Clustering of Air Objects based on Path, Missile Path Prediction Using Deep Learning

For a given instance, the location of the object in air can be demonstrated by quadruples there are latitude, altitude, longitude and velocity. Trajectory is continuous stream time, latitude, altitude and longitude. The proposed work classifies the air objects based on trajectories and group them into meaningful objects such as flying objects, drones, helicopters, fighter, civilian aircraft, unmanned aerial vehicles, missiles, flying bombs, and so on. In this paper, the proposed work estimates the location of air objects by using their trajectory and these provides a guidance to the missiles to shoot the target based on its location where the missile is going to meets the object based on the trajectory. For calculating missile’s path, one of the most difficult challenges is applying the missile model to various related simulations is determining the missile path. The traditional way for this task is to use models and numerical methods, that also necessitates a large computational power. The data generated by the traditional system has been used to train the network and test the network, and the defect of the network prediction result is evaluated in this study. Recurrent neural network (RNN) and long short-term memory (LSTM) have been used to perform Clustering of Air Objects based on Path and for Missile Path Prediction. For clustering of air objects, LSTM gave r2 score of 0.997 and for missile path prediction, LSTM with 7 epochs gave an r2 score of 0.791.