Rocket Vertical Landing With Artificial İntelligence

In today's rapidly evolving technological age, artificial intelligence and machine learning offer us incredible opportunities. Using these opportunities to the fullest, we were able to develop an artificial intelligence that learns rocket launching capabilities.

The project was implemented using the Unity game engine and the Python programming language.
The learning process is based on the basic principles of Reinforcement Learning. Artificial intelligence does not know how to control the rocket towards a specific target initially. However, with experiments and steps repeated thousands of times, he improves himself and becomes a rocket launcher.

In this project, we used a neural network that processes environmental information at each step. Artificial intelligence analyzes the surrounding data and chooses the most appropriate actions for the rocket to go towards the target. In the beginning, he gains experience by making mistakes, but over time, he progresses towards perfection by minimizing these mistakes.

The learning process, completed in 480,000 steps, produces truly remarkable results. The rocket successfully reaches the target, getting better each time. This project has great potential in space exploration and rocket launch operations in the future. It can offer a more reliable and efficient approach to real-world rocket launch missions.