Synthesis of control algorithms for aircraft trajectories in time-optimal climb and descent

Methods of synthesis of control of aircraft trajectories in time-optimal climb and descent are considered. Methods of construction of climb control are based on tracing the profile of the best energy rate of climbing in the plane “specific energy-altitude”. The synthesis of descent control is performed by tracing extremals that are the result of solution of the time optimal problem for the model of the aircraft motion in which the control variables are the trajectory inclination angle and the propulsion of the power unit. Examples of calculations using algorithms constructed on the basis of the described methods are given. These examples testify the efficiency of approaches to construction of the aircraft control upon climb and descent.     

输入有约束时的飞翼布局无人机姿态控制

针对大展弦比飞翼布局无人机的刚体运动与弹性运动耦合动力学模型,研究了输入有约束时的无人机姿态控制问题,提出了一种采用扩张状态观测器的反步滑模控制方法.首先,采用扩张状态观测器实时估计气动弹性模态和外界未知干扰的影响,并引入跟踪微分器避免了控制律中项数膨胀问题.然后,针对飞翼布局无人机多操纵面的配置和输入约束,给出了基于LMI的在线舵面分配算法.最后,针对指令滤波和输入约束情况下控制指令的滞后问题,设计了辅助补偿器对控制指令进行补偿.根据Lyapunov稳定性理论证明了该控制方法能够保证姿态跟踪误差收敛至有界,仿真表明存在复合干扰和输入约束时该方法具有良好的姿态跟踪性能.     

Steepest descent algorithms for neural network controllers andfilters

A number of steepest descent algorithms have been developed for adapting discrete-time dynamical systems, including the backpropagation through time and recursive backpropagation algorithms. In this paper, a tutorial on the use of these algorithms for adapting neural network controllers and filters is presented. In order to effectively compare and contrast the algorithms, a unified framework for the algorithms is developed. This framework is based upon a standard representation of a discrete-time dynamical system. Using this framework, the computational and storage requirements of the algorithms are derived. These requirements are used to select the appropriate algorithm for training a neural network controller or filter. Finally, to illustrate the usefulness of the techniques presented in this paper, a neural network control example and a neural network filtering example are presented.     

Group coordinate descent algorithms for nonconvex penalized regression

We consider the problem of selecting grouped variables in linear regression and generalized linear regression models, based on penalized likelihood. A number of penalty functions have been used for this purpose, including the smoothly clipped absolute deviation (SCAD) penalty and the minimax concave penalty (MCP). These penalty functions, in comparison to the popularly used Lasso, have attractive theoretical properties such as unbiasedness and selection consistency. Although the model fitting methods using these penalties are well developed for individual variable selection, the extension to grouped variable selection is not straightforward, and the fitting can be unstable due to the nonconvexity of the penalty functions. To this end, we propose the group coordinate descent (GCD) algorithms, which extend the regular coordinate descent algorithms. These GCD algorithms are efficient, in that the computation burden only increases linearly with the number of the covariate groups. We also show that using the GCD algorithm, the estimated parameters converge to a global minimum when the sample size is larger than the dimension of the covariates, and converge to a local minimum otherwise. In addition, we demonstrate the regions of the parameter space in which the objective function is locally convex, even though the penalty is nonconvex. In addition to group selection in the linear model, the GCD algorithms can also be extended to generalized linear regression. We present details of the extension using an example of logistic regression. The efficiency of the proposed algorithms are presented through simulation studies and a real data example, in which the MCP based and SCAD based GCD algorithms provide improved group selection results as compared to the group Lasso.     

Comparing descent heuristics and metaheuristics for the vehicle routing problem

Three improvement heuristics for the vehicle routing problem are considered: a descent heuristic and two metaheuristics Simulated Annealing and Tabu Search. In order to make an in-depth comparison of the performance of these improvement heuristics, their behavior is analyzed on a heuristic, time-sensitive level as well as on a parametric level. The design and the results of the experiments are outlined. The external validity of the conclusions is discussed.Scope and purposeTabu Search (TS) and Simulated Annealing (SA) have demonstrated to be appropriate metaheuristics for solving NP-hard combinatorial optimization problems, such as the vehicle routing problem with side-constraints. In order to compare the performances of both metaheuristics with each other and with a traditional descent implementation, a comparison of the best solution independent of computing times is fundamentally wrong because metaheuristics have no unambiguous stopping criteria, as opposed to traditional descent implementations.     

