Ask Your Question
0

TF_NAN_INPUT errors from use of ekf_localization_node [closed]

asked 2015-06-30 03:56:05 -0600

quantumflux gravatar image

updated 2015-08-20 04:08:51 -0600

UPDATE: I have verified that simply setting odom0_differential to false also solves the problem. It is in fact probably not a problem with the ARM system I'm using, as I eventually got the same error on a x86_64 laptop that I mounted on the robot. It seems as though it works fine for a while, and then the error shows up continuously after a while. Is there maybe some sort of caching applied which keeps erronous values and is reset by changing the settings in the launch file?

ORIGINAL QUESTION:

Whenever I use ekf_localization_node in my project, I get this error:

Error:   TF_NAN_INPUT: Ignoring transform for child_frame_id "base_footprint" from authority "unknown_publisher" because of a nan value in the transform (-nan -nan -nan) (-nan -nan -nan -nan)
         at line 240 in /tmp/buildd/ros-indigo-tf2-0.5.11-0trusty-20150522-1835/src/buffer_core.cpp
I don't get it when I put all values of the (currently only) input of the node to false, but as soon as I put a 'true' at the position for (e.g.) the x input, it shows up again. I am using wheel odometry of a Kobuki for this input. Here is a sample odometry message:
header: 
  seq: 20227
  stamp: 
    secs: 1435652980
    nsecs: 60858344
  frame_id: odom
child_frame_id: base_footprint
pose: 
  pose: 
    position: 
      x: 0.0
      y: 0.0
      z: 0.0
    orientation: 
      x: 0.0
      y: 0.0
      z: 0.0
      w: 1.0
  covariance: [0.1, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.1, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.7976931348623157e+308, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.7976931348623157e+308, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.7976931348623157e+308, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.05]
twist: 
  twist: 
    linear: 
      x: 0.0
      y: 0.0
      z: 0.0
    angular: 
      x: 0.0
      y: 0.0
      z: 0.0
covariance: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
---

This is the launchfile I am currently trying to use for the ekf_localization_node: http://pastebin.com/SPrXDD53

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by tfoote
close date 2016-09-14 14:30:25.969072

Comments

Can you please post a sample odometry input message? Also, please let me know which version of robot_localization you're using. I see you're using Indigo, but did you install the package from the package manager or build it from source?

Tom Moore gravatar imageTom Moore ( 2015-06-30 08:36:54 -0600 )edit

Update: I tried to use a different bag file with your launch file, and I could not reproduce your issue. You'll have to post a sample odometry message.

Tom Moore gravatar imageTom Moore ( 2015-07-06 19:31:12 -0600 )edit

Hey, sorry for the delay. After a bit of testing I noticed that this only happens on one unit, which is the one I was testing it on first. It runs an ARM-based embedded system, so that might have something to do with it. On all x86 based platforms this issue does not exist, and in fact (cont.)

quantumflux gravatar imagequantumflux ( 2015-07-08 01:30:52 -0600 )edit

(cont.) I can run the package fine when I "outsource" it to one of our x86-based workstations instead of running it directly on the robot.

quantumflux gravatar imagequantumflux ( 2015-07-08 01:32:01 -0600 )edit

Which ARM board are you using? Did you build from scratch or use the package manager?

Tom Moore gravatar imageTom Moore ( 2015-07-08 18:51:07 -0600 )edit

Any update on this?

Tom Moore gravatar imageTom Moore ( 2015-07-29 18:29:37 -0600 )edit

Yes, sorry. Vacation and other stuff kept me from posting updates. I will update the question to reflect the changed circumstances I have encountered.

quantumflux gravatar imagequantumflux ( 2015-08-20 04:04:13 -0600 )edit

1 Answer

Sort by ยป oldest newest most voted
0

answered 2015-08-20 10:21:11 -0600

Tom Moore gravatar image

No, there's no caching going on. Is the message you posted above the output of ekf_localization_node or an input message? Can you post a bag file that shows the behavior you're describing?

edit flag offensive delete link more

Question Tools

1 follower

Stats

Asked: 2015-06-30 03:56:05 -0600

Seen: 1,574 times

Last updated: Aug 20 '15