Seoul Tech Society

Intro

Daniel Stonier / d.stonier@gmail.com

Yujin Robot
Australia
Tux Robot

Who Am I?



Yujin Robot Innovation Team co-leader.

Embedded/Control Engineer
Algorithms Monkey
Software Architect

B Science., B Electrical Engineering, Phd. Mathematics.

Who Am I?

Math Geek Arrow Terminator

Who Am I?

Math Geek Arrow Replacement

Contents

  • Where are we with Robotics?
  • Who is doing what?
  • ROS/OSRF and the Robotics Community
  • Why is it so damned difficult challenging?
  • Some stories
  • Do you have what it takes?
  • Forging a new business
  • Yujin's toys

Where Are We?


Been trying to get unstructured since the 60's.


At least two waves of venture capital
...early 90s and 00s


Gocart Sketch
Crashed

Riding the New Wave


Technology Convergence

GPS, 3d sensors, embedded cpu power, batteries, web services, cloud, handhelds, IoT


Hardware to Software Shift

Suddenly people writing alot more on software than hardware
Remember the PC sw revolution?


America Interested in Nuts & Bolts Again

And when they move, they move fast.


-> Venture Capital!!!

Who is doing what?


Recent Success Stories


Warehouse Logistics - Khiva @ Amazon

Manipulation & Mobility - Google Acquisitions

Automated Tractors - Many

Automated Trucks - Australian Mining

Automatic Cars - Google/Bosch/...

Hospitals - Atheon Tugs


This never happened surfing previous waves

What is ROS?

  • One of many robot middlewares that emerged in the 2000's
    • ROS, MS Robotics Studio, OROCOS, MOOS, OPROS...
  • 2007 - emerged from Stanford.
  • 2008 - migrated to Willow Garage.
  • 2012 - migrated to the Open Source Robotics Foundation.
  • Not owned by a company.
  • Not dependent on a university and it's funding imperatives.

Again, what is ROS?

Lets try again:


  • Build tools (cmake/python based)
  • Message protocol architecture
  • Multi-language communication libraries
  • Build farm for thousands of open source packages
  • Installation/deployment tools
  • Visualisation & debugging software
  • Robotic drivers
  • Algorithms
  • Strong community support

No, really, what is ROS?


It is becoming the defacto glue for robot software and hardware


Why do we need glue?


A robot is a complicated morass of sensors, drivers, pipelines...what if everyone did their own thing their own way?

What happens when you have glue?


IBM compatible PC for hardware, html browsers for the web, android for phones

Non-trivial explosions

Why ROS?


Lots of others in the past, so why ROS?


  • Very freeform
  • Focused on just helping people do their job
  • Lets you scale what you do, fast
  • Driven by the world's best reference service robot, pr2

Robotics Teams


Never enough people for the different jobs


Almost always have to become multi-skilled

Robotics is growing up, but still a young field


Hard to get experts in various areas (money/above), but these are starting to filter in

Interesting people get attracted to robotics


We've had a stage theatre actor, bank businessman, mathematician, space agency programmer

Flip side of being an immature field, often doing it for the love of it, not the money

Robotics People


You can and will meet the best and brightest of a generation....

but they're still idiots! :)


Leading C++ Boost developer deleted my linux root


Robotic Web Tools hangout FAIL

Robotics Product


Gocart Sketch

It's a long time getting to something real

But the buzz is in making it real


It's surprising how many research-like problems there are along the way.

Do You Have What it Takes?


You like waking up to challenges every day


Willing to sacrifice some money for your passion


You are not embarrased about smelling, talking to your robot!


Dirty jobs betwixt challenges are ok, it completes the picture


You like to be involved in the direction it takes

Why I Like It

It does it all:


Idea + Implementation + Product


New technology every day, every year


Hundreds of open doorways to explore

Forging a New Business

Gocart Sketch

The road to robot nirvana is littered with 'zombies'


Be wary of past traps

Hollywood expectations, engineering glasses, easy grant money

Get out there

Most successes come from companies with an existing tech/skill who then go out and find real world problems that people have.

How to compete against giants?

Move faster with disrupting solutions, fly under their radar with a niche product or plan on being bought up.

Yujin Innovation

Gophers

Early Gocart Concept

Gocart Concept Gocart

GoCart@IFA

GoCarting

Gocart Sketch Gocart
Rocon

Dorothy Cafe

Technology

THE END

By Daniel Stonier / d.stonier@gmail.com