search

mit-acl / cadrl_ros

ROS package for dynamic obstacle avoidance for ground robots trained with deep RL
556 stars 156 forks source link

How to consider map? #26

Closed CAI23sbP closed 8 months ago

CAI23sbP commented 8 months ago

How are you @mfe7? I have found "Robot designed for socially acceptable navigation", and i read it. I have a single question about diffusion map in page 48. -> How to merge diffusion map and CADRL?

I understood that like under. First using grid map for making Diffusion map, then input them into a MPC. The MPC generate trajectories which are able to avoid collision with map. A CADRL uses Trajectory, best to drive, for subgoal. In this view, the CADRL is used for tracking module?

Thank you for reading this issue.

mfe7 commented 8 months ago

the way we used diffusion map and cadrl together in that thesis (from 2017) is: build a 2d occupancy map of the environment ahead of time, then online: given a global goal coordinate, repeatedly use the diffusion map algorithm to choose a subgoal 5-10m ahead of the robot's position that should lead toward the global goal (accounting for static obstacles in the prior map), then pass the current subgoal to CADRL to select a velocity vector (considering the dynamic obstacles in the present environment).

CAI23sbP commented 8 months ago

Thank you for your reply