HYU 2021 PBL Robot Control

0. Purpose of Development

[Outline]

1. MATLAB-based PMSM Driver Design for Neuromeka-CORE simulation.
2. Connect Matlab to ROS2.
3. GUI production using ROS2.

1. MATLAB-based PMSM Driver Design for Neuromeka-CORE simulation.

1.1 Development Environment

• Window11 && wsl2.
• Ubuntu 64-bit 18.04 && ROS2 Dashing.
• MATLAB R2021b.

1.2 Prerequisite

• install add-on (ROS Toolbox v1.2).
• install add-on (Robust Control Toolbox).
• install add-on (Control System Toolbox).
• install add-on (Signal Processing Toolbox).
• install Multiparametric toolbox (MPT), using (install_mpt3.m) file.

1.3 MATLAB-based PMSM Driver Version list

• Version_0 : Current & Velocity Control.
• Version_1 : Select [Current / Velosity / Position] Control Mode. (Same control frequency)
• Version_2 : Select [Current / Velosity / Position] Control Mode. (Different control frequency)
• Version_3 : Model Predictive Control(MPC).
• Version_4 : Multi-Parametric Toolbox(MPT) based MPC.
• Version_5 : Disturbance Observer(DOB).
• Version_6 : H-infinity Control.
• Version_alpha : Integrated Controller.
• Version_beta : The latest version.

1.4 Overall Structure

[Block Diagram]

• You can use a Combination of Control Modes.

  • 0: Not use 1: Current(PI) 2: Current(MPC)
  • 0: Not use 1: Velocity(PI) 2: Velocity(MPC) 3: Velocity(H-inf)
  • 0: Not use 1: Position(PI)
  • 0: Not use 1: Current(PI+DOB)

• Velocity(MPC) & Velocity(H-inf) -> Both controllers control the velocity and current at the same time.

• Therefore, when you select the corresponding velocity control mode, the lower current control mode must be selected as (0: Not use).

2. MATLAB to ROS2 && GUI Design

• This meta-packages is for bridging between matlab and local ros2 dashing to control PMSM Driver by using gui interface.

2.1 Prerequisite

• Ubuntu 64-bit 18.04 && ROS2 Dashing : Follow ROS Installation
• QT5 : Follow QT Installation
• Docker : Follow Docker Installation

2.2 Package Structure

• ROS2 PMSM Studio

  pmsm_studio
  ├── CMakeLists.txt
  ├── include
  │   ├── main_window.hpp
  │   ├── qcustomplot.h
  │   └── qnode.hpp
  ├── package.xml
  ├── src
  │   ├── main.cpp
  │   ├── main_window.cpp
  │   ├── plot_window.cpp
  │   ├── qcustomplot.cpp
  │   └── qnode.cpp
  └── ui
      ├── main_window.ui
      └── plot.ui
  
  3 directories, 12 files
  

• ROS2 MATLAB Node

  matlab_node
  ├── BOUND_PI.m
  ├── callback_CurrentMode.m
  ├── callback_DOBoption.m
  ├── callback_ModeSelection.m
  ├── callback_PositionMode.m
  ├── callback_Reference.m
  ├── callback_VelocityMode.m
  ├── mpcgain.m
  └── Test_Code_Final.m
  
  0 directories, 9 files
  

2.3 Environment Setting

• pmsm_studio

  $ cd ~/PBL_RobotControl/matlab2ros/env_settings
  $ ./build.sh
  $ ./docker_run.sh
  $ cd ~/colcon_ws/src
  $ git clone https://github.com/BreadOak/PBL_RobotControl.git 
  

• matlab_node

  Open MATLAB
  move to ~/PBL_RobotControl/Matlab_PMSM_Drive/Test_Code
  Run install_mpt3.m
  

2.4 UI Design

2.5 How to Use

## in docker workspace
$ docker exec -it gpu_ros2 /bin/bash
$ cm
$ sb
$ ros2 run pmsm_studio pmsm_studio

# matlab workspace
Run ~/PBL_RobotControl/matlab2ros/matlab_node/Test_Code_Final.m

2.6 ROS2 Connections

2.7 Results

3. Future Work

• Improve Performance of the Controller.
• Add More Control Modes.
• Add More GUI Function Tabs.
• ...