Make Ant environment

[Discordius]

Should be based on this: https://github.com/openai/gym/blob/master/gym/envs/mujoco/ant.py

I think we probably just need to change the reward function part of that code. So either just copying that, or overwriting the step function should work.