ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

ROS frontier exploration: Robot doesn't move

asked 2016-06-13 01:40:28 -0500

bemery94 gravatar image

Hi there. I'm currently trying to use move_base, base_local_planner and frontier_exploration on a turtlebot 2 to autonomously explore a room. I currently am using a static map (from hector mapping) for the global costmap and building an obstacle layer in my local costmap as I go with my horizontal hokuyo. The config files that I am using are based off the husky_navigation config files.

These are my parameter files for the costmaps (I can add the rest of my config files for the planners etc if they're needed):

costmap_common.yaml

#footprint: [[-0.5, -0.33], [-0.5, 0.33], [0.5, 0.33], [0.5, -0.33]]
robot_radius: 0.25
footprint_padding: 0.01

robot_base_frame: base_footprint
update_frequency: 4.0
publish_frequency: 3.0
transform_tolerance: 0.5

resolution: 0.05

obstacle_range: 5.5
raytrace_range: 6.0

#layer definitions
static:
    map_topic: /map
    subscribe_to_updates: true

obstacles_laser:
    observation_sources: laser
    laser: {data_type: LaserScan, sensor_frame: /laser_lsl, clearing: true, marking: true,
    topic: /scan_lsl, inf_is_valid: true}

inflation:
    inflation_radius: 1.0

costmap_local.yaml

global_frame: odom
rolling_window: true
update_frequency: 5.0
publish_frequency: 2.0
static_map: false
rolling_window: true
width: 20.0
height: 20.0
resolution: 0.05
transform_tolerance: 0.5

plugins:
  - {name: obstacles_laser,           type: "costmap_2d::ObstacleLayer"}
  - {name: inflation,                     type: "costmap_2d::InflationLayer"}

costmap_global.yaml

global_frame: map
static_map: true
rolling_window: false
track_unknown_space: true
update_frequency: 1.0
publish_frequency: 0.5
transform_tolerance: 0.5

plugins:
  - {name: static,                  type: "costmap_2d::StaticLayer"}
  - {name: inflation,               type: "costmap_2d::InflationLayer"}

I am able to get the correct robot pose using hector mapping and can also get the correct static map, global costmap and local costmap. Images here

The issue is that when I try and send the robot movement commands using either a 2D nav goal or set a region to be explored, the robot doesn't move as it thinks it's stuck. It prints:

[ WARN] [1465787788.724254866]: Clearing costmap to unstuck robot (3.000000m).
[ WARN] [1465787799.324231262]: Rotate recovery behavior started.

I've run the program in debugging mode and this is what it prints for the various nodes that are running (note. I've deleted some of the repeated lines to keep the output more readable):

Local planner

[DEBUG] [1465797982.956315651]: Path/Goal distance computed 
[DEBUG] [1465797982.958396979]: Trajectories created 
[DEBUG] [1465797982.987519030]: Cost PointCloud published 
[DEBUG] [1465797982.987591393]: A valid velocity command of (0.10, 0.00, -0.22) was found for this cycle. 
[DEBUG] [1465797983.141464666]: Nearest waypoint to <0.000000, 0.000000> is <-0.025508, -0.019038>

costmap_2d

[DEBUG] [1465798020.176874025]: Updating area x: [293, 485] y: [174, 365] 
[DEBUG] [1465798020.180376176]: Map update time: 0.004092932 
[DEBUG] [1465798020.203780619]: MessageFilter [target=odom laser_lsl ]: Message ready in frame laser_lsl at time 1465798020.106, count now 3 
[DEBUG] [1465798020.203981715]: MessageFilter [target=odom laser_lsl ]: Message ready in frame laser_lsl at time 1465798020.131, count now 2 
[DEBUG] [1465798020.215840402]: MessageFilter [target=odom laser_lsl ]: Added message in frame laser_lsl at time 1465798020.181, count now 2 
[DEBUG] [1465798020.217410425]: Received a 1024 X 1024 map at 0.050000 m/pix 
[DEBUG] [1465798020.224168259]: MessageFilter [target ...
(more)
edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
1

answered 2016-06-13 03:07:59 -0500

bemery94 gravatar image

Turns out I made a really dumb mistake. I didn't remap the cmd_vel topic that is published from move_base to mobile_base/commands/velocity which is subscribed to by the mobile base nodelet. Once I did this it started working.

edit flag offensive delete link more

Question Tools

2 followers

Stats

Asked: 2016-06-13 01:40:28 -0500

Seen: 1,268 times

Last updated: Jun 13 '16