Skip to content

Giggus/fri

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

42 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Fast Robot Interface

The Fast Robot Interface (FRI) library allows for hard real time communication to the KUKA LBR Med, see KUKA's paper. This folder contains the C++ client side library to setup the UDP connection to the LBR Med. However, hard real time is only supported if the FRI is compiled on a real time OS.

Build and Installation

Other than what KUKA provides, our FRI comes with CMake support. This allows for cross-platform compatability, therefore, the Cross Platform Build and Installation instructions are the same for Linux and Windows. However, the prerequisites are differntly obtained. How to obtain the prerequisites is explained in Linux Prerequisites and Windows Prerequisites, respectively.

Linux Prerequisites

Skip any of these if already satisfied.

  • Install Git. In a terminal, run sudo apt install git
  • Install CMake. In a terminal, run sudo apt install cmake

Windows Prerequisites

Skip any of these if already satisfied.

  • Install the new Windows Terminal. Search for Microsoft Store in Start and open it. Therein, search for Windows Terminal, and install it
  • Install Chocolatey by following the instructions
  • Install Git. In an administrative Windows Terminal, run choco install git.
  • Install CMake. In an administrative Windows Terminal, run choco install cmake. Add CMake to your Path. Therefore, search for Edit the system environment variables in Start, and open it. Open Environment Variables..., and double click on Path under System variables. Click on New, and add C:\Program Files\CMake\bin (usually CMake is installed there, might differ)

Cross Platform Build and Installation

Clone this repository

git clone https://github.com/KCL-BMEIS/FastRobotInterface.git

Build and install the FRI library

cd FastRobotInterface
mkdir build
cd build
cmake ..
cmake --build . --config Release --target install # builds and installs the FRI library in Release mode

Usage

The FRI comes with example apps. These apps require some prerequisites that are explained below.

Prerequisites

Follow the Build and Installation instructions for your OS. Install Sunrise Workbench on your computer. This step requires Windows as OS.

Sunrise Workbench

Sunrise Workbench is KUKA's Java IDE that allows you to program the LBR Med.

  • Download it from the RViM shared folder
  • Extract the .zip file and run the Sunrise Workbench Setup
  • Follow the install instructions

Example Apps

Exemplary applications for the C++ client side are located inside the apps folder. Each of these apps has a Java equivalent for the server side.

Connect Laptop

  • Connect your laptop, therefore, establish an Ethernet connection to connector X66 at the KUKA controller
  • The KUKA controller's default IP is 172.31.1.147. Configure the same network on your laptop, therefore
    • On Windows
      • Search for View network connections in Start and open it
      • Right click on the Ethernet connection and open Properties
      • Double click Internet Protocol Version 4 (TCP/IPv4) and set the IP address to 172.31.1.148 and the Subnet mask to 255.255.0.0
    • On Ubuntu 16 (might differ for other Linux distributions)
      • Search for System Settings and open it
      • Go to Network -> Wired -> Options
      • Go to the IPv4 Settings tab and set the IP address to 172.31.1.148 and the Netmask to 255.255.0.0

Server Side - KUKA Controller

You will have to follow the instructions in Connect Laptop. The FRI has to be installed on the controller. Therefore, the Sunrise Workbench IDE is used.

  • Open the Sunrise Workbench
  • Create a new project, therefore
    • Click File -> New -> Sunrise project
    • Leave the default IP (172.31.1.147) and click Next
    • Give your project a name, e.g. FRI and click Next
    • Select LBR Med 7 R800 and click Next
    • Under Media Flange, select Medien-Flansch Inside elektrisch, and click Next
    • Click Finish (might take some time). Select RoboticsAPI Application, and press Next
    • Click Finish
  • Setup the KUKA controller, therefore
    • Double click the StationSetup.cat under FRI in the Package Explorer and add the LBR Med 7 R800 to the Topolgy (below left)
    • Remove the LBR_Med_7_R800_2, which has no Media Flange, and which we are not using (below right)

    • In the Software tab, make sure that the checker boxes for the FRI extensions are selected (below left)
    • In the Configurations tab, make sure that the IP is set to 172.31.1.147 (below right)
    • Save the StationSetup.cat via Ctrl + s

    • In the Installation tab, click Install (below left)
    • Synchronize your project with the KUKA controller (below right)

Client Side - Laptop

You will have to follow the instructions in Connect Laptop, and make sure that the FRI was installed to the KUKA controller, according to Server Side - KUKA Controller. The client side requires to build the apps, which requires that the FRI was installed according to Build and Installation. Then, open a terminal and do

cd apps
mkdir build
cd build
cmake -DCMAKE_PREFIX_PATH='path/to/lib' .. # on Windows this should be 'C:\Program Files (x86)\FastRobotInterface'
                                           # on Linux this should be '/usr/local'
cmake --build . --config Release           # builds the apps in Release mode

Run the Apps

To run one of the example apps, they have to be started on the smartHMI (KUKA's smartpad), as well as on the laptop.

  • On the smartHMI
    • Select an Application (left below), e.g. the LBRJointSineOverlay
    • Press one of the enabling switches half way (grey buttons on the back of the smartHMI). The joint control A1-A7 will light up (center below)
    • Press and hold the play button (below right), and the enabling switch

  • On the laptop open a terminal, and run
cd apps/build         # on Linux 
cd apps/build/Release # on Windows
./lbrjointsineoverlay

The robot should now be controlled by your Laptop, well done 😄! For open questions please leave an Issue.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • POV-Ray SDL 63.3%
  • C++ 31.6%
  • C 4.3%
  • CMake 0.8%