Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

PixHawk timestamp instead of ROS time #185

Closed
thedinuka opened this issue Jan 7, 2015 · 13 comments
Closed

PixHawk timestamp instead of ROS time #185

thedinuka opened this issue Jan 7, 2015 · 13 comments
Assignees
@mhkabir
Labels
enhancement PX4
Milestone
Release 0.10

Comments

@thedinuka
Copy link

I was looking at imu_pub.cpp and realized that imu/data_raw topic is published with the header timestamp set to ROS::Time::now()
(https://github.com/mavlink/mavros/blob/master/mavros/src/plugins/imu_pub.cpp#L250)

The MAVLINK_MSG_ID_HIGHRES_IMU message from PixHawk does contain a time_usec filed which actually represent the time at which the imu data was captured. For certain applications like for synchronizing a camera and IMU, it maybe useful to have that as the timestamp instead of the ros time.

I'm wondering whether this would qualify as a new feature request? I could do a PR for this, but the change in the code is really simple anyway.
@mhkabir
Copy link
Member

mhkabir commented Jan 7, 2015

Hi,
We have timesync working. I will be updating the plugins soon to take advantage of this. We synchronise Pixhawk timestamps to ROS time and vice-versa.
What you propose would not be ideal, as ROS wall time is calculated from Unix Epoch and Pixhawk time is calculated from boot.

@mhkabir mhkabir self-assigned this Jan 7, 2015
@thedinuka
Copy link
Author

@mhkabir timesync would be ideal, I agree. What sort of accuracies can you achieve with it?

@mhkabir
Copy link
Member

mhkabir commented Jan 7, 2015

We get microsecond precision. Most message timestamps are milliseconds since boot.

@thedinuka
Copy link
Author

ahh, excellent. would you be able to comment on when this will be available?

@mhkabir
Copy link
Member

mhkabir commented Jan 7, 2015

Pretty soon hopefully. Synchronisation works already (check system_time.cpp ) The plugins just need to be updated.
I don't have much time for this now though. Will try to do it soon.

@thedinuka
Copy link
Author

@mhkabir I'm still reading through system_time.cpp (actually it's sys_time.cpp right?) but as a preliminary question, is the idea to maintain the difference between ros time and PixHawk time and then use that offset in other plugins to set the correct timestamp?

@mhkabir
Copy link
Member

mhkabir commented Jan 7, 2015

Yes

On Wed, Jan 7, 2015 at 1:40 PM, Dinuka notifications@github.com wrote:

@mhkabir https://github.com/mhkabir I'm still reading through
system_time.cpp (actually it's sys_time.cpp right?) but as a preliminary
question, is the idea to maintain the difference between ros time and
PixHawk time and then use that offset in other plugins to set the correct
timestamp?

—
Reply to this email directly or view it on GitHub
#185 (comment).

@vooon vooon added this to the Release 0.10 milestone Jan 7, 2015
@mhkabir mhkabir mentioned this issue Jan 11, 2015
Closed
@vooon
Copy link
Member

vooon commented Jan 16, 2015

I testing PR #186.

Setup:

  • Wandboard Quad with Ubuntu 14.04 and latest ROS for UbuntuARM
  • Pixhawk with latest master

Results:
First run without custom -m onboard mode.
Got timsync, but strange timestamps:

/mavros/imu/data

---
header: 
  seq: 4316
  stamp: 
    secs: 4294967295
    nsecs: 4294967295
  frame_id: fcu
orientation: 
  x: -0.0430093572484
  y: 0.00982332692738
  z: -0.307822874007
  w: 0.950420315269
orientation_covariance: [1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0]
angular_velocity: 
  x: 0.0105894114822
  y: -0.0485693067312
  z: -0.010234859772
angular_velocity_covariance: [1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07]
linear_acceleration: 
  x: 0.0641356706619
  y: -0.864281475544
  z: 10.0234670639
linear_acceleration_covariance: [8.999999999999999e-08, 0.0, 0.0, 0.0, 8.999999999999999e-08, 0.0, 0.0, 0.0, 8.999999999999999e-08]
---
header: 
  seq: 4317
  stamp: 
    secs: 4294967295
    nsecs: 4294967295
  frame_id: fcu
orientation: 
  x: -0.0430188834137
  y: 0.00986130963341
  z: -0.307839425409
  w: 0.950414129948
orientation_covariance: [1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0]
angular_velocity: 
  x: 0.00931540410966
  y: -0.0476437397301
  z: -0.00976792909205
angular_velocity_covariance: [1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07]
linear_acceleration: 
  x: 0.0641356706619
  y: -0.864281475544
  z: 10.0234670639
linear_acceleration_covariance: [8.999999999999999e-08, 0.0, 0.0, 0.0, 8.999999999999999e-08, 0.0, 0.0, 0.0, 8.999999999999999e-08]
---

