An intelligent system for identifying track hauls requiring out-of-turn control of the railroad bed is proposed, the transmitting part of which is installed in one of the cars of rolling stock. The result of control of the railroad bed malfunctions through "means of communication" from the rolling stock is transmitted to the "system of establishing the presence of malfunctions on the track hauls" installed at the track maintenance department, or at the operations control service, which after a day after the formation of databases by analyzing it determines the hauls that need to be controlled out of turn. Taking into account that each state of the railroad bed corresponds to a group of diagnostic signs of the dynamic process occurring during the movement of trains on the railroads, using the estimates of the characteristics of noise and other characteristics of vibration signals, monitoring and signaling of the presence of a malfunction on the railroad bed is carried out. Here, if the current informative attributes do not exceed the reference ones, then it is considered that the technical condition of the track has not changed, and if it exceeds then it is assumed that there is the beginning of the latent period of track malfunction. The relevance of the proposed system is due to the fact that nowadays, despite the "scheduled control" of the railroad bed of track hauls from the influence of various factors, even a day after the control, there may be certain malfunctions that can cause a catastrophic derailment.

The paper presents an autonomous intelligent mobile robot for washing surfaces with a high-pressure liquid flow. There are structure, device and operation principle of a robot for selective cleaning of surfaces with a compact laminar jet and its main subsystems functionality considered. These subsystems consist of navigation subsystem in the conditions of additional objects presence in the room, subsystem of searching point or continuous contamination and determining their coordinates, control subsystems for the detergent flow by evaluation the rotation angles of the nozzle holders. Authors introduce the concept and solutions for direct and inverse task of positioning the end point of a fluid flow on a surface. The application of the Denavit-Hartenberg transformation and the parabolic kinematic law of fluid motion while flowing out of a point nozzle with a minimum spray angle is a base for the obtained direct task solution of positioning the end point of fluid flow on the treated surface. An algorithm for the inverse task solving of positioning the end point of a fluid flow on a surface is presented using Newton’s numerical method for solving systems of nonlinear equations. There was constructed a model of the movement kinematics of a laminar fluid flow section in the MSC Adams environment. This article considers a mobile robot model, which makes it possible to simulate the dynamics of the robot’s movement simultaneously with the movement kinematics of the jet cross section. The article noted advantages and prospects for further use of the proposed solutions in various industries.

The article focuses on development and modeling of neural network sliding mode (NSM) algorithms for controlling three-axis gimbal (TAG) orientation with camera of an unmanned aerial vehicle (UAV). The NSM algorithm is based on kinematic equations that describe the rotation and interaction of three rigid TAG components: yaw channel frame (YCF), roll channel frame (RCF), and pitch channel frame (PCF). The mathematical model of the TAG takes into account the interaction of three rigid components (YCF, RCF and PCF) are the TAG components, and the influence of unknown disturbances on the TAG. Disturbances (centrifugal force, gravity, friction that arise during TAG working) significantly complicate the mathematical model of the TAG. The problems are solved by synthesizing an adaptive sliding mode control (ASMC) using an artificial neural network (ANN) RBF. In RBF ANN, radial basis functions (Gaussoids) serve as nonlinear activation functions. In the description of the sliding control mode, disturbances are introduced that contain unknown parameters: influence of gravity, influence of the centrifugal force of inertia of the frames, etc. Unknown parameters of disturbances in NSM are estimated by using RBF ANN. The combination of sliding mode control and RBF neural network implements neural network sliding control for the TAG orientation. The modeling results of neural network sliding mode control in MATLAB prove that system using NSM has a large stability margin, is characterized by a higher quality of control processes and working quite stably in unknown and random disturbances environment compared to the classic sliding mode controller.

