The policy does not learn in the unicycle mode

[ehsankf]

Hi,

I trained the model with the unicycle kinematic and using 20 humans as obstacles and I noticed that the policy does not learn. Is there any requirement for the number of humans with unicycle kinematic?
I believe the interval (-0.06, 0.06) for the change of angle i.e., \Delta \theta is too small for the agent to navigate this dense environment.

Thanks.

[Shuijing725]

If you have a strong computer, you can increase the number of parallel environments by changing args.num_processes and the number of training steps args.num_env_steps in arguments.py . Otherwise, you can reduce the task difficulty by fixing the starting and goal positions of robot and/or humans.

[wangbei16]

Hello, have you solved the problem that the policy dose not learn?

[Shuijing725]

Training a unicycle policy is much more stable in our new repo: https://github.com/Shuijing725/CrowdNav_HEIGHT

[wangbei16]

Training a unicycle policy is much more stable in our new repo: https://github.com/Shuijing725/CrowdNav_HEIGHT

Thank you!