AN ANN-BASED INTEGRATED MODEL FOR AUTONOMOUS UAV FLIGHT CONTROL CONSIDERING EXTERNAL FORCES

This study presents an artificial neural network (ANN)-based integrated model designed to tackle the challenges of autonomous flight control in unmanned aerial vehicles (UAVs), with a particular focus on external forces such as wind speed. The proposed model offers multiple contributions to the field, including a reduction in UAV operation costs, simplified UAV control model establishment, and the ability to handle uncertainties and nonlinearities in different system environments. The model achieves high prediction accuracy (R2 2 0.9710 and 0.9480) for UAV acceleration and path prediction, making it suitable for various UAVs including aviation systems. A dual-model approach is introduced, with Model 1 predicting the path with acceleration and wind speed, and Model 2 predicting the acceleration of the UAV with path and wind speed. This comprehensive approach enhances the autonomous flight control process. The proposed model enables the prediction of future UAV paths and stable control using established autonomous flight mechanisms even when following a new path. Although the study focuses on wind speed as the primary external force, there is potential for further improvement by incorporating additional external forces and data sources, such as gyro sensors, temperature, barometric pressure, and image data. In conclusion, the proposed model provides a valuable contribution to the field of autonomous UAV control, and future work can include refining the model with other external forces and data sources to enhance its accuracy and reliability in various environments.