I have being using cob_collision_velocity_filter along with cob_linear_nav for controlling holonomic robot other than care-o-bot family robot (Since these are generic enough for use in other robots as well)
But I couldn't set something like odom or map frame as the global frame for local costmap and collision_velocity_filter.
Following are my settings (only relavent fields are shown)
collision_velocity_filter_params.yaml
#where to look for costmap parameters
costmap_parameter_source: "/collision_velocity_filter/anti_collision_costmap/costmap" #default "/local_costmap_node/costmap"
#Parameters specifying collision velocity filter
global_frame: /odom
robot_base_frame: /base_footprint
local_costmap_params.yaml
# global information
global_frame: /odom
robot_base_frame: /base_footprint
update_frequency: 4.0
publish_frequency: 4.0
# local map settings
rolling_window: true
width: 10.0
height: 10.0
resolution: 0.2
This will cause collision_velocity_filter to only correctly recognize relevent obstacles when robot is at 0,0 only. So assume cost grid of cost map is not correctly transformed from map frame to robot frame and compared.
The reason I'm trying to use odom frame as costmap global frame is because when robot link is used for costmap it rotates along with robot and smears the obstacles around our robot (Since it does have very limited FoV of LIDAR) - May be we should simply try to resolve that rather, but I believe above problem in collision_velocity_filter also a genuine limitation
Please let me know if any additional information is required.
Hi,
I have being using cob_collision_velocity_filter along with cob_linear_nav for controlling holonomic robot other than care-o-bot family robot (Since these are generic enough for use in other robots as well)
But I couldn't set something like
odom
ormap
frame as the global frame for local costmap and collision_velocity_filter.Following are my settings (only relavent fields are shown)
collision_velocity_filter_params.yaml
local_costmap_params.yaml
This will cause collision_velocity_filter to only correctly recognize relevent obstacles when robot is at 0,0 only. So assume cost grid of cost map is not correctly transformed from map frame to robot frame and compared.
The reason I'm trying to use
odom
frame as costmap global frame is because when robot link is used for costmap it rotates along with robot and smears the obstacles around our robot (Since it does have very limited FoV of LIDAR) - May be we should simply try to resolve that rather, but I believe above problem in collision_velocity_filter also a genuine limitationPlease let me know if any additional information is required.