salvadorvalisena's profile - activity

Hi all,

I'm using move_base for 2D navigation and it's working.the only thing is that i keep getting the a warn about having meter_scoring not set. the warn is the following:

Trajectory Rollout planner initialized with param meter_scoring not set. Set it to true to make your settings robust against changes of costmap resolution

And in my base_local_planner_params i set it to true. I'm having good and bad results from time to time, could this be the problem? what can i do to stop having this warn??

Thank you for your time,

Salvador

@Humpelstilzchen thank you for taking time to look at my question. It's resolved, I was publishing into the odom_combined frame and it has to be into the base_footprint frame.

I'm new at this so i don't know if a did it. i set all the parameters for the .yaml with http://wiki.ros.org/move_base#Parameters http://wiki.ros.org/costmap_2d#costma... http://wiki.ros.org/base_local_planne...

Yes i know that, the thing i need to know is ow to make my robot_pose_ekf to go faster

Hi all

I'm using the navigation stack in my robot and every time i send a goal using rviz (2D Nav Goal) it throws me the following:

[ WARN] [1431013577.749213787]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ WARN] [1431013578.746803719]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ WARN] [1431013579.747597482]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ WARN] [1431013580.747194408]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ WARN] [1431013581.747623215]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ WARN] [1431013582.747012998]: The goal sent to the navfn planner is off the global costmap. Planning will always fail to this goal.

[ERROR] [1431013583.745273259]: Aborting because a valid plan could not be found. Even after executing all recovery behaviors

my global costmap is the following:

global_costmap:
  global_frame: /rosphere/odom_combined  
  robot_base_frame: /rosphere/base_footprint
  transform_tolerance: 5.0
  update_frequency: 0.5
  publish_frequency: 0.0
  static_map: false
  rolling_window: true
  width: 100.0
  height: 100.0
  resolution: 0.2

and the same for the local costmap, only changing the width, height and resolution (20,20, 0.05 respectively)

The goal I'm sending is just one meter from the base on x-axis so the action should be just to move straight forward. On Rviz the grid is set to 20 plane cell count and 1meter per cell.

I've changed the resolution, width and height of both costmaps and it's always the same. Any ideas? if you need more info just ask.

Salvador

Yes right now it's inside a room so it is just sending me zero coordinates but when i use it outside (receiving real data from satellites) it have the same rate.

I'm using an EM406 gps receiver with an integrated antena. It's conected to an ardupilot mega 2.5. The porpouse of it is to use gps navigation using ros navigation stack but i need to make my filter a bit faster for a better connection of both stacks. It's installed in the top of my robot. Salvador

Hello,

I'm using move_base with a non-holonomic, non-rotational robot. Everything is working as it should but i have a big problem and it's that my robot can't make in-place rotations. It's a robot sphere as you can see in the following links:

https://www.youtube.com/watch?v=1DlGn...

https://www.youtube.com/watch?v=XrpZt...

I set the following base local planner parameters:

## Base local planner parameters. 
TrajectoryPlannerROS:
  acc_lim_x: 1.0
  acc_lim_y: 0.0
  acc_lim_theta: 0.1
  max_vel_x: 2.2
  min_vel_x: 0.55
  max_vel_theta: 0.2
  min_vel_theta: -0.2
  min_in_place_vel_theta: 0.0
  escape_vel: -0.55
  holonomic_robot: false
  yaw_goal_tolerance: 0.1
  xy_goal_tolerance: 5.0
  sim_granularity: 0.1
  angular_sim_granularity: 0.04
  vx_samples: 5
  vtheta_samples: 8
  meter_scoring: true
  pdist_scale: 0.6
  gdist_scale: 0.8
  occdist_scale: 0.01
  dwa: false
  global_frame_id: odom_combined
  oscillation_reset_dist: 0.0

Is there a way to set off the in-place rotations? The other thing is that every time the program send an in-place rotation into the /cmd_vel topic, it puts a value of 1.0 or -1.0 on the angular z as shown below

---
linear: 
  x: 0.0
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: -1.0
---

Why is it doing this if i have set the max_vel_theta to 0.2 and the min_vel_theta to -0.2?

Also, when passing short distancies to move form base_footprint it send commands to cmd_vel like the following:

---
linear: 
  x: 0.414027588531
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.894736842105
---

being this unrelated to the values set in the base_local_planner_params.yaml where min_vel_x is 0.55 and max_vel_theta is 0.2.

I've tried changing the min_in_place_vel_theta from 0.0 to 0.5 to see if it takes it as the in-place rotational value but it keep showing me the same thing. Tried setting from true to false the dwa for small values of vel_theta and got the same thing.

If anyone have an answer to all of the questions or even a single one it'd be really good!. If you need some more information just ask. Thanks for your time.

Salvador