Setting up ROS network with Velodyne 32e

Question

I am trying to run ROS across multiple computers, with one processing data from the velodyne lidar and another computer receiving the published messages. I followed the instructions here from the wiki tutorial (Getting Started with the HDL-32E) in order to connect to the velodyne, but now I cannot see any of the published messages from my other machine. I am able to ping back and forth between the computers and set the ROS_IP in both .bashrc files. Everything is running on groovy.

---

Originally posted by sfergs on ROS Answers with karma: 11 on 2013-09-14

Post score: 1

---

Original comments

Comment by tfoote on 2013-09-16:
Can you rostopic pub and subscribe between the machines for a simple test?

Comment by sfergs on 2013-09-17:
I am able to rostopic pub and subscribe to all topics on both machines except the velodyne points.

Comment by joq on 2013-09-17:
What does rostopic list show on each machine?

Comment by tfoote on 2013-09-17:
Do you have enough bandwidth and computation available to send the topic across the wire? The Velodyne produces a lot of data.

Answer 1

I was trying to do something similar and even had long discussion with Velodyne engineers. They are still investigating the issue. The conclusion is that home used router/AP might not be able to support large traffic required by the velodyne. Anyway Here is a list of known configuration that will/will not work:

1. Host connected to a wireless router, lidar connected to the same router, client connect to the router wirelessly. Your ROS client will not be able to communicate with host, nor receive velodyne packets.

2. Host, lidar, client are all connected to the router through RJ45 cable. It will work as long as your network setup is correct.

3. Host connected to the lidar through RJ45 cable, then connect to the router wirelessly, client also connect to the router wirelessly. You will loss velodyne packets, but it might "appears" working.

Hope that helps...

---

Originally posted by BenMa with karma: 76 on 2013-09-23

This answer was ACCEPTED on the original site

Post score: 1

Answer 2

This is not an answer to your question, only a suggestion.

Instead of running both the driver and point cloud conversion nodes on the same machine you can reduce the communication bandwidth significantly by running the driver locally and the point cloud conversion on your other machine. To see what the effect would be:

$ rostopic bw /velodyne_packets
$ rostopic bw /velodyne_points

The point cloud is about four times bigger than the raw packets, and publishes at the same frequency.

---

Originally posted by joq with karma: 25443 on 2013-09-17

This answer was NOT ACCEPTED on the original site

Post score: 1