Multi-sensor data fusion and nonlinear programming-based path prediction for escaping from engagement in combat

Abstract

One of the most important factors that bring success in modern warfare is to show air superiority. Unmanned aerial vehicles (UAVs) have now become an essential component of military air operations. UAVs can be operated in two ways: by pilots from remote control stations or by flying autonomously. Under the condition of disconnection from the control station, UAVs have trouble maintaining navigation and maneuverability. By applying multisensor data fusion, an escape path prediction algorithm was developed and presented as an engagement escape method in this study. To develop the algorithm for prediction of the optimal escape route, data from various sensors are collected and processed under the influence of noise. The data from the distance and angle sensors are interpreted in the Extended Kalman Filter and estimations are made. The instant optimal escape route is created by applying the constrained optimization method on the estimations made. The main motivation of this study is developing a deterministic-based method to get the certification of it in aviation. Therefore, instead of stochastic-based learning approaches, a deterministic approach is preferred. Nonlinear programming is used as the constraint optimization method because the constraints and objective function are nonlinear. In the selected scenarios, it can be seen in the simulation results that the proposed method shows a promising result in terms of escape from engagement.