/diagnostics

  - 
    level: 0
    name: mavros: Time Sync
    message: Normal
    hardware_id: Mavlink
    values: 
      - 
        key: Events in window
        value: 100
      - 
        key: Events since startup
        value: 1361
      - 
        key: Duration of window (s)
        value: 10.007385
      - 
        key: Actual frequency (Hz)
        value: 9.992621
      - 
        key: Last dt (ms)
        value: 0.550172
      - 
        key: Mean dt (ms)
        value: -0.000695
      - 
        key: Last system time (s)
        value: 204.922809000
      - 
        key: Time offset (s)
        value: 1421389669.363838673

@vooon
Copy link
Member

vooon commented Jan 16, 2015

I placed this to /fs/microsd/etc/extras.txt:

#!nsh
#
# USB MAVLink start
#

mavlink start -r 20000 -d /dev/ttyACM0 -x -m onboard
# Enable a number of interesting streams we want via USB
mavlink stream -d /dev/ttyACM0 -s PARAM_VALUE -r 200
usleep 100000
mavlink stream -d /dev/ttyACM0 -s MISSION_ITEM -r 50
usleep 100000
mavlink stream -d /dev/ttyACM0 -s NAMED_VALUE_FLOAT -r 10
usleep 100000
mavlink stream -d /dev/ttyACM0 -s OPTICAL_FLOW -r 10
usleep 100000
mavlink stream -d /dev/ttyACM0 -s VFR_HUD -r 20
usleep 100000
mavlink stream -d /dev/ttyACM0 -s ATTITUDE -r 20
usleep 100000
mavlink stream -d /dev/ttyACM0 -s ATTITUDE_CONTROLS -r 30
usleep 100000
mavlink stream -d /dev/ttyACM0 -s RC_CHANNELS_RAW -r 5
usleep 100000
mavlink stream -d /dev/ttyACM0 -s SERVO_OUTPUT_RAW_0 -r 20
usleep 100000
mavlink stream -d /dev/ttyACM0 -s POSITION_TARGET_GLOBAL_INT -r 10
usleep 100000
mavlink stream -d /dev/ttyACM0 -s LOCAL_POSITION_NED -r 30
usleep 100000
mavlink stream -d /dev/ttyACM0 -s MANUAL_CONTROL -r 5

# Exit shell to make it available to MAVLink
exit

And checked that mavlink starts from fcu bootup (px4 quirk not needed, but enabled).
But get similar resuilt: 4294967295 constant in timestamp

Also strange operation errors:

