Home – Home – Jornals – Avtometriya 2019 number 4

The previously developed technology of multicriteria optimization of deterministic parabolic distributed parameter systems can be applied in situations with interval uncertainties of parametric characteristics of controlled objects. Minimax tasks of mathematical programming are formed in accordance with a control strategy based on the principle of the best guaranteed result on an expanded set of arguments, which includes all permissible realizations of uncertain factors, in the case of estimation of a given approximation accuracy to a required finite state of the system in uniform metric. These tasks are solved by an alternance method that uses the Chebyshev properties of desired extermals and fundamental laws of the problem domain. Of independent interest is the given example of multicriteria optimization of temperature regimes of induction heating of metallic semifinished products before plastic deformation operations in technological complexes for pressure treatment of metal

The problem of estimating the state vector of a bilinear dynamic system with a vector entrance is under consideration. It is shown that, with fulfillment of multiple restrictions, a Luenberger state observer can be formed, which ensures that the estimation error tends to zero. A step-by-step algorithm for determining the matrices of observer coefficients is proposed, based on matrix canonization and modal control methods

This paper describes the problem of automation of choosing the parameters of an electric hydraulic drive controller for a single-channel strength load stand for airframe elements. A parameter configuration method for a proportional-integral stand controller is proposed, based on the results of relay feedback tests and motion separation method. The results obtained make it possible to automatically configure the controller parameters of the strength test stand and improve the accuracy of implementation of force loading cyclograms of structures, which eventually makes it possible to accelerate strength tests. The results of analyzing the implementation of this technique on an experimental stand are given.

Results of studying the system used to control the flight of a hybrid aerial vehicle prototype in transitional regimes are presented. The hybrid vehicle is equipped with a quadrotor and main propulsion units. The control system developed in the study includes subsystems for attitude stabilization, vertical takeoff, hovering, and landing, as well as a control subsystem operating in transitional regimes from quadrotor to aircraft modes and back.

Operating and engineering characteristics of a hybrid aerial vehicle are presented. A mathematical model in the MATLAB/Simulink environment is described. Results obtained by means of hardware-in-the-loop simulations and confirming the efficiency of the proposed control method are reported

A procedure for designing PID controllers for the class of second-order nonlinear non-stationary objects is proposed. To reduce possible control jumps, its proportional differential components are transferred to the system feedback channel. It is shown that the presence of a special fast response differentiating device in the system generates fast processes against the background of slow operating processes, which are identified using the technique of separation of movements. The controller designed on this basis ensures invariance of the system under external uncontrolled disturbances as well as under a change in the parameters of the object. The results of numerical simulation of the system in MATLAB illustrate its basic properties.

The issues of local navigation of mobile robots are considered. The method of artificial potential fields as a popular method for local navigation of robots is analyzed. The issue of designing attractive potential fields is investigated. The problem of local minima (traps) is discussed. A new method for avoiding obstacles by a robot moving to a target is proposed

A numerical method for designing operating software for digital controllers of hybrid continuous-discrete control systems is considered. The method is based on implementing digital controller control algorithms in the form of rule systems and uses a fuzzy optimization procedure to determine the controller operation rules. The continuous part of hybrid control systems consists of multi-operational process chains, which are nonlinear, multidimensional and multi-loop controlled objects, and the discrete part is digital controllers. The designed digital controllers ensure satisfaction of the most general requirements for controlled processes, including fuzzy ones.

This paper presents a comprehensive simulation model of underground coal mining processes in a fully mechanized long mining face. A number of simulation experiments have been performed. The face performance is estimated as a function of various factors such as the technical parameters of mining machines, length of lava, gas factor, shearer operation process chart, distributed geomechanical characteristics of coal seam.

This paper presents the results of analysis of the influence of interfering factors on changes in the informative parameter of a single-turn eddy current sensors, which are used as initial data in the study of a system for measuring radial clearances with self-compensation of temperature effects on the sensor. Conversion functions of the measuring circuit with a connected single-turn eddy current sensor are obtained in analytical form. The conversion functions are dependences of the extreme values of codes at the measuring circuit output on radial clearances and temperature effects on the sensor taking into account the influence of interfering factors (neighboring blades, blade temperatures, axial displacements of the gas turbine engine compressor rotor). Quantitative estimates are obtained for the influence functions of these interfering factors on the calculated difference of the extreme values of codes in the system, which characterizes the measured parameter.

Time-optimal algorithms of spatial localization of random pulsed-point sources are proposed, and their parameters are calculated. These sources have a two-step probability density in the search interval and are detected due to generation of instantaneous pulses at random time instants (delta functions). The advantage of the developed optimal localization algorithms over the search procedures that ignore specific features of the spatial distribution of the detected pulsed-point objects is demonstrated.

A problem of detection of 3D objects on the basis of imagery information provided by two or three different sensors is considered. The efficiency of object detection with the use of algorithms of integration and data fusion of information from different sensors and also by means of formation of a unified synthetic image from two initial images is estimated. Conditions are determined at which the use of the information integration algorithm and the method of forming a single synthetic image provides a lower efficiency as compared to the use of only one image among those available for detection.

This paper describes a neural network architecture of edge detection. Different filters for first-layer neurons are compared. Neural network learning based on a cosine measure algorithm shows much worse results than an error backpropagation algorithm. Optimal parameters for first-layer neuron operation are given. The proposed architecture fulfills the stated tasks on edge selection.