Full installation instructions: Difference between revisions

From Rsewiki
 
(65 intermediate revisions by the same user not shown)
Line 7: Line 7:


[[Install on raspberry]] first section (prerequisites).
[[Install on raspberry]] first section (prerequisites).
===Opencv 3.4===
It should be OK just to install libopencv-dev - in December 2020 this installs version 3.2.0
sudo apt install libopencv-dev
To find the installed version of OpenCV try
pkg-config --modversion opencv
To install newer OpenCV version 3.4  (as of Jan 2020 it is version 3.4.9)
cd
mkdir -p git
cd git
# this method will install version 4.2 - that is not fully compatible (jan 2020)
# git clone https://github.com/opencv/opencv.git
# git clone https://github.com/opencv/opencv_contrib.git
# use this to get a version compatible with opencv 3.4
# get source (jan 2020)
wget -O opencv-3.4.zip https://github.com/Itseez/opencv/archive/3.4.zip
unzip opencv-3.4.zip
wget -O opencv_contrib-3.4.zip https://github.com/Itseez/opencv_contrib/archive/3.4.zip
unzip opencv_contrib-3.4.zip
# compile the lot, but ommit python3 support, as the compile will fail (as of jan 2020)
cd opencv
mkdir build
cd build
# create makefiles, also for all modules in the opencv_contrib (e.g. ArUco code detect)
# cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_opencv_python3=OFF -D OPENCV_EXTRA_MODULES_PATH=~/git/opencv_contrib/modules ..
cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_opencv_python3=OFF -D OPENCV_EXTRA_MODULES_PATH=~/git/opencv_contrib-3.4/modules ..
# inspect printout to see if needed features are enabled
# use -j2 option not to overload or overheat the processor (j2 means use 2 cores only)
# this probably takes more than an hour.
# if the compilation crashes about halfway or so, it is probably due to lack of memory
#    then just start the compile again, and it is likely to continue (at least for a while)
make -j2
# install into /usr/local
sudo make install
# load environment settings
sudo ldconfig


===Raspicam===
===Raspicam===
NB! this should not be needed


A rather short raspberry pi camera API is available from https://www.uco.es/investiga/grupos/ava/node/40.
A rather short raspberry pi camera API is available from https://www.uco.es/investiga/grupos/ava/node/40.
Line 15: Line 59:
or get the file from sourceforge with no GUI, with:
or get the file from sourceforge with no GUI, with:


  wget --no-check-certificate -O raspicam-0.1.4.zip https://downloads.sourceforge.net/project/raspicam/raspicam-0.1.4.zip?r=https%3A%2F%2Fsourceforge.net%2Fprojects%2Fraspicam%2F%3Fsource%3Dtyp_redirect&ts=1486483484&use_mirror=netix
cd ~/Downloads
  wget --no-check-certificate -O raspicam-0.1.9.zip https://downloads.sourceforge.net/project/raspicam/raspicam-0.1.9.zip?r=https%3A%2F%2Fsourceforge.net%2Fprojects%2Fraspicam%2F%3Fsource%3Dtyp_redirect&ts=1486483484&use_mirror=netix


Once fetched and available on the raspberry, then
Once fetched and available on the raspberry, then
Unpack and install:
Unpack and install:
  unzip raspicam-0.1.3.zip
  unzip raspicam-0.1.9.zip
  cd raspicam-0.1.3
  cd raspicam-0.1.9
  mkdir build
  mkdir build
  cd build
  cd build
Line 31: Line 75:
The library is installed in /usr/local, and therefore to make cmake find it the path needs to be added to startup configuration, in file ~/.bashrc  
The library is installed in /usr/local, and therefore to make cmake find it the path needs to be added to startup configuration, in file ~/.bashrc  


nano ~/.bashrc
add at the end:
  export CMAKE_PREFIX_PATH=/usr/local/lib
  export CMAKE_PREFIX_PATH=/usr/local/lib


(needs a re-logon to be activated, or repeat on the command line)
(needs a re-logon to be activated, or repeat on the command line)
===Userland-master===
To be able to install mobotware (not used pt on robobot)
Follow the guide under REGBOT for userland-master install - [[Install on raspberry]]


===DNSMASQ===
===DNSMASQ===


To enable the robobot to be connected directly to a PC, then it is easier if the robobot provides an IP for the PC.
To enable the ROBOBOT to be connected directly to a PC, then it is easier if the ROBOBOT provides an IP for the PC.


