design is its verification. Verification ensures that the controlled system will behave as expected under any circumstances it may operate. In this context, the purpose of this paper is to introduce a new method for the verification of aircraft control systems. The focus of this method is on aircraft systems that are characterized as hybrid, i.e., that merge continuous and discrete dynamics. The method proposed is divided into two main parts: the system modeling and the verification of behavioral properties. In the first part, Petri net, differential equation systems, and object oriented concepts are used concurrently in order to model complex hybrid systems. In the second part, the distributed nature of the model is explored in order to decompose a complex verification problem into series of simple local problems. Linear logic is used as a basis of a theorem-proving approach for the verification from the discrete-event point of view. The verification method has been applied to a number of case studies. Among them is the landing system of a military aircraft, which is described in this paper     

Sensitivity Functions for Linear Discrete Systems with Applications

z-transform techniques are employed to establish general symmetry and simultaneity properties of the first sensitivity functions of the phase-canonical form of single-input, nth-order, linear, constant, discrete-time, controllable systems. It is demonstrated that computation of the first sensitivity function requires one nth-order model in addition to the system model. This simultaneity property is extended to arbitrary single-input, nth-order, linear, constant, discrete systems. In complete analogy with results presented for continuous systems, symmetry and simultaneity properties may be established for the computation of the /th sensitivity function begin{equation*}^{l}beta^{y} triangleq frac{partial^{l}y_{i}}{partialalpha_{Jl}partialalpha_{Jl-1}cdotspartialalpha_{J2}partialalpha_{J1}}|_{alpha=alpha_{0}} {rm for} substack{i = 1,2,ldots,n\ J_{k} = 1,2,ldots,n\ k = 1,2,ldots,l.}end{equation*} Extension of these results to multi-input systems is also mentioned. The usefulness of the simultaneity property is illustrated by applying the results to the design of a low-sensitivity optimal control law.     

Optimum selection of error control coding using neural networks

Significant performance improvements may be obtained in digital communication systems if error control coding is properly applied. However, selection of a coding scheme for specific applications is often a complicated task. The choice is affected by a set of system design goals. Some of these goals impose case-dependent conflicting requirements. Similar scheme selection problems exist in many engineering system design processes. A knowledge-combined neural network approach is developed and applied to optimum coding selection. The proposed approach utilizes a neural network trained not only by precedent examples but also by knowledge rules to draw conclusions. It is shown that artificial neural networks (ANNs) can provide effective solutions to the problems encountered in building systems that emulate a coding specialist's expertise     

基于Heun法的模糊控制系统

针对有些常微分方程求解难且不易用图形直观观察的特点,根据求解常微分方程的Heun法构造了一种模糊控制系统以辨识由常微分方程所描述的系统。这种模糊控制系统在预测状态的同时也能够逼近系统微分方程中的函数。     

控制系统的神经网络鲁棒设计方法

拓展了基于一个给定的模型考虑系统的摄动问题的这种传统的鲁棒控制设计方法,给定多个模型,设计多个鲁棒控制器,将有更大的容许摄动范围,针对多个不同控制器用传统方法很难同时实现的状况,用具有定性和定量多模态控制能力的神经网络同时实现了多个传统控制器的功能,方法简单,实现容易。将该方法用于飞机增稳系统的控制设计,仿真结果证明控制效果良好。     

Flight-critical distributed systems: design considerations [avionics]

With the proliferation of so-called "smart" components and the availability of small, low-cost, and high-speed data networks, avionics that have traditionally been centralized are becoming distributed. A distributed approach offers many potential benefits, such as reduced development time and cost, simplified system installation, increased flexibility for system expansion or modifications, and greater reuse of proven components. The distributed approach can also reduce the risk associated with design errors by splitting complex hardware and software into more manageable components. However, distributed systems also introduce new challenges in meeting real-time deadlines and providing fault tolerance. This paper examines the many design considerations and identifies the strengths and weaknesses of each. Emerging automotive drive-by-wire alternatives are compared for application to aerospace systems. This paper is based on a Draper Laboratory-sponsored effort to look at flight-critical distributed systems and to evaluate emerging hardware and software for building them.     

气动弹性剪裁中的响应值敏度

进行了一种用解析方法求气动弹性剪裁中的结构响应值敏度的研究。推导了结构的气动弹性响应(位移,振动形态和频率,静气动弹性效率和发散,颤振 )对设计变量导数的封闭解析表达式。设计变量可包含复合材料结构的铺层方向角。广义气动力导数计及正交振动形态导数的影响。经实例验算,所求得的敏度,与差分法相比,数值上吻合得很好。此方法已成功地应用于 CAE倡导发展的结构优化设计程序系统。     

Optimization Design System for Composite Structures Based on Grid Technology

To solve the topology optimization of complicated multi-objective continuous/discrete design variables in aircmit structure design, a Parallel Pareto Genetic Algorithm （PPGA） is presented based on grid platform in this paper. In the algorithm, the commercial finite element analysis （FEA） software is integrated as the calculating tool for analyzing the objective functions and the filter of Pareto solution set based on weight information is introduced to deal with the relationships among all objectives. Grid technology is utilized in PPGA to realize the distributed computations and the user interface is developed to realize the job submission and job management locally/remotely. Taking the aero-elastic tailoring of a composite wing for optimization as an example, a set of Pareto solutions are obtained for the decision-maker. The numerical results show that the aileron reversal problem can be solved by adding the limited skin weight in this system. The algorithm can be used to solve complicated topology optimization for composite structures in engineering and the computation efficiency can be improved greatly by using the grid platform that aggregates numerous idle resources.