Robot Operating System

The Robot Operating System (ROS) is a collection of software for all kind of robot applications. In this article I write down some informations and guides for various ros-packages.

Using RPLIDAR in ROS Melodic

If you want to use a cheap laser scanner e.g. RPLIDAR A1 in ROS it is quite simple. The manual can be found here.

First make sure that you have installed python-rosinstall, if not install it.

sudo apt-get install python-rosinstall

Next create a ROS workspace if you not have one already.

mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/src
catkin_init_workspace

Then clone the ROS node for the RPLIDAR in the catkin workspace src directory

git clone https://github.com/robopeak/rplidar_ros.git

Next build it with catkin

cd ~/catkin_ws/
catkin_make

Don't forget to set the environment and it variables when build is complete

source devel/setup.bash

and to check the USB devices e.g. ttyUSB0

ls -l /dev |grep ttyUSB

and set its right to write

sudo chmod 666 /dev/ttyUSB0

Launch the rviz demo and test if everything works

roslaunch rplidar_ros view_rplidar.launch

Using "joy-node" in ROS Melodic

So when you want to use thejoystick-driver with there joy-node you need to install some packages first.

libusb missing

sudo apt-get install  libusb-dev

libspnav / spnav.h missing

sudoapt-get install libspnav-dev

libbluetooth / bluetooth.h missing

sudo apt-get install libbluetooth-dev
sudo apt-get install bluez (not always needed)

libcwiid / cwiid.h missing

sudo apt-get install libcwiid-dev

Now you can clone the source

cd ~/catkin_ws/src
git clone https://github.com/ros-drivers/joystick_drivers.git

and then make it with catkin

cd ~/catkin_ws
catkin_make

If you want to test your joystick first, have a look under what name you joystick is listed:

ls /dev/input/js*

You can cat the output and should see weird signs as an output if the joystick works.

cat /dev/input/js*

As an alternative, use the tool "jstest" therefore you need the joystick package

sudo apt-get install joystick

You can then use it like this

jstest /dev/input/js*

Some more informations about testing Joysticks can be found here!

Using "hector-mapping" in ROS Melodic

The hector-mapping nodes depend on Qt4, so you need to install it first.

sudo apt-get install qt4-qmake qt4-dev-tools

Move into catkin_ws/src, clone the source files and then make them

cd ~/catkin_ws/src
git clone https://github.com/tu-darmstadt-ros-pkg/hector_slam.git
cd ~/catkin_ws
catkin_make

First run of "hector-mapping"

For a quick test of the hector_mapping two launch files needs to be modified.

Edit the "mapping_default.launch" file

nano ~/catkin_ws/src/hector_slam/hector_mapping/launch/mapping_default.launch

and uncomment the second last line to look like this.

<node pkg="tf" type="static_transform_publisher" name="base_to_laser_broadcaster" args="0 0 0 0 0 0 base_link laser 100" />

Modify the third line to his

<arg name="base_frame" default="base_link"/>

and the fourth line to this.

<arg name="odom_frame" default="base_link"/>

Also edit the "tutorial.launch" file

nano ~/catkin_ws/src/hector_slam/hector_slam_launch/launch/tutorial.launch

and replace the third line to look like so

<param name="/use_sim_time" value="false"/>

To start the mapping process use

roslaunch hector_slam_launch tutorial.launch

Some more informations can be found here!

Using "gmapping" in ROS Melodic

(sudo apt-get install ros-melodic-openslam-gmapping)

cd ~/catkin_ws/src
git clone https://github.com/ros-perception/openslam_gmapping
git clone https://github.com/ros-perception/slam_gmapping
(rosdep install --from-paths -i)
cd ~/catkin_ws
catkin_make

SLAM (Simultaneous Localization and Mapping) approaches based on ROS

gmapping

http://wiki.ros.org/gmapping

https://openslam-org.github.io/gmapping.html

Additional informations about gmapping and ROS:

Jackal_navigation

hector_slam

http://wiki.ros.org/hector_slam

https://github.com/tu-darmstadt-ros-pkg/

Additional informations about hector and ros:

SLAM with hector_mapping

ROS and Hector SLAM for Non-GPS Navigation

A comparison of ROS-based visual SLAM methods can be found in the article from Comparison of ROS-based Visual SLAM methods in homogeneous indoor environment written by Ilmir Z. Ibragimov and Ilya M. Afanasyev.

Lidar SLAM without ROS for less than $200

Using "sound_play" in ROS Melodic

To use the sound_play package in ROS Melodic, the easiest way is to install it via apt-get.

sudo apt-get install ros-melodic-sound-play

It is also possible to build the source, but then you need to install all neccessary dependencies by hand.

