Currently we look up each robot pose from ROS before giving it to the planner. This makes things very straightforward to implement, but causes problems when the total number of robots we might have is large. Instead we should only look up robot poses that are necessary during plan grounding. Basically we need to turn this
|
robot_poses = self.get_robot_poses(self.dsg_frame) |
from a dictionary into a function.
Currently we look up each robot pose from ROS before giving it to the planner. This makes things very straightforward to implement, but causes problems when the total number of robots we might have is large. Instead we should only look up robot poses that are necessary during plan grounding. Basically we need to turn this
Omniplanner/omniplanner_ros/src/omniplanner_ros/omniplanner_node.py
Line 304 in 291fe5d