编号 3 (2025)

A set $\Xi$ of conservative systems is considered, allowing oscillations of a given period. The problem of aggregating the set $\Xi$ into a coupled system with an attractive (in a large) cycle, in which the phase control law in the systems $\Xi$ is realized, is solved. An approach is developed in which a Van der Pol oscillator with an adjustable frequency is used as a control signal generator, acting through one-way links--control on the systems of the set $\Xi$: in the aggregated system, there are no direct links between the systems $\Xi$. The control system is described by autonomous equations.

A new analytical approach to static output feedback pole placement for linear time-invariant systems having the order equal to the product of inputs and outputs, and controllability and observability indices take the highest and the lowest possible values is proposed. The approach uses the Cayley–Hamilton theorem in relation to a closed-loop control system matrix. The applicability of the proposed method of static output feedback pole placement is not limited only by fourth-order systems with two inputs and two outputs. Besides, the method applicability does not depend on whether the system is reducible to modal control by state with fewer inputs or to modal observation with fewer outputs. On the examples of sixth-order multi-input multi-output systems not reducible to systems with fewer inputs or outputs, the process of obtaining analytical solutions to the problem of modal control by output, providing desirable pole placement is demonstrated. For systems from the considering class, the obtaining solutions are unique.

Continuous change of the support wheels radius of the vehicle provides the possibility of continuous control of its operational properties. These properties include: driving stability, controllability, cross-country ability, etc. It change is rational to make while minimizing heat losses in the actuator. The motion without slip of a rigid wheel with a continuously changing radius on a flat horizontal non-deformable surface is considered.

A mathematical model is proposed to describe the dynamics of constant-speed plane-parallel motion of the specified wheel while controlling the value of its radius. It is based on the hypothesis that there is not slippage in contact between the wheel and the supporting plane. The obtained mathematical dependencies allow us to determine the optimal control mode when changing the radius based on the criterion of minimum heat loss in the actuator that ensures this change. The results of the research can be used to control the parameters of vehicles operational properties, as well the wheeled robots moving on unorganized surfaces.

The problem of controlling groups of moving objects is considered. Each group is composed of simple objects of control and moves as a single (composite) object of control. During the movement, each group can be divided into subgroups or, conversely, connected with other groups. In this case, the dynamic capabilities of the group change, since they depend on its composition. The initial and final states of each simple object are specified. The problem of the fastest achievement of all final states, i.e. the problem of group speed, is solved.

Models and languages for descriptions and control of real time execution of discrete algorithmic processes are considered in the paper. Processes are divided into two groups: free and of real time processes. New operations for composition of parallel processes and language for their descriptions are proposed. The methods of parallel processes complexity evaluation are introduced.

The problem of group control of aircraft in conditions of active counteraction is considered. In this case, different rates of losses of one’s own forces, superior to the enemy or inferior to him, are taken into account. When assessing the effectiveness of paired counteraction, a multiplicative convolution is proposed. It takes into account the losses of both sides and their cost. A minimax algorithm for group target distribution of different types of aircraft is formed, taking into account the degree of their compliance in each pair. The use of artificial intelligence in the form of several neural networks in the formation of examples of their training on a modeling complex is proposed. Keywords: aircraft, dynamic priorities, survivability losses, minimax algorithm, artificial intelligence

The costs of using a cloud computing infrastructure depend on its optimal configuration. The task is to reduce these costs and to support service quality at agreed level at the same time. To solve these tasks, we need methods for predicting the quality of the network service provided. Such predictions based on the logs of computing infrastructure usage, machine learning and methods for estimating service execution times are the subject of these work. These logs data are obtained through measurements and previously collected data on the operation of the telecommunications infrastructure. Measurements of infrastructure performance and service performance generate large amounts of data. The article discusses various methods for reducing dimensions and isolating significant variables in order to estimate discrepancies between target and predicted characteristics. In experiments, combining the model of a random forest with the method of reducing the dimensions via the principal component analysis has shown the best way.

An approach to improving the structural consistency is considered. The purpose of the study is to select a method for combining identical objects into groups, since it is identical objects that can effectively exchange information and use the information obtained as a result of this exchange. To achieve this goal, a number of experiments with different methods were conducted, after which the best one was selected in terms of the target quality measure and latency. The proposed approach allows taking into account various characteristics of objects and the relationships between them. This ensures accurate determination of identical objects. The proposed approach also has an efficient implementation for distributed computing systems. This makes it fast even on large amounts of data. The comparison of the approaches under consideration is made using the example of the problem of searching for identical products for managing assortment and supplies.

The problem of stabilizing the origin for dynamical systems written as a chain of integrators of arbitrary length is solved, taking into account constraints on absolute values of the state variables. It is shown that linear stabilizing state feedback based on the modal control principle ensures the fulfillment of the specified constraints on the state variables when choosing roots of the characteristic polynomial of the closed-loop system to satisfy nonlinear inequalities. The derived conditions on the roots of the characteristic polynomial are based on the results obtained using the integrator backstepping design combined with logarithmic barrier Lyapunov functions. As an illustrative example, the solution to the problem of stabilizing a given angular position of a single-link manipulator is considered. It is guaranteed that magnitude constraints on values of the angular coordinate, angular velocity, angular acceleration and jerk of the manipulator end-point are satisfied.

The article describes the mechanical design, electronics and control software of the new quadruped robot MORS. The robot is intended for education and research fields. It is relatively small and lightweight and has modular design to simplify production and assembly. The robot is driven by brushless DC motors governed by the Field-Oriented Control. The trot gait is chosen as the basic one. The locomotion control algorithm is based on Zero Moment Point Preview Control method. The desired zero moment point is generated by tracking the intersection of projections of diagonally opposite legs. Software of the robotic platform is completely open sourced. To test the performance of all robot components, a number of experiments were conducted both in simulation and on the hardware platform. The experimental results demonstrate the efficiency of the robot walking in different directions with velocity up to 1.2 m/s and a load capacity of up to 7 kg, which is almost equal to its own weight.