基于鲁棒特征结构配置的飞行器姿态角控制

针对传统特征结构配置方法鲁棒性不强的缺点,提出一种基于$L_{1     

Block coordinate descent algorithms for large-scale sparse multiclass classification

Over the past decade, ? ₁ regularization has emerged as a powerful way to learn classifiers with implicit feature selection. More recently, mixed-norm (e.g., ? ₁/? ₂) regularization has been utilized as a way to select entire groups of features. In this paper, we propose a novel direct multiclass formulation specifically designed for large-scale and high-dimensional problems such as document classification. Based on a multiclass extension of the squared hinge loss, our formulation employs ? ₁/? ₂ regularization so as to force weights corresponding to the same features to be zero across all classes, resulting in compact and fast-to-evaluate multiclass models. For optimization, we employ two globally-convergent variants of block coordinate descent, one with line search (Tseng and Yun in Math. Program. 117:387–423, 2009) and the other without (Richtárik and Taká? in Math. Program. 1–38, 2012a; Tech. Rep. arXiv:1212.0873, 2012b). We present the two variants in a unified manner and develop the core components needed to efficiently solve our formulation. The end result is a couple of block coordinate descent algorithms specifically tailored to our multiclass formulation. Experimentally, we show that block coordinate descent performs favorably compared to other solvers such as FOBOS, FISTA and SpaRSA. Furthermore, we show that our formulation obtains very compact multiclass models and outperforms ? ₁/? ₂-regularized multiclass logistic regression in terms of training speed, while achieving comparable test accuracy.     

Calculation of radiation in gas around a descent space vehicle

The problem of numerical modeling of the effect of radiation processes on gas dynamic parameters around a descent space vehicle in the Earth’s atmosphere is studied. A relatively new approach is suggested of direct modeling of radiation processes oriented to the use of high performance multiprocessor computer systems. The physical model of radiation transfer is described by a diffusion approach with a discretization of radiation and absorption spectra and a set of quasigasdynamic equations. The computations were carried out on both triangular and hybrid locally refined grids. The gas-flow modeling results taking into account radiation processes are represented for bodies of different shapes.     

Automated guided vehicle systems,state-of-the-art control algorithms and techniques

Automated guided vehicles (AGVs) form a large and important part of the logistic transport systems in today's industry. They are used on a large scale, especially in Europe, for over a decade. Current employed AGV systems and current systems offered by global manufacturers almost all operate under a form of centralized control: one central controller controls the whole fleet of AGVs. The authors do see a trend towards decentralized systems where AGVs make individual decisions favoring flexibility, robustness, and scalability of transportation. Promoted by the paradigm shift of Industry 4.0 and future requirements, more research is conducted towards the decentralization of AGV-systems in academia while global leading manufacturers start to take an active interest. That said, this implementation seems still in infancy. Currently, literature is dominated by central as well as by decentral control techniques and algorithms. For researchers in the field and for AGV developers, it is hard to find structure in the growing amount of algorithms for various types of applications. This paper is, to this purpose, meant to provide a good overview of all AGV-related control algorithms and techniques. Not only those that were used in the early stages of AGVs, but also the algorithms and techniques used in the most recent AGV-systems, as well as the algorithms and techniques with high potential.     

基于遗传算法和梯度下降的RBF神经网络组合训练方法

在使用基于梯度下降的径向基函数（RBF）神经网络学习方法时，由于网络目标函数误差曲面极其复杂，因而产生了网络收敛速度慢，且容易陷入局部极小，网络初始值的设置对网络训练结果影响很大等问题。基于遗传算法的训练方法能够摆脱陷入局部最优的困扰，但遗传算法的局部搜索能力不够，从而影响网络的训练效果。为了解决上述问题，在研究两种算法特点的基础上，提出一种组合训练方法，用提出的训练方法对UCI中的部分数据集进行了仿真实验，并将实验结果与传统方法下的结果进行了比较，实验结果表明新方法是有效的。     

基于非线性L1自适应动态逆的飞行器姿态角控制

钊对常规动态逆控制器不能有效抵消系统中的不确定性这一缺点,提出了一种非线性L_1自适应动态逆控制方法.该方法能够克服常规动态逆的不足,在保证系统鲁棒性的前提下,提升飞行器姿态角控制效果.首先,采用时标分离原理,将姿态角控制系统分为内外两个回路:外回路采用常规动态逆控制器,用于姿态角的跟踪控制;内回路采用非线性L_1自适应控制器,用于角速率的控制.其中,L_1自适应控制器由静态反馈控制器和自适应控制器组成:静态反馈控制器通过状态反馈实现,用于保证内回路的稳定和具有期望的闭环特性;自适应控制器由状态观测器、自适应律和控制律组成,用于抵消系统中的不确定性.其次,对所提控制方法的稳定性进行了分析,结果证明了该控制方法能够保证内回路的稳定和外回路的误差有界.最后,在综合考虑多种不确定性的情况下,将本文提出的非线性L_1自适应动态逆控制方法用于某无人飞行器姿态角控制,仿真结果验证了该控制方法的有效性和鲁棒性.     

