LunarLander-v2 is a scenario developed by Oleg Klimov, an engineer at OpenAI, inspired by the original Atari Lunar Lander (https://github.com/olegklimov). In the implementation, you have to take your landing pod to a lunar pad that is always located at coordinates x=0 and y=0. In addition, your actual x and y position is known since their values are stored in the first two elements of the state vector, the vector that contains all the information for the reinforcement learning algorithm to decide the best action to take at a certain moment.
This renders the task accessible because you won't have to deal with fuzzy or uncertain localization of your position with respect to the objective (a common problem in robotics). ...
