search

aqeelanwar / PEDRA

Programmable Engine for Drone Reinforcement Learning Applications
MIT License
269 stars 60 forks source link

Error while running simulation in training mode --> msgpackrpc.error #106

Closed rakehsaleem closed 2 years ago

rakehsaleem commented 2 years ago

Hi @aqeelanwar, I'm trying to use this library to train and test the basic functionality of PEDRA but I'm running into some issues. At the moment, my config settings are like this:

[general_params]
run_name:                 Tello_indoor
env_type:                 outdoor
env_name:                 utdoor_oldtown
ip_address:               127.0.0.5
algorithm:                DeepQLearning
mode:                     train

[drone_params]
SimMode:                  ComputerVision
num_agents:               1
drone:                    DJIPhantom
ClockSpeed:               20

[camera_params]
width:                    320
height:                   180
fov_degrees:              80

and I'm trying to train, while I run the command python main.py, the simulation starts but right after that I see the following error:

msgpackrpc.error.RPCError: rpclib: client error C0002: Function 'simGetImages' was called with an invalid number of arguments. Expected: 3, got: 2

although a similar issue to this was posted to AirSim repository and I check that and tried to make changes but still, couldn't get it working. Any help on this would be much appreciated, thank you!

rakehsaleem commented 2 years ago

Post Error: I tried to add an argument '0' which I read in the AirSim repository so the function:

responses1 = client.simGetImages( 
    [
         airsim.ImageRequest('front_center', airsim.ImageType.Scene, False, False)
    ], vehicle_name=vehicle_name)  # scene vision image in uncompressed RGBA array

changed to 

responses1 = client.simGetImages(0, 
    [
         airsim.ImageRequest('front_center', airsim.ImageType.Scene, False, False)
    ], vehicle_name=vehicle_name)  # scene vision image in uncompressed RGBA array

after this, the simulation gives me the following error.

File "main.py", line 105, in eval(name) File "", line 1, in File "C:\PEDRA\algorithms\DeepQLearning.py", line 72, in DeepQLearning current_state[name_agent] = agent[name_agent].get_state() File "C:\PEDRA\network\agent.py", line 129, in get_state camera_image = get_MonocularImageRGB(self.client, self.vehicle_name) File "C:\PEDRA\aux_functions.py", line 404, in get_MonocularImageRGB ], vehicle_name=vehicle_name) # scene vision image in uncompressed RGBA array TypeError: simGetImages() got multiple values for argument 'vehicle_name'

I'm trying to debug the code and fix the problem but I'll update if I get success and if somebody figured it out, please drop the suggestion here. Thanks again!

rakehsaleem commented 2 years ago

@aqeelanwar Still struggling with this error, and can not get it working. Any help or potential lead would be much appreciated. Thank you!

xiaoshengjianan commented 1 year ago

Can you tell me how to solve this problem?