RL_rocket

Repository for the development of my master thesis on control of launch vehicles descent and landing through reinforcement learning actors.

Separate Gym environment available.

6DOF problem

In this repository a full 6DOF rocket landing environment is developed, with realistic dynamics equation modeled on a rigid body assumption.

Continuous action space

The environment employes a continuous action space, with the engine allowed to throttle between maxThrust and minThrust. The thrust was normalized to lie in the range [-1, +1] as best practice for convergence of the algorithms suggest. The engine is gimbaled by two angles $\delta_y$ and $\delta_z$ around two hinge points, respectively moving the engine around the z and y axis.

Docker

To run the algorithm in a Docker container follow these steps:

1. Clone the repository

2. Build the docker image docker build -t rl_rocket_docker .

3. Get your Wandb API key from wandb.ai/authorize

4. Start the docker container passing the API key as an environmental variable (paste it in place of $YOUR_API_KEY$) docker run -e WANDB_API_KEY=$YOUR_API_KEY$ -it rl_rocket_docker