Please use Python

Scenario

We are on the way to setting foot on Mars. To do so$,$ we need to refuel our space capsule in a space station thousands of miles away from Earth.

We need to perform a docking operation similarly as shown in this video.

Goal

Build a reinforcement learning model to perform the docking operation autonomously while respecting safety and time constraints.

Time constraint

The docking operation must be done in $5$ minutes $(300$ s$)$

Safety constraint

For the docking operation to be performed safely, the space capsule needs to be aligned $+/-15$ cm from the docking center of the space station.

Actions

You only have $3$ actions: go left, do nothing, go right.

Observations

Space capsule alignment with the space station. It is assumed the alignment can range from $-2$ m to $2$ m$,$ where $0$ m means perfect alignment. We also assume that negative values describe a misalignment to the left and positive values represent a misalignment to the right.

Tasks

Create a new $.$py file called SpaceCapsuleEnvironment.py and build a model that keeps the space capsule in the safety range for as long as possible using SARSA algorithm.

$1)[15$ points$]$ Create a SpaceEnv$()$ class with the following function

$[5$ points$]\_\_$init$\_\_($self$),$ an initialization function that defines the initial values for the environment.

$[5$ points$]$ step$($self$,$ action$),$ a step function that provides the next observation $($state$)$ of the system, a reward, and stopping criteria $($done$)$ given an action.

$[5$ points$]$ reset$($self$),$ a function to reset the values of the environment after each episode.

$2)[5$ points$]$ Discretize the alignment range in $100$ evenly spaced numbers over the interval $-200$ cm and $200$ cm$.$ You should use the function np$.$linspace. Also, update the function getState$($observation$).$

$3)[10$ points$]$ Run SARSA algorithm with the SpaceEnv$()$ you've just created. You should save the plot containing the results.

$4)[10$ points$]$ Change the learning rate parameter ALPHA to $0\mathrm{.}2$ and then to $0\mathrm{.}5,$ and discuss your findings.