Duckiebot zelfrijdende auto

Duckiebot zelfrijdende auto

MIT-project met Computervision

Bouwen van de zelfrijdende MIT Duckietown Auto's. Samen met een aantal enthousiastelingen bouwt CreativeHubs de zelfrijdende auto van MIT Duckiebot genaamd....

Aanleiding voor dit project van zelfrijdende auto's komt uit een brainstorm avond met Volker Infra. Zij zijn op zoek naar nieuwe technologieen en visies op de toekomst van ons vervoer.

Vervolgens is een groep bouwers, waaronder CreativeHubs, aan de slag gegaan met het bouwen van Duckietown Auto's van MIT. Deze autos beschikken over visuele mogelijkheden om banen op een weg te volgen en obstakels te omzeilen....


 

Follow lines:

Around the corner:

BUILDING OF DUCKIETOWN BOT

CHECKLIST VRROOM
https://docs.google.com/document/d/1L3xzNKYlcETUcaJawpp4PtKPBKXncKqEOChT4OhjVEs
CHECKLIST DUCKIETOWN:
https://docs.google.com/document/d/1UgCuAcE0WJKja-y0WTXR8pFo9f3srQ4TNtAiX8Nhp08/edit#
CHECKLIST DUCKIETOWN-BUNNY:
https://www.duckietown-bunny.com/tutorials


HARDWARE

STORAGE (around € 250,00 incl VAT and shipping)
Exclusive HDMI-cable, keyboard, mouse, screen, needed for the first setup

1 x RPI
https://www.kiwi-electronics.nl/raspberry-pi/board-and-kits/raspberry-pi-3-model-b
1 x RPI camera
https://www.kiwi-electronics.nl/raspberry-pi-camera-board-v2-8mp?search=pi%20cam
1 x RPI camera behuizing
https://www.kiwi-electronics.nl/raspberry-pi-camera-behuizing-transparant?search=pi%20cam
1 x DC STEPPER MOTOR HAT
https://www.kiwi-electronics.nl/adafruit-dc-stepper-motor-hat-voor-raspberry-pi?search=DC%20hat
1 x SERVO HAT 16 CHANNEL
https://www.kiwi-electronics.nl/adafruit-16-channel-pwm-servo-hat-voor-raspberry-pi-mini-kit?search=servo%20hat
2 x GPIO HEADER
https://www.kiwi-electronics.nl/40-pin-gpio-stapel-header-voor-raspberry-pi-2x20?search=gpio
1 x JUMPWIRES
https://www.kiwi-electronics.nl/jumperwires-bundel?search=gpio
4 x DISTANCE HOLDERS HAT
https://www.kiwi-electronics.nl/messing-m2-5-afstandhouders-11mm-hoog-zwart-2-pack?search=afstand
1 x USB to DC cable
https://www.allekabels.nl/usb-adapter/1724/1084098/usb-power-kabel-5.5mm-x-2.1mm.html?gclid=CNn9zLz9ldMCFcfgGwod3bQK5A
1 x 32 GB SDCARD
1 x USB 32 GB FLASH DRIVE

https://nl.aliexpress.com/item/100-Original-Genuine-Sandisk-Glide-mini-USB-3-0-Flash-Drive-up-to-130m-s-SD/32400110857.html?source=%7Bifdyn:dyn%7D%7Bifpla:pla%7D%7Bifdbm:DBM&albch=DID%7D&src=google&albch=shopping&acnt=494-037-6276&isdl=y&aff_short_key=UneMJZVf&albcp=664365076&albag=32654332494&slnk=&trgt=87981925924&plac=&crea=nl32400110857&netw=g&device=c&mtctp=&gclid=CI268I-CltMCFeQW0wod8fsMiw
1 x CAR KIT
https://www.aliexpress.com/item/Smart-robot-car-chassis-kit-in-other-electronic-components-Speed-Encoder-Battery-Box-For-Arduino-Free/32603428102.html?spm=2114.40010608.4.28.MTqa3F
1 x SOLAR POWER HUB
https://www.solarpowersupply.nl/productgroepen/solar-chargers/xtorm-lava-solar-charger-am114

ROAD:
black foam board
Isolation tape in RED, BLUE, WHITE and YELLOW


SOFTWARE

CLEAN UBUNTU IMAGE

http://ubuntu-mate.org/download/

choose: Ubuntu MATE 16.04.2 LTS for Raspberry Pi 2 and 3 systems

 

BURN UBUNTU IMAGE ON SD-CARD

ON MAC: brew install xz

WINDOWS: install Win7

xz -d ubuntu-mate-16.04.2-desktop-armhf-raspberry-pi.img.xz

sudo dd if=/Applications/MAMP/htdocs/ubuntu-mate-16.04.2-desktop-armhf-raspberry-pi.img of=/dev/rdisk2 bs=4m

Or use ApplePiBaker to write RPI-images easily

 

FOLLOW UBUNTU INSTRUCTIONS

Name: {your own}

Computername: alphaduck

Username: ubuntu

Password: ubuntu

 

Type the command: sudo raspi-config

Change settings:

- expand filesystem

- enable i2c

- enable camera

- enable SSH

 

