RobAIR/Pilotage Automatique du Robot

De fablab
Révision de 16 juin 2013 à 08:22 par Belluzj (discussion | contributions) (15/05/2013 Flyer)

(diff) ← Version précédente | Voir la version courante (diff) | Version suivante → (diff)
Aller à : navigation, rechercher

Project definition

Based on an RobAIR environment map (fournie par l'équipe "Cartographie Dynamique", mais simulée dans un premier temps), we should nous devons guide RobAir with an :

  • Automatic way : Only the destination coordinates are given to RobAIR
  • Half-automatic way : RobAIR is directed manually and should only autocorrect the trajectory to avoid collision.

Linked pages


The project team is composed by the following members :

=> Send us an email on the ensimag fablab mailinglist .

Project avancement

14/02/2013 session

16/02/2013 session

  • Discovery of gmapping. This package allows us to simulate the RobAIR environment map creation and then to reproduce the work of the team "dynamic cartography"

19/02/2013 session

  • Metting with RICM5 team(report available here).
  • Project redifinition : we should take into account the kinect data.
  • beginning of the nodes we should implement.

20/02/2013 session

  • Used and produced data listing.

22/02/2013 session

  • Kinect is working on Ubuntu 12.04, see installation here
  • To link ROS and Kinect, we should reinstall ROS fuerte.

Commands to launch :

Launch the Kinect
$ roslaunch openni_launch openni.launch
Get an RGB picture
$ rosrun image_view image_view image:=/camera/rgb/image_color
Get a depth picture
$ rosrun image_view disparity_view image:=/camera/depth_registered/disparity

14/03/2013 session

  • Packages installation to activate human skeletons detection : link
In pi_tracker directory, modify the last line in the file CMakeList.txt into :
$ target_link_libraries(skeleton_tracker glut OpenNI GL GLU)
In skeleton_markers directory, modify the last line in the file CMakeList.txt into :
$ target_link_libraries(skeleton_tracker glut GL GLU log4cxx)

Then it should compile without error on fuerte.

19/03/2013 session

To tell that a directory contains ROS packages :
$ roscd
$ rosws set <repertoire>
To launch openni : replace in skeleton_traker/launch/markers_from_tf.launch the line
<include file="$(find openni_camera)/launch/openni_node.launch"/>
with :
<include file="$(find openni_launch)/launch/openni.launch"/>
  • Skeleton Tracer launch :

Open 3 different terminal. In the first one :

$ roscore

In the second one :

$ roslaunch skeleton_markers markers_from_tf.launch 

In the third one :

$ roscd skeleton_markers/
$ rosrun rviz rviz -d markers_from_tf.vcg

27/03/2013 session

  • Lidar received for the Slam subproject. We also received the tablet.
  • Installation of Ubuntu 12.04 LTS and ROS fuerte on the tablet.

03/05/2013 session

  • RobAIR correctly received at Ensimag on 02/05/2013
  • cable soldering
  • Infrared sensors and Arduino Uno installation and configuration. We have adjusted the height detection to be compliant with RobAIR. We also tested the Arduino with two parallel sensors and we have modified the source code to work with 3 sensors.

Here are some pictures of RobAIR installation and the infrared sensors tests :

07/05/2013 session

$ roscd
$ git clone robair_demo
$ rosws set robair_demo
$ source setup.bash
$ rosmake robair_demo
$ chmod +x robair_demo/scripts/*

In order to make some parts work (the application that sends commands to the motors), we had to add a custom udev rule to relax permissions on the file /dev/ttyUSB0

$ cat /etc/udev/rules.d/60-fablab.rules 
KERNEL=="ttyUSB0" MODE="0666"

08/05/2013 session

  • Intern improvement of RobAIR.
  • Adding 2 ultrasonic sensors : we now have got 3 infrared sensors and 4 ultrasonic sensors.
  • Looking for a way to have a good ultrasonic sensor gestion ( use a bunch in parallel :
  • All the sensors that we will use are now plugged. We could close the box for a while in order to make it roll without cable disconnection risk.
  • Work on the ROS node for the Arduino connected sensors (Ultrasonic + infrared)
  • Development of ROS nodes at based on work from RICM4

09/05/2013 session

  • Total rewriting of the RICM4 source code in order to have something easily comprehensible and reusable.
  • Creation of a script that launch all the needed commands to boot RobAIR and all its components.

gnome-terminal -x bash -c "roscore"
gnome-terminal -x bash -c "rosrun robair_demo"
gnome-terminal -x bash -c "rosrun robair_demo"
gnome-terminal -x bash -c "rosrun robair_demo"
  • We looked for an efficient way to create an bootable iso of Ubuntu with the content of the tablet ssd. That will allow other teams to download the iso, install it on the tablet and have everything configured like it should be.
  • Arduino source code for ultrasonic and infrared sensors completely written.
  • Python source code for infrared sensors completely written : RobAIR stops when a hole is in front of him !
  • Python source code for ultrasonic sensors completely written : RobAIR stops when somebody walks in front of him !

Test infrarouge 1
Test infrarouge 2
Test ultrason

  • English traduction of our project page.

10/05/2013 session

  • ROS for kinect node installation on the tablet. Caution ! If you follow the kinect for ROS tutorial written earlier, you should change the version of openni and nite from x64 to x86 in the url (the tablet runs in 32bits).
  • Experiencing installation issues. Device not detected.

12/05/2013 session

  • Kinect with ROS is working ! We have got the depth and RGB image.

14/05/2013 session

  • Implementation of speed reduction when approaching an obstacle.*Sensor calibration : Different kind of floor influences the sensor sensibility*Increasing RobAir speed when turning
  • Looking for a way to save the current state of the tablet OS : remastersys.

Remastersys Tuto

On récupère la clé d'authentification

$ wget -0 - | sudo apt-key add -

On ajoute le dépôt

$ deb precise main
$ sudo apt-get update

On installe remastersys

$ sudo apt-get install remastersys remastersys-gui ubiquity-frontend-gtk

17/05/2013 session

  • The ISO image has been created. It weights around 1.9 Gb so it could be burned on a DVD or uploaded.

21/05/2013 session

We recorded some videos of our robot:

24/05/2013 session

We uploaded the ISO image we created. Here is the link:

29/05/2013 defense day

Here are our defense day slides.

15/05/2013 Flyer

We made Fichier:Robair-flyer-recto-joined.pdf to show RobAIR's key features and design choices. This work is to be shared with other teams.

Source files: Robair-flyer-recto.svg Robair-flyer.svg