Synthesis of a measurement procedure for estimating the orientation of a small unmanned aerial vehicle under changes in the status of measurement results

The article considers the problem of processing measurement data under the changing status of measurement results from micromechanical sensors used in the intelligent onboard measurement system of a small unmanned aerial vehicle. While an aircraft is in the air, the status of measurement results changes from confirmed to indicative (for example, due to sensor defects, measuring channel degradation, false output measurement signals from micromechanical sensors due to vibrations and shocks caused by movement of air masses). As a result, the probability of stability loss in a small unmanned aerial vehicle increases, requiring improved accuracy of its orientation estimation under changes in the measurement result status. In the study, a Kalman measurement procedure was synthesized, with its equations defined with accuracy to the parameters of transition and state noise matrices characterizing the measurement process, as well as the disturbance vector. The parameters were determined by the mathematical model of measurement data transformation relying on a dynamic mathematical model, which distinguishes the developed measurement procedure from that with the classical transition matrix. In order to find unknown parameters, a neural network was applied. A multilayer perceptron was chosen as the basis for the neural network, which was trained using the error backpropagation algorithm. The mathematical modeling and measurement experiment revealed that the accuracy of the measurement procedure synthesized using a dynamic mathematical model is higher than that of the measurement procedure using the Kalman filter with a classical transition matrix. The study results can be used in the development of intelligent measurement procedures for onboard measurement systems operating under changes in the measurement result status.