Type the command: sudo apt-get update

Type the command: sudo apt-get dist-upgrade

 

Type the command: sudo reboot

 

SSH:

ssh ubuntu@alphaduck.local

pass: ubuntu

 

CLEAN ROS (this may take a while)

http://wiki.ros.org/ROSberryPi/Installing%20ROS%20Kinetic%20on%20the%20Raspberry%20Pi

choose: full install and follow instructions till step 2.3

 

MAKE SD-BACKUP IMAGE ON YOUR OWN SYSTEM

————————————————————————————————————

 

INSTALL LATEST DUCKIETOWN DIRs

https://github.com/duckietown/Software.git

Upload to rpi in dir: duckietown

Type the command: cd ~duckietown

Type the command: sudo dpkg --configure -a

Type the command: sudo apt-get -f install

Type the command: sudo pip install graphviz

Type the command: sudo apt-get install python-sklearn

 

BUILD ROS CATKIN

Type the command: cd ~/duckietown

Type the command: source environment.sh

Type the command: source /opt/ros/kinetic/setup.bash

Type the command: make build-catkin
NOTE: this may take a while… and when it stops more then 5 minutes at 94%, quit process and use the solution beneath. After that, repeat this process

 

IF ERROR (around 94%)  

FIX:

Type the command: sudo dd if=/dev/zero of=/swap0 bs=1M count=512

Type the command: Type the command: sudo mkswap /swap0

sudo nano  /etc/fstab

at the end paste: /swap0 swap swap

Type the command: sudo swapon -a

Type the command: make -j4 -l4 build-catkin

 

CREATE NEW MACHINES-FILE:

Type the command:
sudo nano /home/ubuntu/duckietown/catkin_ws/src/00-infrastructure/duckietown/machines

And paste the next code

 

<launch>

   <arg name="env_script_path" default="~/duckietown/environment.sh"/>

   <machine name="alphaduck" address="alphaduck.local" user="ubuntu" env-loader="$(arg env_script_path)"/>

</launch>

 

Change “alphaduck” to your own veh.name, or use this one for easy use.

CTRL+X and Y to save this machines-file.

 

BEFORE TESTING THE JOYSTICK
Type the command: sudo apt-get install ros-kinetic-joy
Type the command: sudo chmod 777 /dev/input/js0

 

TEST THE JOYSTICK AND WHEEL-TURNS

Type the command: source environment.sh

Type the command: source set_ros_master.sh

Type the command: roslaunch duckietown joystick.launch veh:=alphaduck

 

This is the expected result of the joystick:   

left joystick up = forward

left joystick down = backward

right joystick left = turn left (positive theta)

right joystick right = turn right (negative theta)

 

WHEELS CALIB

STRAIGHT

rosservice call /alphaduck/inverse_kinematics_node/set_trim -- -0.01

SPEED

rosservice call /alphaduck/inverse_kinematics_node/set_gain -- 1.1

SAVE WHEELS CALIB

rosservice call /alphaduck/inverse_kinematics_node/save_calibration

 

CAMERATEST

Type the command: raspistill -t 100000 -o out.jpg

Type the command: roslaunch duckietown camera.launch veh:=alphaduck

 

CAMERA CALIB

Type the command: sudo apt-get install ros-kinetic-camera-calibration

 

Via RPI

Connect screen HDMI

Type the command: roslaunch duckietown intrinsic_calibration.launch veh:=alphaduck

 

Turn duckiebot around chessboard in several positions till button calibrate shows up.

Than click Commit

 

Type the command: roslaunch duckietown camera.launch veh:=alphaduck

Type the command: roslaunch ground_projection ground_projection.launch veh:=alphaduck local:=1

 

Via SSH

Type the command: source environment.sh

Type the command: source set_ros_master.sh alphaduck

Type the command: rostopic list or rosservice list

 

Type the command: rosservice call /alphaduck/ground_projection/estimate_homography

 

VIA RPI

roslaunch duckietown camera.launch veh:=alphaduck

rosrun ground_projection test_projection.py alphaduck

 

rosrun image_view image_view image:=/alphaduck/camera_node/image_mono

rosrun image_view image_view image:=/alphaduck/camera_node/image_rect
 

DEMOS VIA RPI

Type the command: sudo nano Makefile

make XXX (XXX = the name of the demo in the makefile)

 

EXAMPLE

Type the command: make demo-joystick-camera

Type the command: make demo-line_detector

 

And try to run something by clicking your Joystick…

 

FINALLY DRIVE!!!

 

Via SSH

Type the command: cd ~/duckietown

Type the command: source environment.sh

Type the command: source set_ros_master.sh alphaduck

Type the command: make openhouse-dp1

 

Move duckiebot to the camera calibration zone

Press RB then Y to start camera calibration

Press RB again to end verbose mode

 

Initially, it will be in joystick mode. Press LB to toggle between joystick and parallel autonomy.

 

Drive = Joystick START

Stop = Drive = Joystick BACK

 

Or run these separately:

Type the command: make openhouse-dp2-vehicle

Type the command: make openhouse-dp2-obstacle

x
This website is using cookies. Read more... I accept cookies
Creativehubs
X
Add to home screen
Add to home screen