This paper addresses the issue of developing a computerized system for processing information in the construction of the trajectory of an unmanned aircraft (UAC), a remotely-piloted aviation system (RPAS), or another robotic system. Resolving this task involves the neural network learning algorithms based on the mathematical model of movement. The construction of such a trajectory between two specified destinations has been considered that provides for the possibility of bypassing static and dynamic obstacles. The specified trajectory is divided into several smaller parts. The possibility of restructuring when changing the position of obstacles in space has been considered. A UAC flight control algorithm has been developed, which implies training a neural network for bypassing obstacles of different sizes. To predict the development of the situation when an object moves between two specified points in space, it is proposed to use the Q-Learning algorithm. It has been shown that the smallest number of steps required for moving along a specified trajectory is 18, the largest is 273 steps. In case of distortion during data transmission, the training of the neural network makes it possible to reduce the possibility of collision with obstacles by improving the accuracy and speed of information transfer between the on-board computer and operator. A system of the video support to moving objects was modeled; dependence charts of the normalized frame size at different parameter values were built. Using the charts makes it possible to determine the function of the maneuver intensity. Existing neural network learning methods such as CNN and LSTM were compared. It has been proven that the success rate reaches 74 % when using CNN only, while it amounts to 92 % at the hybrid application of CNN+LSTM. The simulation results have demonstrated the high efficiency of the developed algorithm

No abstract

An analogue interface circuit that contains a mea­suring chain, a transmitter, the input of which is connected to a DC source through an analog demultiplexer, and the outputs of the measuring chain through the analog multiplexer are connected to the measuring amplifier, and a two-channel analog-digital converter with simultaneous sampling is developed. The measuring chain is made in the form of three resistive current dividers, where one divider is formed by a resistive strain gauge and adjusting resistor, and two others — exemplary resistors. This work is aimed at finding possibilities for increasing the accuracy of measurements and interfering analog interfaces for remote measurements using resistive strain gauges by introducing a new structural scheme, which is shown on the Fig.2.

The object of research is to refine the linear sizes of the obtained 3D models via scanning, and reducing the numbers of errors when obtaining the model. For now, there is no accuracy method for transferring the actual sizes of an object to a 3D model. One of the most problematic places in the existing methods of transferring sizes from the object to the model is the error in the placement of dimensional markers due to inaccuracy, or poor quality of the received surface via scanning. A model of the instrument complex is used to implement an improved method of 3D scanning, based on the photogrammetric method. The advanced technology of construction and measurement of 3D models on the basis of photos on the principle of stereo pairs in combination with image projection is based on a combination of existing scanning methods. As well as the introduction of new functionalities, such as maintaining the actual sizes of an object, its textures, color and light characteristics, as well as improving the accuracy of linear sizes. As a result of the use of a standard, reference projections, and a new method of comparing photographs to build a 3D model, a 60 % increase in the accuracy of linear dimensions was achieved. This is due to the fact that the proposed new combined method incorporates all the existing most important aspects of scanning. And also has a number of features, such as the definition of boundary surfaces, automatic sizing, detection, and processing of glass and mirror surfaces. Due to this, this method eliminates the main disadvantages of the usual photogrammetric method – inaccuracies in the surface quality of the models, and inaccuracies in the transfer of linear dimensions. It is estimated that the combined method will allow to transfer the real size of simple objects in 3D with an accuracy of 99.97 % of the actual size of the object. It will also improve the quality of complex surfaces (boundary, glass, mirror) by at least 40–60 %, compared to other existing methods.

In the article is offered the structural scheme of error correction of the precision measuring channel of average active power for researches in laboratory conditions and exclusively within the limits of changes of the basic frequency of a network. A feature of the scheme is the use of calibration of functional transducers with piecewise linear approximation. The input voltages of these converters are a triangular voltage, which is formed at the output of the integrator by integrating rectangular bipolar meanders, which are formed from the output signals of the frequency divider phase shifter synchronized with the network by a device based on the original precision amplitude-pulse system of phase frequency tuning. Compensatory small-sized low-voltage transformers using measuring amplifiers with differentially split inputs are used as primary converters, which increases the linearity of the characteristic in a wide dynamic range, due to which additive-multiplicative correction of errors of the whole measuring path by two points is realized. The article presents the results of computer modeling of the main functional components of the measuring channel, which confirm its precision and high metrological characteristics. References 10, Figures 2.