Design and analysis of lateral motion control algorithms for an unmanned aerial vehicle with two control surfaces

This paper considers control of an unmanned aerial vehicle with two horizontal aerodynamic control surfaces in the modes of angular stabilization and coordinated turn. We analyze the linear yaw angle control algorithms based on roll angle variation in terms of their performance. In addition, we suggest an algorithm that ensures a required performance of yaw angle control. Some simulation results are given.     

Hybridization of gradient descent algorithms with dynamic tunnelingmethods for global optimization

An algorithm based on gradient descent techniques with dynamic tunneling methods fur global optimization is proposed. The proposed algorithm consists of gradient descent for local search and a direct search scheme, based on dynamic tunneling technique, for repelling away from local minimum to find the point of next local descent. This search process applied repeatedly finds the global minimum of an objective function. The convergence properties of the proposed algorithm is validated experimentally on benchmark problems. A comparative computational results confirm the importance of dynamic tunneling in gradient descent techniques     

Intelligent control for a remotely operated vehicle

This article considers the tracking control problem of an underwater vehicle used in the exploitation of combustible gas deposits at great sea depths. The vehicle is subjected to different load configurations that introduce considerable variations of its mass and inertial parameters. In this work it is assumed that the possible vehicle configurations are known, but the time instants when the changes occur and the new vehicle configuration following the change are unknown. A neural network-based switching control is proposed for the considered mode-switch process. This solution simplifies the control scheme implementation and reduces the control signal chattering.     

Heuristic algorithms for a three-dimensional loading capacitated vehicle routing problem in a carrier

In this paper, we present heuristic algorithms for a three-dimensional loading capacitated vehicle routing problem arising in a real-world situation. In this problem, customers make requests of goods, which are packed in a sortment of boxes. The objective is to find minimum cost delivery routes for a set of identical vehicles that, departing from a depot, visit all customers only once and return to the depot. Apart of the usual 3D container loading constraints which ensure that the boxes are packed completely inside the vehicles and that the boxes do not overlap each other in each vehicle, the problem also takes into account constraints related to the vertical stability of the cargo and multi-drop situations. The algorithms are based on the combination of classical heuristics from both vehicle routing and container loading literatures, as well as two metaheuristic strategies, and their use in more elaborate procedures. Although these approaches cannot assure optimal solutions for the respective problems, they are relatively simple, fast enough to solve real instances, flexible enough to include other practical considerations, and normally assure relatively good solutions in acceptable computational times in practice. The approaches are also sufficiently generic to be embedded with algorithms other than those considered in this study, as well as they can be easily adapted to consider other practical constraints, such as the load bearing strength of the boxes, time windows and pickups and deliveries. Computational tests were performed with these methods considering instances based on the vehicle routing literature and actual customers’ orders, as well as instances based on a real-world situation of a Brazilian carrier. The results show that the heuristics are able to produce relatively good solutions for real instances with hundreds of customers and thousands of boxes.     

一种无振荡误差的车载惯导系统测姿方法

车载GPS/INS组合导航系统中的卫星信号易受环境影响出现失锁,而惯导独立工作时处于无阻尼状态,其特有的机械编排导致解算的姿态信息中存在舒勒振荡和傅科振荡误差,因此无法满足测量系统对姿态信息的高精度需求。为此,提出一种基于罗经方案的无振荡误差的车载惯导姿态确定方法,通过将加速度计测量的载体运动加速度进行低通滤波处理,引入惯性坐标系并间接地建立载体与导航坐标系之间的姿态转换矩阵,改变惯导无阻尼解算模式并实现系统无振荡性误差的姿态信息的求解。试验结果表明,论文提出的无振荡误差测姿方法具有为车载平台提供准确姿态信息的可行性与可靠性。     

Uzawa algorithms for coupled Stokes equations from the optimal control problem

Coupled Stokes equations obtained from the optimal control problem subject to Stokes equations can be approximated by Uzawa finite element methods. The linear convergence is provided with an optimal relaxation parameter for the Richardson pressure update. This work was supported by KRF-2005-070-C00017.     

Robust convergence of the steepest descent method for data-based control

Iterative data-based controller tuning consists of iterative adjustment of the controller parameters towards the parameter values which minimise an H ₂ performance criterion. The convergence to the global minimum of the performance criterion depends on the initial controller parameters and on the step size of each iteration. This article presents convergence properties of iterative algorithms when they are affected by disturbances.