hh2564 / franka_feedback_controller

This is a ROS package of PD feedback controller to control Franka Emika Panda arm in Gazebo simulation environment.

Franka Feedback Controller

Package Description

This is a ROS PD feedback controller package to control Franka Emika Panda arm in Gazebo simulation environment.

Getting Started

This package was tested in simulation environment Gazebo with:

• Ubuntu 20.04.4 LTS
• Ros Noetic
• Libfranka 0.9.0
• franka_ros 0.9.0

You can go to the their official website to find more instruction on installing these packages.

PD Feedback Controller

For feedback controller approach, we implemented a PD controller: desired joint velocity is proportional to joint position error $\theta_e=\theta_d-\theta$ and joint velocity error $\dot\theta_e=\dot\theta_d-\dot\theta$, where $\theta_d$ and $\theta$ are the desired and measured joint position, and $\dot\theta_d$ and $\dot\theta$ are the desired and measured joint velocity respectively.

We can obtain the desired joint toque $\tau_{feedback}$:

$$\begin{align} \tau_{feedback} = K_p\theta_e+K_d\dot\theta_e+G(\theta) \end{align}$$

Whereas $K_p$ and $K_d$ are the appropriate p-gain and d-gain, and $G(\theta)$ is gravity compensation.

Run Controller

To run the feedback controller in Gazebo simulation, run the following code in command line:

roslaunch franka_feedback_controller franka.launch