treep_robproj

This is a demo related to the software tutorial "management of robotics project, part 2". See: https://owncloud.tuebingen.mpg.de/index.php/s/CNFA32EEY2eNNaE

Setup of the workspace

setup a software folder (name it was you wish)

mkdir ~/software
cd ~/software

setup a python virtual env

python -m venv ./env
source ./env/bin/activate

install the python dependencies

pip install treep colcon-common-extensions pyside2 empy catkin-pkg sphinx breathe matplotlib

install the system-wide dependencies

sudo apt install git python3-pip cmake libcereal-dev libboost-all-dev libgtest-dev libeigen3-dev libedit-dev libncurses5-dev freeglut3-dev libxmu-dev doxygen libxcb-xinerama0 libxrandr2 libxcursor1 libxinerama1 libglfw3-dev

clone this repo:

git clone git@gitlab.is.tue.mpg.de:robotics/treep_robproj.git

clone all the repos of the project:

treep --clone ROBPROJ

compile the workspace:

cd workspace
colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release -DPYBIND11_TEST=OFF --no-warn-unused-cli

sourcing the workspace

source ./install/setup.bash

optionally, have your python virtual env and the workspace sourced for all new terminals, by adding a the end of the file ~/.bashrc something like:

echo "- sourcing robotics management virtual env"
source ~/software/env/bin/activate
echo "- sourcing robotics management workspace"
source ~/software/workspace/install/setup.bash

it is also convenient to be able to compile the workspace from any folder. Can also be added to the ~.bashrc file:

echo "- adding robproj_compile function"
robproj_compile() {
    start_dir=$(pwd)
    cd ~/software/workspace/
    colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release -DPYBIND11_TEST=OFF --no-warn-unused-cli
    cd ${start_dir}
}

Run the demo

More info regarding the demo can be found in the slides https://owncloud.tuebingen.mpg.de/index.php/s/CNFA32EEY2eNNaE

start the realtime process

In a terminal

robpro2_run_cos

This spawns a thread which will write data to the shared memory at a desired frequency of 200Hz.

This process also pings a shared memory array with information about the frequency its main loop is executed.

check if the process is running realtime

While robproj2_run_cos runs, launch in another terminal:

robproj2_period_report

This reads the shared memory array and displays the average period at which robproj2_run_cos runs.

read the data from any python code

import robproj2
data = robproj2.get_cos_value()

display the data live

In another terminal:

robproj2_plot_cos_values

There is the issue that the spawned matplotlib window is difficult (impossible?) to exit. Close the terminal to terminate the program.

Issues, realtime and more infos

executable hanging

If something hangs, stop all the executables and delete the files

/dev/shm/cos_signal
/dev/shm/cos_signal_PERIODS
/dev/shm/sem.cos_signal_PERIODS_mutex

They correspond to the shared memory used by all the program above.

realtime

The code will compile for realtime if the terminal command uname -a displays something containing one of the following (see mpi_cmake_module):

xenomai
lowlatency
preempt-rt
preempt_rt
preempt rt

understand what is going on

Code in the project ~/software/workspace/src/robproj2 is heavily commented.

MPI-IS/treep_robproj