Improving of firing performance of rocket systems of helicopter by machine learning method

In modern battlefields, the capability of versatile platforms such as helicopters, which play a critical role, to accurately engage targets is of paramount importance. Traditional ballistic calculations are employed to determine the azimuth and elevation angles required for rockets to impact. However, these calculations inherently involve certain risks. Machine learning, a sub-branch of artificial intelligence, minimizes these risks by predicting the impact points of rockets launched from helicopters. In this thesis, it is aimed to predict the hit points of rockets using a dataset from the Cobra attack helicopter. 20% of the dataset was allocated as test data to compare the results with trajectory simulations and those predicted by the Random Forest Algorithm. As a result of this comparison, it was determined that the machine learning method can be used as an effective part of the decision support system in determining the hit points of the rockets. In this way, it is anticipated that this will significantly reduce the work load on weapon system operators during the decision-making process, enhance mission safety, and improve hit probability. The Random Forest Algorithm, tested five times, yielded a mean absolute error rate of approximately 6%. This study demonstrates the potential of employing machine learning algorithms to substantially enhance the firepower capabilities of helicopters in modern battlefields.