RVFLNN UAV FLIGHT TRAJECTORY PREDICTION ALGORITHM BASED ON BASIC MOTION MODEL

Abstract

During the flight, the drone is identified through the monitoring platform and monitored in real time. However, the existing monitoring platform is limited to monitoring collaborative drones. If the drone data collected by the monitoring platform can be used to forecast it in real time, it can better adapt to the air traffic of drones. management requirements. Considering that the prediction accuracy of RVFLNN decreases sharply with the prediction step size, the Kalman filter theory is used to combine RVFLNN with the basic motion model to implement the UAV trajectory prediction method based on RVFLNN. By matching the motion pattern with the flight pattern of the UAV at the next expected time, RVFLNN showed better results in multi-stage forecasting compared with the traditional joint prediction method of RVFLNN.