NatNet 2 support

[omardarwish1995]

I get the following error when I run "rosrun natnet_ros client _server:=10.195.79.98":

_```
radlab@DESKTOP-TSTVI8U:~/catkin_ws$ rosrun natnet_ros client _server:=10.195.79.98
[INFO] [1551755108.133212]: Connecting to 10.195.79.98
Traceback (most recent call last):
File "/home/radlab/catkin_ws/src/natnet_ros/natnet_ros/scripts/client", line 217, in
node.run()
File "/home/radlab/catkin_ws/src/natnet_ros/natnet_ros/scripts/client", line 207, in run
client = natnet.Client.connect(server, logger=self.log)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/comms.py", line 415, in connect
return cls._simple_connect(server, logger, timeout)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/comms.py", line 393, in _simple_connect
timeout=timeout)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/comms.py", line 154, in wait_for_message_with_id
return protocol.deserialize_payload(message_id, payload), received_time
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/protocol/common.py", line 271, in deserialize_payload
return _registry.deserialize_payload(*args, **kwargs)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/protocol/common.py", line 221, in deserialize_payload
message = message_type.deserialize(payload_data, version)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/protocol/ServerInfoMessage.py", line 67, in deserialize
high_resolution_clock_frequency = data.unpack(uint64_t)
File "/home/radlab/catkin_ws/src/natnet_ros/python_natnet/python_natnet/src/natnet/protocol/common.py", line 92, in unpack
value = struct_type.unpack(self.data[self.offset:self.offset + struct_type.size])
struct.error: unpack requires a string argument of length 8

Motive Version: 1.10.3
ROS Version: Kinetic
Ubuntu: 16.04
Python: 2.7.12

Note: It works when using the fake client and crashes when I enter the Motive IP Address

Thanks in advance! 

[mje-nz]

Thanks for letting me know!

Unfortunately I don't have access to a Motive version that old, so I don't have any test data for the old protocol. If you give me a packet capture of another NatNet client successfully communicating with your version of Motive I'll put it on my to-do list.

[omardarwish1995]

We were able to get it working with Natnet 2.10. The attached data was captured after we ran the "WinFormTestApp.exe" in the NatnetSDK samples.

Is that what you were asking for or do you need more information?

[mje-nz]

Nice work. What I need is a recording of your raw network traffic, captured with a network sniffing tool like Wireshark.

First, start the capture with the filter "port 1510 or port 1511" so you only include NatNet traffic.

Then, open WinFormTestApp.exe and do whatever you did to get it going. You should see packets appearing in Wireshark. Give it a few seconds, then stop capturing in Wireshark, save the capture file, and upload it here.

It would be ideal if you could run WinFormTestApp.exe on a different computer to Motive, but if you can't then I can probably still figure out what's going on.

[omardarwish1995]

WireSharkData.zip

Thank you for your reply. Unfortunately we were not able to run it on two separate machines.

Please see attached the captured data requested

[mje-nz]

That's great, thanks! The error you reported is in the skeleton parsing code, which I've never been able to test as we only use our mocap for robots. The packet format doesn't seem to be what NaturalPoint says it is, would you mind collecting a little more data for me?

Open a command prompt in the NatNet SDK samples folder. Start the capture in Wireshark as before. Run SampleClient.exe in the command prompt, press 1 to connect, wait a few seconds while data flashes past, then press q to quit. It should look something like this:

Select all the output (Ctrl-A), copy it (Enter), and attach it here along with the capture file.

[omardarwish1995]

Please see attached the captured data and the SampleData.exe data

Thank you,

WireShark_03-13-2018.zip

[mje-nz]

That's perfect, thanks again.

I've made some changes in the dev branch which should fix your issue. Please try it out and let me know how it goes. Depending on your git config you might have to run git submodule update when you change branches.

At the moment I have each bone in the skeleton published as a PoseStamped. What do you want to be able to do with the data? I'm open to suggestions for how to represent it.

[omardarwish1995]

Thank you very much. Unfortunately we get a new error (please see attached.)

and to answer your question, we are trying to capture live stream data and create a live stream simulation.

Thanks in advance

[mje-nz]

Okay, that's an issue with connecting to old versions of Motive. It should be fixed now, please update (git pull and git submodule update) and try again.

[omardarwish1995]

Hi,
We tried the updated library and ran into a similar error as seen in the attached screenshot. As you can see, we are using the right data port and multicast address. Even when we tried to use an IP address instead of setting Motive to loopback, we got the same exact error. Any help fixing this would be appreciated!

[mje-nz]

You have to be in src/natnet_ros when you run git pull and git submodule update or they won’t do anything ;)

[omardarwish1995]

Hi Matthew,
Sorry for not replying earlier as we were on spring break and had no access to the mocap system.

I tried your suggestion in the previous comment and it still gave me the same error as attached.

Is it because the program is using "multicast_address" while it is labeled "Multicast Interface" in the motive application?

Please advise.

[mje-nz]

Ah I see the problem, try again now.

[mje-nz]

Hi Omar, have you had a chance to test these changes?

[omardarwish1995]

Hi Matthew,

We tried out the update and still ran into errors. It seems too tedious of a process to force the library to work with an older version so after speaking with senior members of our lab, we have decided to update our Motive software to the latest version (2.1).
The purpose of our project is to live stream data from Motive in order to create simulations of the human motions and eventually control robots. My teammates and I would like to acknowledge your work and your assistance to us in our paper. Is there a certain affiliation that you would like us to recognize?

Again, we are immensely thankful for your help. I will keep you in the loop once we get our updated software and test your library.

Best,
Omar

[mje-nz]

Sorry Omar, I missed this one! If it's not too late, my affiliation is Electrical and Computer Engineering, University of Canterbury, New Zealand. I'd be interested to see your paper if you don't mind sharing.

How are you getting on with updating Motive?