Install DNSMASQ
Install DNSMASQ
Line 53: Line 93:


  sudo nano /etc/dnsmasq.conf
  sudo nano /etc/dnsmasq.conf
 
find and change/add the following line (a bit down, eg to serch use: ctrl-W dhcp-range)
find and change/add the following line (for eth0)
  dhcp-range=192.168.17.100,192.168.17.150,12h
 
  dhcp-range=eth0,192.168.0.100,192.168.0.150,12h


To work, the eth0 must have an IP, if noone provides one,
To work, the eth0 must have an IP, if noone provides one,
change /etc/dhcpcd.conf (dhcp client deamon) to have a default IP, if no DHCP server is available
change /etc/dhcpcd.conf (dhcp client deamon) to have a default IP, if no DHCP server is available


Add the following lines at the end of /etc/dhcpcd.conf to have a fall back behaviour for eth0 and eth1.
Edit of /etc/dhcpcd.conf to have a fall back behaviour for eth0 and eth1.


sudo nano /etc/dhcpcd.conf
Add the following lines at the end (uncomment and modify):
  # define static profile
  # define static profile
  profile static_eth0
  profile static_eth0
  static ip_address=192.168.0.2/24
  static ip_address=192.168.17.1/24
  static routers=192.168.0.1
  # static routers=192.168.0.1
  static domain_name_servers=192.168.0.1
  # static domain_name_servers=192.168.0.1
   
   
  # fallback to static profile on eth0
  # fallback to static profile on eth0
Line 78: Line 118:
Now, after a reboot, you should be able to connect a PC directly with at network cable to a PC, and the PC should get an IP from the raspberry, so now
Now, after a reboot, you should be able to connect a PC directly with at network cable to a PC, and the PC should get an IP from the raspberry, so now


  ssh local@192.168.0.2
  ssh local@192.168.17.1


===ROBOBOT mission demo C++===
===Sound===


This is an example software in C++ to access both raspberry camera and REGBOT, and with an example mission controlled from the raspberry.
====festival====


Get the robobot software from the svn repository:
Easy understandable voice, but takes time to generate
and require two intermediate files


  svn checkout svn://repos.gbar.dtu.dk/jcan/regbot/mission mission
  sudo apt install festival
echo "The brown fox jumps over lazy dog" > aaa.txt
text2wave aaa.txt -o aaa.wav
aplay aaa.wav


To be able to compile the demo software CMAKE needs also to use the user installed library (raspicam installed above),
====Music files====
so add the following line to ~/.bashrc:


  export CMAKE_PREFIX_PATH=/usr/local/lib
This part works
 
Play sound files, like
 
  sudo apt-get install sox libsox-fmt-all
play -v0.1 music.mp3
 
This line plays an mp3 file with volume reduced by a factor 0.1.
The at startup the blue button on the gamepad plays the file in /home/local/Music/music.mp3.<br>
This is a symbolic link to some real music (Radetzky Marsch in this case):
 
ls ~/Music
    radetzky-marsch_Schloss-Schoenbrunn-Konzerte_Wien_full-length.mp3
cd ~/Music
ln -s radetzky-marsch_Schloss-Schoenbrunn-Konzerte_Wien_full-length.mp3 music.mp3


Then build Makefiles and compile:
===Sound level===


cd mission
On a Raspberry pi version 3+ (not tested on other versions)
cd build
cmake ..
make


Then test-run the application:
$ amixer
  Simple mixer control 'PCM',0
    Capabilities: pvolume pvolume-joined pswitch pswitch-joined
    Playback channels: Mono
    Limits: Playback -10239 - 400
    Mono: Playback 400 [100%] [4.00dB] [on]


./mission
An thus:<br>
Setting the sound level can be done by:


It should print that the camera is open and connected to the robot (bidge).
amixer set PCM 0dB unmute


====Software structure====
Setting to maximum (100%) is about the same as setting 4dB (may be too much for amplifier).


The software is structures as shown in figure 1 below.
amixer set PCM 100% unmute


[[file:mission-software.png | 500px]]
===Regbot GUI===


Figure 1. The example mission software intended as basis for full mission control. In the main.cpp file, three objects are created: bridge (UBridge reg(127.0.0.1)), camera (UCamera cam()) and mission (UMission mission(&reg, &cam)). The bridge handles communication and stores all data from the robot and the gamepad. The camera can capture images and image processing is intended here. The mission object is intended for splitting long missions into mission snippets that the REGBOT can execute.
To run the regbot gui on the raspberry the following package is needed


The mission class has access to robot data and the camera at all times using the pointers "bot->" and "cam->". e.g. the heading (relative to the start position) and if the yellow gamepad button is pressed, can be fetched as:
Additional python packages used


float heading = bot->pose.h;
* python-pyqtgraph
bool yellowButton = bot->joy.button[3];
* python-qt4-gl
* python-serial
* python-scipy


The sound system can be used for debugging, e.g. add a C++ line like:
sudo apt install python-pyqtgraph python-qt4-gl python-serial python-scipy


system("espeak \"bettina reached point 3\" -a30 -s130");


This line makes the robot say "bettina reached point 3" the parameters "-a30" turns amplitude down to 30%, and "-s130" makes the speech a little slower and easier to understand.
=====* no python-pyqtgraph =====
It requires that espeak is installed (sudo apt install espeak).


