Didn't see moving obstacle, am I run it in the wrong way?

anassinator / dqn-obstacle-avoidance

Deep Reinforcement Learning for Fixed-Wing Flight Control with Deep Q-Network

MIT License

74 stars 20 forks source link

Didn't see moving obstacle, am I run it in the wrong way? #2

Open mushixuan opened 6 years ago

mushixuan commented 6 years ago

thank you for your inspiring project I have "director simulator.py" successfully and it looks like the picture below.

all the obstacles are stationary. I think it's an moving obstacle simulation right? Do I do something wrong?

and when I “director simulator.py”, error occurs like this:

how could I fix this?

anassinator commented 6 years ago

Hi @mushixuan

To enable moving obstacles you just need to run:

director simulator.py --moving-obstacle-ratio <RATIO_OF_MOVING_OBSTACLES>

Just replace <RATIO_OF_MOVING_OBSTACLES> with a number between 0 and 1. This controls how many of the obstacles are moving vs stationary.

You might also find some other command-line arguments useful. You can see them all as follows:

director simulator.py --help

Hope this helps.

zhaoxinyizhf commented 6 years ago

I met the same question with you. DirectorPython can‘t import tensorflow while python can. I have installed the Python3.6.3, tensorflow 1.6.0. Is the problem of version? How did you solve it?

anassinator commented 6 years ago

At the moment, I believe this only works with Python 2.7

zhaoxinyizhf commented 6 years ago

I see. Thank you so much.