search

reiniscimurs / DRL-robot-navigation

Deep Reinforcement Learning for mobile robot navigation in ROS Gazebo simulator. Using Twin Delayed Deep Deterministic Policy Gradient (TD3) neural network, a robot learns to navigate to a random goal point in a simulated environment while avoiding obstacles.
MIT License
488 stars 98 forks source link

running error #23

Closed namjiwon1023 closed 1 year ago

namjiwon1023 commented 1 year ago

Thank you for your contribution!!

When I run the command 'python3 velodyne_td3.py', agent's training episode steps is not displayed.

And the following error will appear.

image

How can I solve this problem, waiting for your reply.

reiniscimurs commented 1 year ago

Hi,

Can you show the full output from the terminal? Which ROS version are you using?

namjiwon1023 commented 1 year ago

Hi, thanks for replying so quickly！

OS : Mint 19 (Ubuntu 18.04) ROS version : Melodic

image image

After a period of time, information such as episode, mean_reward can be displayed. But I don't know what the impact of the "Spawn service failed" error will be.

reiniscimurs commented 1 year ago

So as it appears to me the rviz is trying to spawn the robot model but timing out and essentially ignoring the robot model in the end. You can try increasing the time that neural network waits for the environment to start here: https://github.com/reiniscimurs/DRL-robot-navigation/blob/d51438aea0dd8404ae7a633dcfc4f5f96e54f569/TD3/velodyne_td3.py#L210

See if that helps. What is the output from the rviz screen?

Also, it seems that it is training just fine, just the robot model is not getting loaded in rviz. You could check this by opening the gazebo simulator during run time. Start the training, then open a separate terminal window and run command gzclient This should open gazebo and robot motion should be visible there.

namjiwon1023 commented 1 year ago

I didn't make any changes.

Now the agent has been trained for 20 hours, The agent has learned a policy to find the target.

image

rviz screen output:

image

I will continue my research in your direction.

I'm very much obliged to you.