The scientific work consists of an introduction, four chapters, conclusion, 4 applications. The total volume is 95 pages, of which 80 basic text, 50 figures, a list of 4-page links and 4 applications on 11 pages.

**Relevance of the topic. **Research in the direction of orientation systems aimed at addressing one of the main objectives-improved accuracy due to system upgrades (reduction of instrumental errors) or the use of new data-processing algorithms (decrease methodological errors). Progress in micromechanical systems (MEMS) makes it possible to use miniature inertial sensors in various devices. The accuracy characteristics of devices based on these sensors are not very high. There is a problem of using new methods of modeling and using data from various sources of information - the problem of integration. Therefore, to achieve the maximum possible efficiency of such systems, great attention must be paid to their calibration, investigation of possible errors of a different nature, the introduction of filtering algorithms, and also to the evaluation of the accuracy with the use of sensitive elements that are part of the orientation system.

Proceeding from this, such a systematic approach to solving the problem of increasing the accuracy of the system will allow us to obtain the maximum possible exact characteristics on the basis of which the algorithm consists and therefore is an actual research work.

**Target** dissertation work is to develop an orientation system and its calibration, development and implementation of a neural network model of the Earth's magnetic field to increase the accuracy of the system.

Achieving the goal involves the following tasks:

- Analysis of the current state of methods for increasing the accuracy of orientation systems by introducing filtering algorithms;
- Calibration of sensing elements;
- Determination of the coefficients of the complementary filter, for filtering and integrating information from sensors;
- Finding a neural network for approximating the mathematical model of the Earth's magnetic field.
- Use of the system algorithm to determine the orientation of the object based on the output signals of the primary sensors;
- Experimental confirmation of the results.

**Research Methods.** Calibration of inertial sensing elements. Finding the spectral density and determining the filter coefficients. Simulation of the neural network for the Earth's magnetic field model. Determination of the angular position of the orientation system using the orientation algorithm.

**Scientific novelty of the dissertation** is to improve the accuracy of the orientation system for aircraft, vehicles, satellites, mobile phones, control devices ("gamepad", "joystick"), robotics by introducing the Earth's magnetic field, reducing the time of its calculation with a neural network, using the calculated complementary filter and the integration of the output signals from the sensors for mutual noise filtering.

**Keywords:** algorithm, orientation system, neural network, IGRF, calibration, complementary filter, integration.