[RL-baseline] Model v3, experiment #2

[ziritrion]

Policy network is almost identical to v2, but we removed the final FCL from the main network and added a new one to both the actor and the critic to allow them to learn more.

We also tweaked the actions compared to the experiments with v2. The main difference is that a new no-move action was added to allow the network to use the car's momentum. Here is the action list for this experiment:

[0.0, 0.0, 0.0], # no action
[0.0, 0.8, 0.0], # throttle
[0.0, 0.0, 0.6], # break
[-0.9, 0.0, 0.0], # left
[-0.5, 0.0, 0.0], # left
[-0.2, 0.0, 0.0], # left
[0.9, 0.0, 0.0], # right
[0.5, 0.0, 0.0], # right
[0.2, 0.0, 0.0], # right

Around the 15k episode mark, the entropy collapsed and never recovered. Running reward for all of the following episodes was negative. Final running reward is -34, with a max reward of 512 around step 9k.

Sample video below. Similar results to experiment #1: the car goes forward but never turns.
https://user-images.githubusercontent.com/1465235/112730896-6de92780-8f34-11eb-9438-c7fe699a179c.mp4