search

mit-acl / cadrl_ros

ROS package for dynamic obstacle avoidance for ground robots trained with deep RL
573 stars 157 forks source link

Is ford_msgs an open package on ROS? #1

Closed yzxky closed 6 years ago

yzxky commented 6 years ago

I cannot find ford_msgs on ROS website, is it a customized package? Where can I get that package? Thx.

mfe7 commented 6 years ago

good point. that's actually a custom pkg i haven't uploaded to github yet.

i suggest you look into the cadrl_node.py file to see where those custom msgs are being used (i think clusters, plannermode, safeactions) and see if there is a way to modify the python code to handle the input your system current uses. for example, we have our laserscan clusters in a custom message format, but any existing message format you have to define other agents' positions, velocities etc could be fit into this sample code without too much trouble. that is probably much easier than trying to modify the rest of your system to use our custom message formats to work with this script.

i will try to remove those dependencies in the next few weeks

yzxky commented 6 years ago

Thx. I am Kaiyuan Xu from CFINS, Tsinghua University, China. I have read your paper and think you have done a great job. Our lab is very interested in the topic you have introduced and are trying to do some extended research related to yours. We can get in touch and push forward the research of robot control in people-rich environment.

mfe7 commented 6 years ago

Sorry for the delay in response. Please find the ford_msgs repo at this link and use the dev branch. Let me know if that does not work and I will re-open the issue.

mudassirej commented 5 years ago

Hi Mfe7. i also read your paper and i am trying to continue your work. i want to know which simulation software you used for this project?

mfe7 commented 5 years ago

Thank you for the interest in the work. We wrote our own simulator for the agent dynamics, hoping to release it at some point but no timeline yet. This repo just contains the pre-trained network for inference, not the training code.

On Mon, Apr 1, 2019 at 6:47 AM mudassirej notifications@github.com wrote:

Hi Mfe7. i also read your paper and i am trying to continue your work. i want to know which simulation software you used for this project?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/mfe7/cadrl_ros/issues/1#issuecomment-478529905, or mute the thread https://github.com/notifications/unsubscribe-auth/ADWgpZndGvSt93arelMjkQO1_VhZO_tnks5vceOYgaJpZM4SmLnf .