The development of environmentally efficient high-thrust aviation engines requires the improvement of automatic control systems. The operation and increased number of aircraft make aviation one of the largest sources of emissions of harmful substances, particularly nitrogen oxides, during fuel combustion. A particular challenge is the control of the combustion chamber of aviation gas turbine engines (GTE), as it is necessary to simultaneously meet the fundamental requirements for ensuring the stable operation of the engine and minimizing the emissions of nitrogen oxides. This paper presents a new approach to the control of the combustion chamber of aviation GTE. The proposed solution involves adjusting the fuel consumption between the collectors of the combustion chamber by introducing feedback on NOx into the GTE automatic control system using an adaptive virtual neural nitrogen oxide measurer, considering real-time "hard" operation mode and ensuring gas-dynamic stability of the combustion chamber. Gas-dynamic stability of combustion in the combustion chamber is ensured by the uniform distribution of the fuel-air mixture through transverse pulsations of concentration using homogeneous and diffusive dosers. When redistributing fuel, the engine operates in a stable mode, preventing flameout in the combustion chamber and the "vibrating combustion" mode. The simulation results in the MATLAB software package confirm the effectiveness of the new approach to designing the fuel consumption control system of the combustion chamber using an adaptive virtual neural nitrogen oxide measurer. The proposed system has a high potential for reducing the concentration of nitrogen oxide emissions, thereby enhancing the ecological efficiency of the aviation GTE combustion chamber operation.

One of the promising areas of joint use of unmanned aerial vehicles (UAVs) is the group air patrol of large territories. An important stage in the organization of this process is the planning of UAV flights. The paper considers the problem of optimal planning of flight routes for a group of UAVs when patrolling large-scale territories with several depots based on drones. An example of such territories can be hard-to-reach territorial waters or narrow border areas (coast, mountain and forest masses) of a State. It is assumed that the patrolled area has an elongated shape and can be divided into a chain of adjacent patrol zones prescribed by a separate UAV. The drone’s flight route passes through adjacent zones. The flight task performed periodically by each drone consists in moving it to a given flight zone, collecting and transmitting operational data to the control center. The optimization aspect of UAV flight route planning is to minimize the maximum route length when flying over all patrolled zones. The problem under consideration is mathematically formalized as a multiple traveling salesman problem (MZK) with several depots. Since it belongs to the class of NP-hard combinatorial optimization problems, approximate heuristic and metaheuristic approaches to its solution are of practical interest. A metaheutristic method for solving MZK using genetic algorithms is proposed. As model examples, the tasks of patrolling the land and sea borders of Vietnam are considered, the solution of which was obtained in the MATLAB environment using the Global Optimization Toolbox mathematical package.

The paper considers the task of coaxial unmanned Martian helicopter motion simulation in virtual environment systems. To solve this task, a nonlinear mathematical model of helicopter dynamics with blade flapping motion is presented. This model also includes nonlinear equations for computing rotor thrust coefficients. The proposed solution for helicopter control is to use the feedback linearization method of nonlinear equations. Based on this approach, a control of the vertical and horizontal helicopter motion was implemented. Regarding this, expressions for the roll and pitch angles of helicopter relative to required rotor thrusts and aircraft accelerations were derived. The control results in the relationship for collective and cyclic pitch angles for both rotors. To ensure these angles, PD regulators were used, implementing computation of voltages supplied to the helicopter’s swashplate actuators. The approbation of methods and algorithms proposed in the paper was carried out in our virtual environment complex, exemplified on the control of virtual coaxial helicopter using a training remote controller. Aiming this, software modules for helicopter dynamics simulation and control were developed and added to the complex. For this, a semi-implicit Euler scheme to integrate differential equations, and Newton’s numerical method to solve nonlinear equations, were used. The helicopter control is implemented by means of a functional diagram scheme which inputs are commands from the remote controller and measurements from virtual sensors, and the voltages for actuators are generated at the output. Approbation results showed the adequacy of the solutions proposed in the paper, which can be further used in creation of simulators designed for teaching operators to control a coaxial Martian helicopter.