Using of the SPSA method to improve the accuracy of the UAV following a given route under the action of wind loads

Abstract

In the modern world, UAVs (unmanned aerial vehicles) are increasingly used in everyday life in solving civilian tasks. One of the main applications of UAVs is data collection with their reference to a given coordinate system. For example, for the task of aerial photography, it is necessary to accurately link each image to the global coordinate system. In addition to the exact location of coordinates, it is worth the exact movement of a given route, to collect data of exactly those places that are needed. Thus, it is very important to ensure the minimum deviation of the UAV from the given route under the conditions of external disturbances (wind disturbances) acting on it. The article describes a procedure for assessing wind speed and direction for a UAV control system using the SPSA method. The simulation results of the algorithm operation, confirmed during flight tests on an ultralight UAV with an ardupilot autopilot, are presented.

Multidisciplinary Performance Enhancement on a Fixed-wing Unmanned Aerial Vehicle via Simultaneous Morphing Wing and Control System Design

“…In the case of noisy loss function data, there are outperforming gradient-free models of SA such as Simultaneous Perturbation SA or Random-direction SA (RDSA) that come to the fore rather than finite-difference approaches to reach the global minimum value [28]. SPSA is well-known to be successfully applied in many aerospace applications such as flight path planning [20,26] and aerodynamic shape and control system designs [40,46,47]. The algorithm is exhibited to estimate the gradient of a multivariate differentiable cost function.…”