If you get a "No package 'gstreamer-video-1.0' found" error during catkin_make process, try to install

sudo apt-get install libgstreamer-plugins-base1.0-dev

If you run ROS Melodic on Ubuntu 18.04 on a Raspberry Pi 3 B+ you need to enable the sound in the config.txt file.

sudo nano /boot/firmware/config.txt

Add the following line to the config file.

dtparam=audio=on

If the volume is to low, it's possible to change it with the amixer set command like this.

amixer set PCM -- -1200

It is also possible to set the volume with a percentage value.

amixer sset PCM,0 85%

TF - Error was Lookup would require extrapolation into the future

This error can be caused by delayed processing in some of the nodes, or if a master-slave system is used it is most likely that the used systems do not have the same time. In that case it is a good practice to use a tool like "chrony" to synchronize the times.

apt install chrony

To start chrony daemon during boot use this command

systemctl enable chrony

To synchronize the times, it is neccessary to run chrony on one device as a server. A simple server config is placed at /etc/chrony/chrony.conf and can look like this

# Use time server pools to synchronize the time
pool ntp.ubuntu.com        iburst maxsources 4
pool 0.ubuntu.pool.ntp.org iburst maxsources 1
pool 1.ubuntu.pool.ntp.org iburst maxsources 1
pool 2.ubuntu.pool.ntp.org iburst maxsources 2

# Use a time server to synchronize the time
server de.pool.ntp.org iburst

keyfile /etc/chrony/chrony.keys
driftfile /var/lib/chrony/chrony.drift
logdir /var/log/chrony
maxupdateskew 100.0
rtcsync
makestep 1 3
local stratum 8

# allow an ip to synchronize with server
allow 192.168.0.132

# allow a net to synchronize with server
allow 192.168.1.0/24

A config for a client is shorter.

keyfile /etc/chrony/chrony.keys
driftfile /var/lib/chrony/chrony.drift
logdir /var/log/chrony
maxupdateskew 100.0
rtcsync
makestep 1 3

# server in the network to synchronize the time with
server 192.168.1.191 minpoll 0 maxpoll 5 maxdelay .05

To check the synchronization the "tracking" argument is very useful

chronyc tracking
chronyc -a tracking

Additional informations about "chrony" can be found on Ubuntu Bionic: Using chrony to configure NTP or on docs.oracle.com - Time Synchronization (Chrony)

Rviz does not show robot model

A known problem with ROS melodic and newer Ubuntu system like Bionic Beaver where the robot model does not be visualized in Rviz. Floating point values in the urdf description file will not be parsed correctly, only the integer part of the floating point value gets parsed correctly. To fix it, use

export LC_NUMERIC="en_US.UTF-8"

It often happens if you use german locales in your setup.

Tutorial zum erstellen von Robotermodellen (Building a Visual Robot Model with URDF from Scratch)

Wer sich durch dieses Tutorial durchgearbeitet hat wird sehr schnell feststellen, dass es also auch möglich ist, komplette Mesh-Modelle zu laden. Diese bieten den Vorteil, dass sie wesentlich komplexer sein können als die direkt per URDF erstellten Modelle. Dieser Vorteil ist jedoch auch ein Nachteil, da bei jeder Zustandsänderung die komplexen Mesh-Modelle in rviz neu gezeichnet werden müssen.

Erstellen von einem Mesh-Modell z.B. für rviz mittels SketchUp

Als komfortables und kostenloses Programm hat sich SketchUp von Goolgle herrausgestellt. Mit diesem Programm kann man recht intuitiv einfache dreidimensionale (3D) Objekte erzeugen.

Zur Veranschaulichung habe ich das Bild eines schlichten 3D-Modells (Sick Laserscanner PLS101-312) angehängt. Ich persönlich finde, detailreicher muss solch ein Model auch nicht sein. Zur Not täte es eben auch ein einfacher quadratischer Klotz.

SketchUp Programm Sick PLS101-x12

Nachdem das Modell erstellt wurde lässt es sich als Google Earth-Datei (*.kmz) oder als Collada-Datei (*.dae) exportieren.

Für rviz ist nur die Collada-Datei interresant, da diese dort geladen werden kann.

Es kann manchmal zu Problemen beim Laden von Collada-Dateien kommen. Dann hilft es oft, die *.dae Datei in eine *.stl Datei umzuwandeln. Dafür kann das kostenlose Open-Source Programm MeshLab.

Sick PLS101-x12

SketchUp Sick PLS101-x12

Sick PLS Mount

SketchUp Sick PLS Mount

Pan Tilt Sick Mount

SketchUp Pan Tilt Sick Mount (Download SketchUp File)

Meyra Genius Front Wheel

SketchUp Meyra Front Wheel

Signal Light

SketchUp Signal Light

Einbinden von mesh Modellen in rviz

TBD

<address>
  <strong>Twitter, Inc.</strong><br>
  795 Folsom Ave, Suite 600<br>
  San Francisco, CA 94107<br>
  <abbr title="Phone">P:</abbr> (123) 456-7890
</address>

<address>
  <strong>Full Name</strong><br>
  <a href="mailto:#">Diese E-Mail-Adresse ist vor Spambots geschützt! Zur Anzeige muss JavaScript eingeschaltet sein!</a>
</address>