joint_group_ff_controllers

Overview

This package implements basic controllers (similarly to ros_controllers).

License

The source code is released under a MIT license.

Author/Maintener: EtienneAr
Affiliation: INRIA Paris

This package package has been tested under [ROS] Melodic on respectively Ubuntu 18.04. This is research code, expect that it changes often and any fitness for a particular purpose is disclaimed.

Installation

Building from Source

To build from source, clone the latest version from this repository into your catkin workspace and compile the package using

cd catkin_workspace/src
git clone https://github.com/inria-paris-robotics-lab/joint_group_ff_controllers.git
cd ../
rosdep install --from-paths . --ignore-src
catkin build

Controllers

Effort controller : joint_group_effort_ff_controller

This controller tracks position, velocity and effort commands for a set of joints using a pd-feedforward controller.
It expects a EffortJointInterface type of hardware interface. The control is done as followed, for each joint :

Subscribed Topics

• command ([joint_group_ff_controllers/setpoint])

  The positions, velocities and efforts set point for each joint, aswell as the timeout value.
  Note: if timeout < 0, then the timeout is ignored and the command will be executed indefinitely.

Parameters

• joints (string[])

  The list of joints to control.

• <joint> (associative array)

  Key/Value pairs for the controller gains. Expected keys are kp, kd, kp_safe, kd_safe.

Controller configuration example

arm_right_controller_test:
  type: joint_group_ff_controllers/JointGroupEffortFFController
  joints:
  - arm_right_1_joint
  - arm_right_2_joint
  - arm_right_3_joint
  - arm_right_4_joint
  - arm_right_5_joint
  - arm_right_6_joint
  - arm_right_7_joint
  arm_right_1_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_2_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_3_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_4_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_5_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_6_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }
  arm_right_7_joint: { kp: .01, kd: .01, kp_safe: 1000, kd_safe: 64. }

Note : Usually kp and kd are set to small values as the effort feed-forward should do most of the "work". kp_safe and kd_safe on the other hand are bigger as they should be able to keep the robot stable in it's current position (without any feed-forward), when timeout occurs.

Velocity controller : joint_group_velocity_ff_controller

This controller tracks position and velocity commands for a set of joints using a p-feedforward controller.
It expects a VelocityJointInterface type of hardware interface. The control is done as followed, for each joint :

Subscribed Topics

• command ([joint_group_ff_controllers/setpoint])

  The positions and velocities set point for each joint (efforts field is ignored in this case), aswell as the timeout value.
  Note: if timeout < 0, then the timeout is ignored and the command will be executed indefinitely.

Parameters

• joints (string[])

  The list of joints to control.

• <joint> (associative array)

  Key/Value pairs for the controller gains. Expected keys are kp, kp_safe.

Controller configuration example

right_arm_ff_controller:
  type: joint_group_ff_controllers/JointGroupVelocityFFController
  joints:
  - right_shoulder_pan_joint
  - right_shoulder_lift_joint
  - right_elbow_joint
  - right_wrist_1_joint
  - right_wrist_2_joint
  - right_wrist_3_joint
  right_shoulder_pan_joint:  { kp: .1, kp_safe: 100 }
  right_shoulder_lift_joint: { kp: .1, kp_safe: 100 }
  right_elbow_joint:         { kp: .1, kp_safe: 100 }
  right_wrist_1_joint:       { kp: .1, kp_safe: 100 }
  right_wrist_2_joint:       { kp: .1, kp_safe: 100 }
  right_wrist_3_joint:       { kp: .1, kp_safe: 100 }

Note : Usually kp is set to small values as the effort feed-forward should do most of the "work". kp_safe on the other hand is bigger as they should be able to keep the robot stable in it's current position (without any feed-forward), when timeout occurs.