Robotic Deep Rolling GUI

A GUI for doing deep rolling using ABB industrial robots.

Simulator Installation

GUI Installation

System Tuning (WIP)

Toolpath Generation (WIP)

GUI Usage (WIP)

How to use the ABB Robot

Control goal:

Design and track a position / force profile

Control structure:

• Generalized damper force control using force/torque sensor feedback, with filtering at sensor and robot
• Look-ahead control for robot delay compensation

Implementation:

Parametric toolpath generation

• 16 parameters for geometry, force, and velocity
• Custom toolpath protocol designed for hybrid motion/force control

GUI for toolpath execution

• Interface with F/T sensor and robot controller
• Toolpath execution state machine covering multiple command types
• Designed for use with curved parts as well

High fidelity simulations for testing and tuning

• Connect to RobotStudio to simulate robot dynamics
• Linear spring model for joint flexibility modeling
• Can be expanded to bypass RobotStudio and use delay+LPF model for full robot simulation in one piece of software

---

Sources

Older deep rolling code https://github.com/ShuyoungChen/force_motion_control_physical

Reference for general robotics toolbox https://github.com/rpiRobotics/rpi_general_robotics_toolbox_py

Folder breakdown

rapid

Code which runs on ABB robot controller

src > controller

Executes commands, from the GUI or from the toolpath file

src > ft_sensor

Reads in force/torque data from ATI sensor

This file is based on this python3 code https://github.com/hehonglu123/MARS_Sample_Return

That code is based on this python2 code https://github.com/rpiRobotics/rpi_ati_net_ft

Code has been modified to be nonblocking

src > ft_simulation

Robot deflection simulation for simulating force/torque sensor readings when using RobotStudio

src > gui

Qt GUI made in Qt Designer

If you install pyqt5-tools using python -m pip install pyqt5-tools you will see Qt Designer at

Python39\Lib\site-packages\qt5_applications\Qt\bin\designer.exe

src > robot_kin

Forward and inverse kinematics for ABB IRB-6640

src > rpi_abb_irc5

Interface to ABB's EGM

Based on https://github.com/rpiRobotics/rpi_abb_irc5

Ported to python3, and added a function for Cartesian robot control

src > toolpath_gen

Generate toolpath files for the controller to execute