===USB as SOCKET for REGBOT client===
On my 14.04 32 bit ubuntu the python-pyqtgraph package was not found using apt-get, so I


This is not compatible with use of ROBOBOT_BRIDGE, so don't use it here
downloaded from http://www.pyqtgraph.org/
the debian package ''python-pyqtgraph_0.9.10-1_all.deb'' and installed using:


To use the REGBOT client through (not on) the raspberry pi, the serial connection /dev/ttyACM0 on the raspberry can be converted to a network port by SOCAT. So install:
sudo dpkg -i python-pyqtgraph_0.9.10-1_all.deb


  sudo apt install socat
  sudo apt install python-pyqtgraph


Add the following line to /etc/rc.local
To start the GUI


  socat TCP-LISTEN:24001,fork,reuseaddr FILE:/dev/ttyACM0,raw,echo=0 &
  cd ~/regbot/qtgui
python regbot.py


This creates a TCP socket server, listening to port 24001 and piping /dev/ttyACM0 to the socket (with no local echo). This only takes the data from the /dev/ttyACM0 when a client is connected to the socket.
Connect to network localhost or kill the bridge and connect directly to "/dev/ttyACM0" (not com5 as is the default serial connection).


Works after a reboot, or the same line on the command prompt.
== File sharing for development ==


This is good for configuring the REGBOT (in REGBOT client connect wifi to IP of robot, e.g. 192.168.0.2).
Mirror the local disk on the Raspberry to a PC, make all file editing here, and then on a terminal connection to the robot to compile the software.


===Share files from windows===
===Share files from windows===
Line 152: Line 214:
===Share files on Linux===
===Share files on Linux===


On the Linux PC use sshfs for sharing.
On a Linux PC use sshfs for sharing.
Make an empty directory for the mapped disk, and then mount
Make an empty directory for the mapped disk, and then mount



Latest revision as of 13:55, 24 December 2020

Back to Robobot


Installation instruction on the raspberry

Install raspberry pi version of Linux on a micro-SD, instructions can be found here:

Install on raspberry first section (prerequisites).

Opencv 3.4

It should be OK just to install libopencv-dev - in December 2020 this installs version 3.2.0

sudo apt install libopencv-dev

To find the installed version of OpenCV try

pkg-config --modversion opencv

To install newer OpenCV version 3.4 (as of Jan 2020 it is version 3.4.9)

cd
mkdir -p git
cd git
# this method will install version 4.2 - that is not fully compatible (jan 2020)
# git clone https://github.com/opencv/opencv.git
# git clone https://github.com/opencv/opencv_contrib.git
# use this to get a version compatible with opencv 3.4
# get source (jan 2020)
wget -O opencv-3.4.zip https://github.com/Itseez/opencv/archive/3.4.zip
unzip opencv-3.4.zip
wget -O opencv_contrib-3.4.zip https://github.com/Itseez/opencv_contrib/archive/3.4.zip
unzip opencv_contrib-3.4.zip
# compile the lot, but ommit python3 support, as the compile will fail (as of jan 2020)
cd opencv
mkdir build
cd build
# create makefiles, also for all modules in the opencv_contrib (e.g. ArUco code detect)
# cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_opencv_python3=OFF -D OPENCV_EXTRA_MODULES_PATH=~/git/opencv_contrib/modules ..
cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_opencv_python3=OFF -D OPENCV_EXTRA_MODULES_PATH=~/git/opencv_contrib-3.4/modules ..
# inspect printout to see if needed features are enabled
# use -j2 option not to overload or overheat the processor (j2 means use 2 cores only)
# this probably takes more than an hour.
# if the compilation crashes about halfway or so, it is probably due to lack of memory
#    then just start the compile again, and it is likely to continue (at least for a while)
make -j2
# install into /usr/local
sudo make install
# load environment settings
sudo ldconfig

Raspicam

NB! this should not be needed

A rather short raspberry pi camera API is available from https://www.uco.es/investiga/grupos/ava/node/40. And can be downloaded from https://sourceforge.net/projects/raspicam/files/,

or get the file from sourceforge with no GUI, with:

cd ~/Downloads
wget --no-check-certificate -O raspicam-0.1.9.zip https://downloads.sourceforge.net/project/raspicam/raspicam-0.1.9.zip?r=https%3A%2F%2Fsourceforge.net%2Fprojects%2Fraspicam%2F%3Fsource%3Dtyp_redirect&ts=1486483484&use_mirror=netix

Once fetched and available on the raspberry, then Unpack and install:

unzip raspicam-0.1.9.zip
cd raspicam-0.1.9
mkdir build
cd build
cmake ..
make
sudo make install
sudo ldconfig

The library is installed in /usr/local, and therefore to make cmake find it the path needs to be added to startup configuration, in file ~/.bashrc

nano ~/.bashrc
add at the end:
export CMAKE_PREFIX_PATH=/usr/local/lib

(needs a re-logon to be activated, or repeat on the command line)

DNSMASQ

To enable the ROBOBOT to be connected directly to a PC, then it is easier if the ROBOBOT provides an IP for the PC.