[ INFO] [1421390635.313224536]: MAVROS started. MY ID [1, 240], TARGET ID [1, 50]
[ WARN] [1421390635.318253761]: TM: Clock skew detected (-1421390607.381244898 s). Hard syncing clocks.
[ INFO] [1421390635.388229088]: CON: Got HEARTBEAT, connected.
[ WARN] [1421390646.390623651]: PR: request list timeout, retries left 2
[ WARN] [1421390647.392587524]: PR: request list timeout, retries left 1
[ WARN] [1421390648.393595986]: PR: request list timeout, retries left 0
[ WARN] [1421390651.391841876]: WP: timeout, retries left 2
[ INFO] [1421390651.907876513]: WP: seems GCS requesting mission
[ INFO] [1421390651.909298605]: WP: sheduling pull after GCS is done
[ WARN] [1421390652.401791065]: WP: timeout, retries left 2
[ INFO] [1421390652.414816520]: WP: seems GCS requesting mission
[ INFO] [1421390652.415519210]: WP: sheduling pull after GCS is done
[ INFO] [1421390652.936081782]: WP: seems GCS requesting mission
[ INFO] [1421390652.936835230]: WP: sheduling pull after GCS is done
[ WARN] [1421390653.404124672]: WP: timeout, retries left 1
[ INFO] [1421390653.451881471]: WP: seems GCS requesting mission
[ INFO] [1421390653.452306833]: WP: sheduling pull after GCS is done
[ INFO] [1421390653.973946167]: WP: seems GCS requesting mission
[ INFO] [1421390653.974379332]: WP: sheduling pull after GCS is done
[ WARN] [1421390654.405906896]: WP: timeout, retries left 0
[ INFO] [1421390654.479033400]: WP: seems GCS requesting mission
[ INFO] [1421390654.479452293]: WP: sheduling pull after GCS is done
[ INFO] [1421390654.995982050]: WP: seems GCS requesting mission
[ INFO] [1421390654.997416385]: WP: sheduling pull after GCS is done
[ERROR] [1421390655.409012207]: WP: timed out.
[ INFO] [1421390655.499020671]: WP: seems GCS requesting mission
[ INFO] [1421390655.499445457]: WP: sheduling pull after GCS is done
[ INFO] [1421390656.006200442]: WP: seems GCS requesting mission
[ INFO] [1421390656.006822432]: WP: sheduling pull after GCS is done
[ERROR] [1421390656.452320919]: FCU: Operation timeout
[ WARN] [1421390662.010983050]: WP: timeout, retries left 2
[ WARN] [1421390663.012967000]: WP: timeout, retries left 1
[ WARN] [1421390664.015000102]: WP: timeout, retries left 0
[ERROR] [1421390665.017205422]: WP: timed out.

@vooon
Copy link
Member

vooon commented Jan 16, 2015

Bug in syncronise_stamp(). I will fix it.

vooon added a commit that referenced this issue Jan 16, 2015
@vooon
uas: Fix ros timestamp calculation.
d0b3ed4 
Issues: #186, #185.
vooon added a commit that referenced this issue Jan 16, 2015
@vooon
plugin: gps: #185 use synchronized timestamp
49763b6 
common.xml tells that GPS_RAW_INT have time_usec stamps.
vooon added a commit that referenced this issue Jan 16, 2015
@vooon
plugin: rc_io: #185 Use synchronized timestamp.
03e590d
@vooon
Copy link
Member

vooon commented Jan 16, 2015

@mhkabir pull request closed, your patch rebased in master.

Time is close to ros wall time in range ±5 ms by my quick test.

@mhkabir
Copy link
Member

mhkabir commented Jan 16, 2015

Awesome.
On 16 Jan 2015 15:17, "Vladimir Ermakov" notifications@github.com wrote:

@mhkabir https://github.com/mhkabir pull request closed, your patch
rebased in master.

Time is close to ros wall time in range ±5 ms by my quick test.

—
Reply to this email directly or view it on GitHub
#185 (comment).

vooon added a commit that referenced this issue Jan 20, 2015
@vooon
Merge branch 'master' into coverity_scan
03af3b3 
* master: (52 commits)
  travis: update coverity scan
  plugin: sts_time: Code cleanup and codestyle fix.
  plugin: command: Quirk for older FCU's (component_id)
  plugin: rc_io: #185 Use synchronized timestamp.
  plugin: gps: #185 use synchronized timestamp
  uas: Fix ros timestamp calculation.
  plugin: add synchronisation to most plugins (fixed)
  readme: Add notes about coordinate frame conversions #49
  travis: add gitter notification
  Added Gitter badge
  mocap_pose_estimate: Switched from pose to poseStamped.
  0.9.4
  Prepare correction release 0.9.4
  plugin: sys_time: enable EMA
  0.9.3
  Prepare minor release 0.9.3 #182
  Initiliser fix
  plugin: visualisation - Fixes CI build
  plugin: visualisation
  plugin: visualization minor patch
  ...
@vooon
Copy link
Member

vooon commented Jan 24, 2015

I think we may close this.

@vooon vooon closed this as completed Jan 24, 2015
@fnoop fnoop mentioned this issue Feb 20, 2017
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@mhkabir mhkabir

Labels
enhancement PX4
Projects
None yet
Milestone
Release 0.10
Development

No branches or pull requests
3 participants
@vooon @thedinuka @mhkabir