Erle Robotics: "Linux Drones & the App Store for Drones We - - PowerPoint PPT Presentation

erle robotics
SMART_READER_LITE
LIVE PREVIEW

Erle Robotics: "Linux Drones & the App Store for Drones We - - PowerPoint PPT Presentation

Erle Robotics: "Linux Drones & the App Store for Drones We make artificial robotic brains that power robots and drones. Today with you Vctor Mayoral Vilches Alejandro Hernndez Iigo Muguruza Robotics Robotics Robotics


slide-1
SLIDE 1

Erle Robotics: 


"Linux Drones & the App Store for Drones”

slide-2
SLIDE 2

We make artificial robotic brains that power robots and drones.

slide-3
SLIDE 3

Today with you

Víctor Mayoral Vilches

Robotics Engineer

Alejandro Hernández Iñigo Muguruza

Robotics Engineer Robotics Engineer

slide-4
SLIDE 4

Erle-Brain

A Linux-based computer for making robots and drones

Erle-Brain is an artificial Linux brain for making robots. It puts together more than 25 sensors, a 1 GHz Cortex-A8 processor and 4 GB of flash memory running state of the art algorithms.

Brains

slide-5
SLIDE 5

Erle-Brain

Buzzer: piezoelectric device for sound emitting UART serial: GPS, serial console, additional sensors, … Power module support: Power Erle-Brain and all the sensors through traditional Power Modules I2C bus: How about connecting a humidity sensor

  • r maybe a H2O one ;)?

1 GHz processor running Linux distributions, several IMUs (3 axis accelerometers, 3 axis magnetometers, 3 axis gyroscopes), pressure sensors, temperature sensors, … PWM signals: Motors control, PPM (radio) reception, servo motor control, …

Brains

slide-6
SLIDE 6

Robots overview

Erle-HexaCopter

12.2.2015

Erle-Rover

20.03.2015

Erle-Copter

1.1.2015

Erle-Plane

07.3.2015

Robots

slide-7
SLIDE 7

?

Robots overview

Erle- HexaCopter Ubuntu drone

5.5.2015

Erle-Copter Ubuntu Core

5.5.2015 30.7.2015*

Robots

?

slide-8
SLIDE 8

RGB LED: Light your environment with an Erle- shaped light Buzzer: piezoelectric device for sound emitting Li-Po battery Radio Control Link (2.4 GHz)
 A traditional RC way of securing your flights. Telemetry link (433 / 915 MHz) Dedicated link for vehicle information and mission data GPS 
 UBlox 7M and magnetomer that includes a patch antenna Flight computer 
 A1 GHz Cortex-A8 processor that runs all the flight logic in Linux Power Module
 A standard way of powering your drones with batteries from different sizes. WiFi (2.4 / 5 GHz)
 We support 802.11 ac both in ad hoc (hotspot) and infrastructure modes. 30+ sensors 
 accelerometers, magnetometers, gyroscopes, temperature sensors, barometers, …

Anatomy of a Linux drone

slide-9
SLIDE 9

Anatomy of a Linux drone

slide-10
SLIDE 10

Erle-Brain+

Brains

the smartest brain for drones

slide-11
SLIDE 11

Robotics has long been a field that overpromised and underdelivered. We think drones are the most likely way to rectify that.

  • C. Dixon
slide-12
SLIDE 12

Why now?

slide-13
SLIDE 13
slide-14
SLIDE 14

Last 30 days at Dronecode:

– 150,000 downloads of the Copter code 
 – 30,000 downloads of the Plane code 
 – 300,000 downloads of the GCS 
 – 400,000 downloads of terrain data

– data on 12-15,000 flights per day
 – 200,000+ systems worldwide
 –Growth of 10,000 systems per month

slide-15
SLIDE 15

Linux

SDK: ROS

App Store

slide-16
SLIDE 16

Linux drones and robots

slide-17
SLIDE 17

Worldwide community

Out of the 3B Internet users 80M use Linux and 25M of them Ubuntu.

