Velodyne 32E IMU support

450 views
Skip to first unread message

Ioannis Havoutis

unread,
Jul 9, 2012, 6:16:41 AM7/9/12
to utexas-art-ros-pkg-users
Hi all,
I have been playing with the velodyne stack with a 32E model over the
past month.
I didn't have any problems setting it up or getting data out of the
driver and pointcloud nodes.

I would like to access the IMU data coming out of the on-board
sensor.
I have seen that in the review this is listed as future work - is
anyone working on it or has anyone worked on this before?

cheers,
Ioannis


ps; a big thanks to all that have contributed to this stack!

Jack O'Quin

unread,
Jul 9, 2012, 10:21:18 AM7/9/12
to utexas-art-r...@googlegroups.com
On Mon, Jul 9, 2012 at 5:16 AM, Ioannis Havoutis
<i.hav...@googlemail.com> wrote:
> Hi all,
> I have been playing with the velodyne stack with a 32E model over the
> past month.
> I didn't have any problems setting it up or getting data out of the
> driver and pointcloud nodes.

Glad it's working for you.

> I would like to access the IMU data coming out of the on-board
> sensor.
>
> I have seen that in the review this is listed as future work - is
> anyone working on it or has anyone worked on this before?

Not that I know about. I wasn't planning to work on new features until
someone asked for them. Now you have.

I don't see IMU mentioned in the 32E manual from Velodyne. Do you mean the GPS?

The 32E provides GPS packets on a separate port, 8303. I think that
allows us to write a separate ROS driver node to handle those packets
and publish the appropriate sensor_msgs/NavSatFix and
sensor_msgs/TimeReference messages. Since the data are in NMEA format,
it may be possible to adapt one of the existing ROS GPS drivers for
that purpose. There is a collection of them here:

https://kforge.ros.org/projects/gpsdrivers/

We don't have access to that device here. There is probably enough
information in the device manual for writing code that should
more-or-less work. But, I can only test it on PCAP data. I did the 32E
support that way with people providing test feedback, so it is a
workable approach, although it takes a little longer.

Another solution would be for you or someone else with the correct
device to provide a patch. That was done for the 64E S1 model, which
worked fine.

> ps; a big thanks to all that have contributed to this stack!

We appreciate your kind words.
--
joq

Ioannis Havoutis

unread,
Jul 12, 2012, 6:19:09 AM7/12/12
to utexas-art-ros-pkg-users
Hi Jack,
on port 8303 there is the positioning packet that outputs the rotation
and acceleration of the sensor base (gyro and accelerometter values).
The packet begins with gyro/tempature/accel, while the GPS timestamp
and NMEA sentence follows.
What I would like to do is use this information to correct for the
base attitude, in essence to rotate the readings to a 'global'
reference frame.

A separate ROS node publishing this data would be ideal - I would be
happy to provide PCAP data and test feedback.
I can also help in coding though I'm quite new to ROS.

Best,
ioannis

Jack O'Quin

unread,
Jul 12, 2012, 8:54:03 AM7/12/12
to utexas-art-r...@googlegroups.com
That sounds workable.

There is already a nice NMEA GPS driver written in Python by Eric
Perko. I suspect the port 8303 packet rate is slow enough we could
adapt that to some code that unpacks the Velodyne packet format.

I hope we can find a good way to calculate the covariance matrices in
the Imu message.
--
joq

Thomas Denewiler

unread,
Sep 12, 2012, 9:18:01 PM9/12/12
to utexas-art-r...@googlegroups.com
Is there any work being done for the IMU and GPS data for the 32E ROS driver? I can provide you with sample PCAP data as well if it would help.

I have been using this driver for both the 64E and 32E models for the past month and have had zero problems with them, either live or as bag files. Thanks for a well written driver!

Jack O'Quin

unread,
Sep 12, 2012, 9:53:30 PM9/12/12
to utexas-art-r...@googlegroups.com
On Wed, Sep 12, 2012 at 8:18 PM, Thomas Denewiler <tdene...@gmail.com> wrote:
> Is there any work being done for the IMU and GPS data for the 32E ROS
> driver? I can provide you with sample PCAP data as well if it would help.

I have not done any work on it. With some PCAP data I could probably
do something useful (I have received none so far).

I think it can and should be implemented as a separate driver,
probably based on Eric Perko's Python NMEA driver implementation.

http://ros.org/wiki/nmea_gps_driver

If so, there is no particular urgency for doing it as part of the
velodyne stack, although it might be more convenient for users if we
package everything together. I would be happy to do that.

> I have been using this driver for both the 64E and 32E models for the past
> month and have had zero problems with them, either live or as bag files.
> Thanks for a well written driver!

Thanks! Glad you have found it useful.
--
joq

Ioannis Havoutis

unread,
Sep 13, 2012, 5:03:04 AM9/13/12
to utexas-art-r...@googlegroups.com
Hi Jack and Thomas,
here is a link to a pcap file (~50mb) of about minute of scans with the 32E.
http://dl.dropbox.com/u/14302656/pcap-desk.tar.gz

This was a static set-up so there will be little to none roll/pitch/yaw
motion unfortunately.

I can capture some more data after next Monday (icra deadline!).
best,
ioannis



ps. Apologies for the delay - we finally used a separate imu for some
preliminary testing.

Jack O'Quin

unread,
Sep 13, 2012, 9:20:30 AM9/13/12
to utexas-art-r...@googlegroups.com
On Thu, Sep 13, 2012 at 4:03 AM, Ioannis Havoutis
<i.hav...@googlemail.com> wrote:
> Hi Jack and Thomas,
> here is a link to a pcap file (~50mb) of about minute of scans with the 32E.
> http://dl.dropbox.com/u/14302656/pcap-desk.tar.gz

Thanks!

> This was a static set-up so there will be little to none roll/pitch/yaw
> motion unfortunately.
>
> I can capture some more data after next Monday (icra deadline!).

Motion would be helpful.

> ps. Apologies for the delay - we finally used a separate imu for some preliminary testing.

Not a problem. Other tasks kept me busy lately. :-)
--
joq

Zdeněk Materna

unread,
Dec 16, 2013, 6:14:42 AM12/16/13
to utexas-art-r...@googlegroups.com, i.hav...@googlemail.com
Hi all,

can someone provide new link to PCAP containing packets on port 8303? The posted link doesn't work anymore. Many thanks in advance.

Best,
Zdenek

Dne pondělí, 9. července 2012 12:16:41 UTC+2 Ioannis Havoutis napsal(a):
Reply all
Reply to author
Forward
0 new messages