Simple tf causes "Message removed because it is too old" about livox_ros_driver

[Livox-SDK]

Hi,satoshimiya,
You can first synchcronize the lidar time with gps, the time of lidar is the absolute time after turning on.

[RakVish1342]

@satoshimiya, I had the same issue with the Livox and was wondering the same thing; Unlike while rosbag files which can publish the simulation clock using the --clock flag, the Livox does not seem to publish into /clock. So how can this issue of livox old messages be resolved?

@Livox-SDK, if I do not have any other sensors along with the Livox (not even a GPS sensor), how can I synchronize Livox messages with the current ROS time?

I used a hacky work around to resolve the issue. I created a node that subscribes to /livox/lidar and /livox/imu. Within the callback functions of each of the topics, I modify the incoming message's msg.header.stamp with the ros::Time::now() or rospy.Time.now() and publish it back into a different topic. I'm not sure if this follows ROS best practices but it seems to work for the isolated sensor of just the Livox.

[gongyue666]

@RakVish1342 is it working for your method? I face the same problem. Does Livox can not show pointclouds in RViz with a static_tf_publisher or it can not work in our codes too?

[RakVish1342]

@gongyue666, I do not use the livox lidar any more. However, the technique that I had mentioned in the workaround should resolve your rviz visualization issue. In my workaround I write a subsrciber that subscribes to the /livox/lidar and /livox/imu topics. The subscriber add timestamps manually to the Header of the messages using ros::Time::now(). Then I re-publish those messages (now containing the manually added time stamps) to a new topic like /livox/lidar_new and /livox/imu_new. I record and visualize only the _new topics. While playing back the rosbag file remember to use the --clock flag to use the rosbag's sim_time rather than the onboard/computer time.

[gongyue666]

@RakVish1342 Thx very much～