The object of this study is a single-channel dual mass mechatronic system with flexible kinematic links that is widely used for controlling the motion of work machines. In the event of difficulty in measuring all state coordinates relating to the object, the required control quality may be ensured by using dynamic "input-output" (polynomial) controllers synthesized by the polynomial modal control method. The problem of studying the influence of the internal parameters of the object on the degree of its controllability and observability, and also on the parametric robustness of the synthesized system, is addressed. Methods are developed for improving the above system properties of the object or its computational model, and for creating on this basis a robust mechatronic system with a polynomial controller. For a comparative evaluation of the degree of controllability and observability, of various kinds of objects, use is made of diagonal forms of representing controllability and observability Gramians, and also the first norms of matrices reflecting the degree of closeness of system properties of the object represented in real coordinates and canonical forms of controllability and observability. The influence of small time constants (rapid motions) of the power converter, electromagnetic circuit of the electric motor and mechanical section, and also the internal friction value of kinematic transmission on system properties of the object and parametric robustness of synthesized systems, is researched. The negative conditions of the above influence are identified and methods of its compensation are proposed, based on further reduction of the above parameters and their exclusion from computational object models to improve initial system properties and synthesis of robust systems, and also on increasing the speed of the main control loop with elimination of additional boosts by correcting the inertia of the outof-loop prefilter. In the presence of internal friction in the kinematics leading to zero in object transfer function, improvement of system properties in the computational model is ensured by using the developed Gramian method, introducing additional virtual control channels and performing modal synthesis of the basic polynomial controller based on the improved object model.

The problem of controlling a single-link technological manipulator on a mobile platform with time delays in the control system is considered. The delays appear, in particular, in the actuators of the manipulator. The mobile robot contains a wheeled group and a tracked group. The movement of the wheeled and tracked groups relative to each other allows reconfiguring the transport system to overcome obstacles. The manipulator is designed to perform technological operations, such as clearing rubble in emergency situations. A bilinear model for an onboard single-link manipulator is investigated. A mathematical model with a delay of the manipulator is constructed to find optimal control. The model of the manipulator control process is described by a nonlinear model. The control moment of the actuator is formed using regulators with an aftereffect from the beginning of the movement to the current moment. The task of optimal control of the manipulator is to determine the admissible control that minimizes the functional describing the deviation of the manipulator movement angle from the specified position and the proximity of its angular velocity to zero. The admissible control is a piecewise continuous function. A control minimizing the functional, depending on time and the measured trajectory is found. The functional and control correspond to the Bellman equation. The problem of synthesizing the optimal control of a single-link manipulator is reduced to finding a solution to a system of differential equations that satisfies the boundary conditions at the ends of the control interval.

In modern automated systems and robotics, machine vision is widely used to solve various types of applied problems without human participation. This is a scientific field of artificial intelligence, and related technologies for obtaining images of real-world objects, processing them and using the results obtained. Most machine vision subsystems use artificial neural networks to solve problems such as detection, classification and segmentation of objects. The effectiveness of machine vision subsystems can be assessed using many criteria, the key one of which is the accuracy of the corresponding problem solution, for example, the accuracy of object classification. The use of traditional methods for increasing accuracy based on optimizing the structure of the neural network and neurons, selecting hyperparameters, does not guarantee the stability of the results when image obtaining conditions change, for example, when changing lighting, shooting angle, noise. The number of false detections of objects in unacceptable positions, or the number of missed objects and classification and segmentation errors increases and becomes unacceptable. An alternative way to increase accuracy is to use data augmentation methods for network training obtaining synthetic images that provide special properties of the training sample. However, studies devoted to augmentation methods do not analyze the image structure in terms of saliency maps, and it does not allow to produce effective augmented data for training.
To solve the problem of increasing the accuracy of neural network image processing, an iterative augmentation method based on new principles of image synthesis has been developed. The method will allow countering false activations of neurons by reducing the influence of non-key features on the image saliency map on the result of object classification. The proposed method was used to train neural networks of a dental robot simulator. Software has been developed that allows synthesizing new images automatically and training the network. When using the proposed augmentation method, an increase in the accuracy of object classification by 2-10 % is observed.

This paper presents a technique for synthesizing a finite-time tracking control law, based on the Lyapunov function, for quadcopters. The dynamic model of quadcopters is an underactuated system with six degrees of freedom. To synthesize a control law for the system, a control architecture is first constructed, followed by the derivation of control laws for each subsystem. The convergence property of each control law is ensured through the use of a virtual system in the form of a strict-feedback system, which is employed to synthesize the control laws. The control law is derived via a diffeomorphism between the subsystem and the virtual system. The finite-time convergence property is guaranteed using a special Lyapunov function. Simulation results validate the effectiveness of the designed control laws.

The paper presents the issue result of searching a non-stationary object in a topologically closed bounded area by a multi-agent system of quadrocopter-type aircraft. Thermal motion equivalent method (TMEM) method is proposed as a control algorithm for the multi-agent system. The TMEM is based on the similarity of the motion of molecules. In the case of TMEM it is possible to estimate swarm stability and performance of swarm problem solving using integral parameters already known from thermodynamics: RMS velocity, frequency of interactions and others. The aircraft motion similar to a molecular dynamic in TMEM is provided by the control system of each agent interacted with others by data exchanging. The paper demonstrates the effectiveness of applying the TMEM method to the search for a dynamic target in a bounded area based on numerous simulations in a specially designed MASPlatform environment representing a virtual polygon. The MASPlatform software implements the motion of thirty agents in a space of 300 by 300 meters. The dynamics of each agent is represented by a system of nonlinear differential equations with a quaternionic controller. The inference of efficiency is made based on multiple simulations of the ICD search issue.
In this paper, estimates of swarm numbers as a function of the search space are obtained. The paper proposes an approach to estimating the search time of a non-stationary object with known dynamics by a swarm, whose agents interact using TMEM in a bounded area, based on the thermodynamic parameter "mean free path length". This parameter is important for real systems where there are constraints on the agent’s energy resource.

The problem of providing the inspection motion of one spacecraft (inspector) relative to another spacecraft (reference spacecraft) moving in a circular orbit is being solved. Inspection refers to the motion of the inspector around the reference spacecraft. In unperturbed motion, it is possible to select initial conditions that provide a closed trajectory of the inspector’s motion relative to the reference spacecraft; however, the influence of the Earth’s oblateness from the poles will lead to the evolution of such a trajectory. The paper presents two approaches that make it possible to select the initial parameters of the motion of the reference spacecraft and the inspector, which provide inspection motion taking into account the second zonal harmonic J2 of the Earth’s gravitational potential (oblateness of the Earth from the poles). In this case, the inspector’s motion is considered in the plane of the circular orbit of the reference spacecraft. The first approach is to select the initial positions of the reference spacecraft in its orbit, at which the oblateness of the Earth does not have a significant effect on the relative trajectory of the inspector. This motion of the inspector is close to the unperturbed relative trajectory. Analytical relationships are obtained that allow one to select the necessary motion parameters. The second approach is to select the relative velocity of the inspector, ensuring inspection motion, taking into account the influence of the Earth’s oblateness. This approach based on the equality of the total orbital energies of the inspector and the reference spacecraft.