Vol 58, No 3 (2019)

The problem of designing a control law that guarantees the asymptotical stability of a moving object in a neighborhood of an instable equilibrium of the isolated system is considered. The desired control law minimizes the energy cost function that is assumed to be proportional to the integral of a positive definite quadratic form of the control vector. If the motion equations can be decomposed into globally asymptotically stable and instable subsystems, then this control law depends only on the state vector of the instable subsystem. The optimal control law can be obtained using time reversal from the motion equations of the instable subsystem.

The two-dimensional problem on the fastest turn of a rigid body by moving an internal mass is considered. It is assumed that the rigid body with an internal movable mass is a closed mechanical system. In the general case, equations for the trajectory of the motion of mass, optimal control, and the Bellman function are represented in terms of elliptic integrals that include two unknown variables. To determine these variables, two nonlinear scalar equations based on boundary conditions must be solved numerically. If the location of the mass at the terminal time is not specified, then the problem is reduced to solving a single scalar equation on the known interval, and the corresponding root always exists and is unique for arbitrary boundary conditions. In the special case when the internal mass is small compared to the mass of the rigid body, optimal trajectories in the form of circular arcs are obtained. The relation between approximate and exact solutions is investigated, and numerical examples of using these relations are discussed.

A synthesis problem for the optimal control on average of a bunch of trajectories of a nonlinear deterministic dynamic system that originates from the given set of initial states is considered, and the control is performed using the results of estimation of the state vector generated by a nonlinear state observer using the incoming measurements. The control laws for the model of the control plant and for the observer are found simultaneously by solving the parametric optimization problem by one of the metaheuristic algorithms. The proposed algorithm for solving the problem is illustrated by two examples of control of the three-dimensional chaotic nonlinear dynamic systems.

The results of a numerical and analytical solution of the optimal control problem for a rigid body rotation with a combined criterion of the control process quality and the use of quaternions are presented. The optimization criterion takes into account the duration of the control process, the total impulses of the squared angular velocity of the body and the magnitude of the control moment applied to the body. The controlled motion of a rigid body may contain special control regions.

Scheduling in multichannel distributed information processing systems is studied. The problem is solved for systems-on-a-chip and is considered in its extended statement when the system architecture is specified from the point of energy efficiency prior to scheduling.

Algorithms for scheduling tasks in real-time systems with integrated modular architecture that are based on finding the maximum flow in a transportation network are proposed. Results of the experimental evaluation of these algorithms for a single processor and multiprocessor version of the computation scheduling problem are discussed.

The upper bound on the complexity of the potential transformation algorithm for solving cyclic games on graphs is improved. This bound is close to the lower bound on the complexity of the potential transformation algorithm. The optimal deviation problem is reduced to a cyclic game on a directed graph.

In this paper, we propose a training system for visual markers that provides the generation and subsequent recognition (under real-world conditions) of stylized images that contain information encoded by a sequence of bits. New types of neural network layers that make the recognition process tolerant to external noise are developed. The training process is based on the end-to-end principle, which enables automatic training for the intermediate stages of the model. The experimental results that demonstrate the performance of the system in encoding and decoding artificial markers are presented.

A method for fusing images of different spectra by using generative adversarial networks is proposed. An original architecture of a FusionNet neural network is developed based on pix2pix. It enables the synthesis of a complex (integrated) image that comprises the most informative fragments of different-spectra images, thus being more informative than any of these individual images. A technique for generating training and test sets, as well as the process of data augmentation, is described. The operation of the proposed image fusion method is demonstrated on some real-world infrared and visible images.

Two problems associated with low-altitude aircraft routing based on digital terrain maps (hidden routing through a mountain range and routing to avoid radar detection zones) are considered. The features and differences of each problem compared to traditional routing problems and solutions based on Dijkstra’s algorithm or its modification are analyzed.

The paper is devoted to the applied aspect of a new technology for calculating the laws of program controls for the orientation of a spacecraft (SC) [1]. The two-point boundary problem of reorienting a SC by gyrodynes is solved. The results of the numerical implementation of a new method of forming the program laws of reorientation that do not generate singular states of the executive gyro system are presented. The numerical implementation is carried out using the example of a dynamic model of a space telescope with real characteristics and six gyrodynes, which make up three pairs with collinear precession axes. The obtained results demonstrate the principal novelty of the new method of calculating the controls: the calculated laws of the telescope’s guidance lack the singular values of the precession angles of the gyro units of the executive gyrodynes. When the gyro units enter these positions, the space telescope loses controllability and the guidance process is interrupted. The absence of singular positions of gyro unit points guarantees the non-stop execution of the calculated laws with the required accuracy and speed.