About | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange

Only released in EOL distros:   groovy hydro indigo jade kinetic

Prerequisites

Running this node requires you to have a calibrated force-torque sensor (bias of the sensor + mass & center of mass of gripper) and an accelerometer whose reference frame has already been calibrated to some reference frame of the robot. You can do this by using the force_torque_sensor_calib package. Please look at the tutorials page of the force_torque_tools metapackage for instructions on how to run the calibration node.

ROS API

gravity_compensation_node

The gravity_compensation_node compensates gravitational forces on the gripper of the robot so that force-torque signals are zeroed. Assumes previously calibrated accelerometer and F/T sensor. It also broadcasts through tf the pose of the center of mass of the gripper from the force-torque sensor calibration.

Subscribed Topics

~/ft_raw (geometry_msgs/WrenchStamped)

• The raw (uncalibrated) F/T sensor readings.

~/imu (sensor_msgs/Imu)

• Measurements from an accelerometer already calibrated to the robot's frame.

Published Topics

~/ft_zeroed (geometry_msgs/WrenchStamped)

• force torque measurements after subtracting the sensor bias.

~/ft_compensated (geometry_msgs/WrenchStamped)

• force torque measurements after subtracting the sensor bias and the gravitational forces.

Parameters

~ns (string, default: "~")

• All published topics are pushed down into the <ns> namespace, i.e., the topics for the compensated force-torque measurements (ft_zeroed and ft_compensated) will be published under <ns> (<ns>/ft_zeroed and <ns>/ft_compensated.

~bias (double array)

• 6x1 double array which contains the ([fx, fy, fz, tx, ty, tz]) force-torque bias of the force-torque sensor.

~gripper_mass (double)

• mass of the gripper attached to the force-torque sensor.

~gripper_com_frame_id (string)

• ID of the frame in which the gripper center of mass is expressed.

~gripper_com_child_frame_id (string)

• ID of the (child) frame of the gripper center of mass. This can be freely chosen by the user.

~gripper_com_pose (double array)

• 6x1 array ([x y z r p y]) with x,y,z translation and roll, pitch, yaw orientation (in meters and radians) containing the pose of the gripper COM expressed in the gripper_com_frame_id frame

~gripper_com_broadcast_frequency (double, default: 100.0)

• broadcast frequency (in Hz) of the gripper COM pose on /tf.

~loop_rate (double, default: 1000.0)

• frequency (in Hz) of the main loop of the gravity compensation node (should be at the same frequency as the publish frequency of the force-torque sensor).

Launching the node

An example calibration file required to run this node can be found under config/ft_calib_data_example.yaml in the force_torque_sensor_calib package.

You can launch the node using the gravity_compensation.launch file contained in the launch folder of the package and setting a few roslaunch arguments, e.g.:

    roslaunch gravity_compensation gravity_compensation.launch node_name:=gravity_compensation calib_data_file:=$HOME/.ros/ft_calib/ft_calib_data.yaml ft_raw:=/ft_sensor/ft_raw imu:=/imu/data loop_rate:=650.0 ns=/ft_sensor gripper_com_child_frame_id:=gripper_com

where the arguments are:

• node_name sets the name of the gravity_compensation node being launched.

• calib_data_file: location of the force-torque sensor calibration data.

• ft_raw: name of topic where raw force-torque sensor readings are published.

• imu: name of the topic where the accelerometer readings are published.

• loop_rate: frequency (in Hz) of the main loop of the node.

• ns: namespace under which to publish the topics for the compensated force-torque measurements (ns/ft_zeroed and ns/ft_compensated).

• gripper_com_child_frame_id: ID of the (child) frame of the gripper center of mass. This can be freely chosen by the user.

Tutorials

Please visit the tutorials page of the force_torque_tools metapackage.

Bug Reports & Feature Requests

We appreciate the time and effort spent submitting bug reports and feature requests.

Please submit your tickets through github (requires github account) or by emailing the maintainers.