80M

Linux users

slide-18
SLIDE 18

Stable OS built over the last 20 years

slide-19
SLIDE 19

Embedded computers

slide-20
SLIDE 20

Demo 
 drone webstore

SSID: Erle-Brain+ try erlerobotics.com

slide-21
SLIDE 21

A toolbox for creativity

slide-22
SLIDE 22

Demo 
 real time visual targeting

slide-23
SLIDE 23

Software autopilot

a.k.a. ardupilot

  • 700,000 loc
  • 17,000 commits
  • 136 commiters
  • 800+ contributors
  • HALs
  • APMrover2
  • AntennaTracker
  • ArduCopter
  • ArduPlane
  • Tools
  • docs
  • libraries
  • AC_AttitudeControl
  • AC_Fence
  • AC_PID
  • AC_Sprayer
  • AP_HAL
  • AP_HAL_Linux
  • mk
  • modules
slide-24
SLIDE 24

Real time

Maximum latency Task

100 ns SPI bus transitions 1 us PWM transitions, PPM-SUM and SBUS 1 ms IMU sensor inputs (gyroscopes and accelerometers) 20 ms Barometer, compass, airspeed, sonar (I2C, SPI and analog). 200 ms GPS and other sensors

slide-25
SLIDE 25

Real time

Kernel latencies Thread priorities

  • vanilla kernel
  • PREEMPT
  • RT_PREEMPT
  • Xenomai

APM application threads are configured with with the scheduling policy SCHED_ FIFO with priorities that assure meeting the required maximum latencies. Set up has proved to be sufficient for among hundreds of flights.

slide-26
SLIDE 26

Demo 
 flying while compiling

slide-27
SLIDE 27

SDK:
 Robot Operating System (ROS)

slide-28
SLIDE 28

Robot Operating System

An SDK for robots and drones

The Robot Operative System (ROS) is an open-source, framework for robot application development maintained by the Open Source Robotics Foundation (OSRF). A ROS system is comprised of a number of independent nodes, each of which communicates with the other nodes using a publish/subscribe messaging model that can be deployed over different computers.

slide-29
SLIDE 29

Robot Operating System

LAMP

Linux
 Apache
 MySQL
 PHP (Python)

slide-30
SLIDE 30

Robot Operating System

slide-31
SLIDE 31

/accel Accels Gyros /gyro Mags /mag AHRS /euler /quater motor1 /pwm1

slide-32
SLIDE 32

Demo 
 ROS music

> rosrun ros_erle_buzzer ros_erle_buzzer_writer <melody>

slide-33
SLIDE 33

Demo: ROS music

/buzzer buzzer

> rosrun ros_erle_buzzer ros_erle_buzzer_main > rosrun ros_erle_buzzer ros_erle_buzzer_writer <melody>

slide-34
SLIDE 34

Demo 
 ROS light

> rosrun ros_erle_ubled ros_erle_ubled_writer <color>

slide-35
SLIDE 35

Demo: ROS light

/ubled ubled_main

> rosrun ros_erle_ubled ros_erle_ubled_main > rosrun ros_erle_ubled ros_erle_ubled_writer <color>

slide-36
SLIDE 36

Demo 
 Autonomous flight

slide-37
SLIDE 37
slide-38
SLIDE 38

http://answers.ros.org/question/198576/ar_pose-doesnt-recognize-tag-in-tum_simulator-please-help/

slide-39
SLIDE 39

The next thing
 ROS 2.0

slide-40
SLIDE 40

The next thing
 ROS 2.0

  • lossy networks
  • embedded and deep

embedded (bare-metal)

  • teams of multiple robots
  • real time
  • production environments
slide-41
SLIDE 41

App store

slide-42
SLIDE 42

The first app store for robots and drones in the cloud supported by Canonical and the Open Source Robotics Foundation.

An app store for robots

App Store

slide-43
SLIDE 43

Sneak peek

App Store

slide-44
SLIDE 44

Demo 
 packing a ROS app

> ./ros2snap ros_erle_takeoff_land

slide-45
SLIDE 45

Sneak peek

App Store

slide-46
SLIDE 46

Sneak peek

App Store

slide-47
SLIDE 47

Sneak peek

App Store

slide-48
SLIDE 48

Sneak peek

App Store

slide-49
SLIDE 49

Sneak peek

App Store

slide-50
SLIDE 50

“The best way to predict the future is to invent it” Alan Kay

slide-51
SLIDE 51

Contact

erlerobot

Erle Robotics S.L.

contact@erlerobot.com twitter.com/erlerobotics http://erlerobotics.com facebook.com/erlerobot