multi_robot_state_publisher

Reduce the compute required to publish frames for multiple robots.

multi_robot_state_publisher_node

Read robot configs from the Parameter Server and publish frames for them.

Synopsis

multi_robot_state_publisher_node ~publishing_rate:=50

Subscribed Topics

• /joint_states: Gets joint data for a robot.

Published Topics

• /tf: Publishes frame data to the dynamic TF tree.

• /tf_static: Publishes latched frame data once to the static TF tree.

Parameters

• ~robots/<robot> (struct, default: {}): Configuration settings for each robot. Has the following form:

  • robot_description (str): The URDF of the robot in XML format. Typically this is loaded either from a file or with the xacro tool to convert a XACRO file into an URDF.

  • publish_frequency (float, default: 50.0): The rate at which this robot will update its joints. If this is higher than the node publish rate, it will be capped. If it is lower, it will still publish at the higher rate but the data for this robot will only be updated at this rate.

  • use_tf_static (bool, default: true): If true, publish fixed transforms to the /tf_static topic instead of /tf.

  • ignore_timestamp (bool, default: true): If true, joint_states messages are accepted, no matter their timestamp.

  • tcp_nodelay (bool, default: false): If true, tell the joint_states publisher to set the TCP_NODELAY socket option, which disables the Nagle algorithm. Doing so causes segments to be sent as soon as they are ready instead of bundling them. This means there will be more frames with smaller packets which translates to poorer network utilization but lower latency for the joint_states topic, which may be important for real-time control applications. See also: tcp(7).

    tf_prefix (str, default: ""): A prefix to add to the tf frame which prevents namespace collisions for multiple robots.