2025 , Volume 30, № 5, p.50-65

During the operation of the UAV, existing external and internal factors are complex, that causes additional interference in the data received from the sensor equipment. It leads to a destabilizing effect on the functioning of the UAV control system. The data obtained from the gyroscopic sensor, which is used to calculate the angular velocities and angles of inclination of the device relative to the ordinate axis, is particularly susceptible to noise. This has a negative effect on the operation of the PID controller, which leads to the formation of incorrect control signals. The spectral characteristic of noise in the UAV control system under consideration is heterogeneous, and contains both high-amplitude noise vibrations at the resonance frequencies of the UAV body with rotary engines rotating, and uniformly distributed vibrations similar to Gaussian noise. The problem is that filtering additive interference with a nonuniform spectral density requires a whole range of solutions for filtering this signal in real-time operation of the entire system. To solve this problem, a new method for filtering noise from a gyroscopic sensor in the UAV control system has been developed, presented as a cascade of filters that removes high-frequency noise from the source signal and guarantees maximum performance due to the use of mathematical filters that effectively suppress noise with small data sample sizes (from 2 to 5 values). The proposed cascade consists of four sequentially connected elements: a low-pass filter, a notch filter, an alpha-beta Kalman filter and an exponential filter, each of which cleans the signal from a certain type of noise. The developed method also allows eliminating the influence of vibrations caused by the rotations of the rotors of the motors and the resonant vibrations of the UAV frame due to the possibility of adjusting its parameters, such as the cutoff frequency of the low-pass filter for a specific UAV control system. During experimental studies of the proposed filter cascade, the signal-to-noise ratio for the processed data reached a maximum value of 8.40 dB, which demonstrates a significant improvement over the 2 dB figure reported in the corresponding literature. The obtained indicator markers the applicability of the developed method in UAV control systems of a multirotor configuration with four rotors, in order to improve control accuracy and flight stability.