Install DNSMASQ

sudo apt install dnsmasq

allow dnsmasq to provide IP to pear-to-pear networks Edit the /etc/dnsmasq.conf (nano is a small text editor, fine for editing configuration files owned by Linux root)

sudo nano /etc/dnsmasq.conf
find and change/add the following line (a bit down, eg to serch use: ctrl-W dhcp-range)
dhcp-range=192.168.17.100,192.168.17.150,12h

To work, the eth0 must have an IP, if noone provides one, change /etc/dhcpcd.conf (dhcp client deamon) to have a default IP, if no DHCP server is available

Edit of /etc/dhcpcd.conf to have a fall back behaviour for eth0 and eth1.

sudo nano /etc/dhcpcd.conf
Add the following lines at the end (uncomment and modify):
# define static profile
profile static_eth0
static ip_address=192.168.17.1/24
# static routers=192.168.0.1
# static domain_name_servers=192.168.0.1

# fallback to static profile on eth0
interface eth0
fallback static_eth0 

This will also give a default IP for an eventual second network (usb-to-cable) if needed

Now, after a reboot, you should be able to connect a PC directly with at network cable to a PC, and the PC should get an IP from the raspberry, so now

ssh local@192.168.17.1

Sound

festival

Easy understandable voice, but takes time to generate and require two intermediate files

sudo apt install festival
echo "The brown fox jumps over lazy dog" > aaa.txt
text2wave aaa.txt -o aaa.wav
aplay aaa.wav

Music files

This part works

Play sound files, like

sudo apt-get install sox libsox-fmt-all
play -v0.1 music.mp3

This line plays an mp3 file with volume reduced by a factor 0.1. The at startup the blue button on the gamepad plays the file in /home/local/Music/music.mp3.
This is a symbolic link to some real music (Radetzky Marsch in this case):

ls ~/Music
   radetzky-marsch_Schloss-Schoenbrunn-Konzerte_Wien_full-length.mp3
cd ~/Music
ln -s radetzky-marsch_Schloss-Schoenbrunn-Konzerte_Wien_full-length.mp3 music.mp3

Sound level

On a Raspberry pi version 3+ (not tested on other versions)

$ amixer
 Simple mixer control 'PCM',0
   Capabilities: pvolume pvolume-joined pswitch pswitch-joined
   Playback channels: Mono
   Limits: Playback -10239 - 400
   Mono: Playback 400 [100%] [4.00dB] [on]

An thus:
Setting the sound level can be done by:

amixer set PCM 0dB unmute

Setting to maximum (100%) is about the same as setting 4dB (may be too much for amplifier).

amixer set PCM 100% unmute

Regbot GUI

To run the regbot gui on the raspberry the following package is needed

Additional python packages used

  • python-pyqtgraph
  • python-qt4-gl
  • python-serial
  • python-scipy
sudo apt install python-pyqtgraph python-qt4-gl python-serial python-scipy


* no python-pyqtgraph

On my 14.04 32 bit ubuntu the python-pyqtgraph package was not found using apt-get, so I

downloaded from http://www.pyqtgraph.org/ the debian package python-pyqtgraph_0.9.10-1_all.deb and installed using:

sudo dpkg -i python-pyqtgraph_0.9.10-1_all.deb
sudo apt install python-pyqtgraph 

To start the GUI

cd ~/regbot/qtgui
python regbot.py

Connect to network localhost or kill the bridge and connect directly to "/dev/ttyACM0" (not com5 as is the default serial connection).

File sharing for development

Mirror the local disk on the Raspberry to a PC, make all file editing here, and then on a terminal connection to the robot to compile the software.

Share files from windows

On the windows computer install "winscp" and use it to copy files to and from the raspberry disk as needed.

There are other methods too.

Share files on Linux

On a Linux PC use sshfs for sharing. Make an empty directory for the mapped disk, and then mount

mkdir robobotdisk
sshfs local@192.168.0.2: robobotdisk

Unmount with

fusermount -u robobotdisk

Or use "sudo unmount robobotdisk"