Limits to Automation
13 views
Skip to first unread message
Nathan Cravens
Jul 20, 2011, 5:13:32 PM7/20/11
to Open Manufacturing, Abundance, Paul D. Fernhout, Eric Hunting, Kevin Carson, doug rushkoff, Chris Anderson
We've had this discussion before, but it may be fun to revisit it.
http://www.bbc.co.uk/iplayer/episode/b012fs74/The_Bottom_Line_Limits_of_Automation/
'
This week Evan asks his panel of top executives about the limits of automation. How far can they go in removing human beings from their business? Which processes are beyond automation? The panel also swap thoughts on the benefits of the corporate awayday.
Evan is joined in the studio by Mike Lynch, founder and chief executive of the software company Autonomy; Colin Drummond, chief executive of waste management firm Viridor; Douglas Anderson, president and chief executive of the global travel management company Carlson Wagonlit Travel.
'
At the start of the interview, Mike Lynch mentions Aurasma (augmented reality software) to identify objects in the user environment, display how they are made and where to locate materials to make it, relates to previous discussions.
It would be wise to develop and have available an open source augmented reality software for survival: identify edible plants and ways to prepare them, location of nearest fresh water source, suggestions on how to improve or rebuild a makeshift structure that can keep a comfortable temperature and remain dry. It could be called 'resource management at the bottom'.
It would show communal spaces like freeshops (ex: cloths or food), tool and materials sheds (ex: bicycle or motorcycle maintenance), nearest skillsharing events (Facebook presently does a good job in informing and coordinating events), free (closed and open) communal or individual living spaces, ect.
On something similar to Google maps, markers will be placed for all the free stuff, such as fruit and nuts from trees in the area with user generated images of each tree and its surroundings. This can already be done using Google maps, of course.
http://en.wikipedia.org/wiki/Augmented_reality#Open_source_software
'
Open source software
- ARToolKit, an open source (dual-license: GPL, commercial) C-library to create augmented reality applications; was ported to many different languages and platforms like Android, Flash or Silverlight; very widely used in augmented reality related projects
- ArUco, a minimal library for augmented reality applications based on OpenCv; licenses: BSD, Linux, Windows
- mixare, Open-source (GPLv3) augmented reality engine for Android and iPhone; works as an autonomous application and for developing other implementations
- OpenMAR, Open Mobile Augmented Reality component framework for the Symbian platform, released under EPL; website is down but there is some information here
- Argon, Augemented reality browser by Georgia Tech's GVU Center that uses a mix of KML and HTML/JavaScript/CSS to allow developing AR applications; any web content (with appropriate meta-data and properly formatted) can be converted into AR content; currently available only for the iPhone, website is down
- Goblin XNA, BSD licensed, a platform for research on 3D user interfaces, including mobile augmented reality and virtual reality, with an emphasis on games. It is written in C# and based on Microsoft XNA Game Studio 3.1
- PTAM, Non-commercial use only
'
Just as augmented reality can help engage people with the physical world, we have the ability to have the hardware do the exercise with an ecology that will include autonomous vehicles, mass production facilities, materials storage, and personal robotics for personal fabrication.
So if you want a bicycle you could send a personal robot, perhaps something like a Turtlebot or Segway platform with two Shadowhands and Barrett WAM arms to an autonomous vehicle that would transport the personal robot to all the materials locations, collect the parts to have them assembled by your robot at home. That description takes renewable energy and long lasting batteries from abundant/recyclable materials for granted. It could do all the work silently or it could tell and show what it was doing, a process Mike Lynch briefly mentions, identifying each part and putting it in the right place with the right tool, or it could teach you how to build by giving you verbal and visual instructions.
Turtlebot
Segway Robotics
Barrett WAM arm
Shadowhand
The fusion of design models above hop into a...
Autonomous Vehicle
...using software based around or related to these areas:
Product Lifecycle
Material Science
Computer-aided Design
Hackerspaces
A good place for...
Knowledge sharing
Freeshop
Personal Fabrication
Community Kitchen
Passing Clouds Model
More automation will continue for market maximizing purposes: make it better, cheaper, faster, closer (once labor in the east is exhausted). Robotics like Makerbot and Turtlebot will ensure they develop for more personal, community, and other positive reasons. People will continue to be pushed out of work and become more dependent on the community and the state. To speculate, in 20-30 years (in westernized areas) it will become clearly unjust to distribute space and material rights using money, since near full automation ensures a very small monied and ownership minority. Japan will have an elderly crisis into the 2020s, ensured for cultural reasons, will have robots do the care rather than loosen immigration policy. Development on Robot Operating System continues to double. So it seems a safe speculation to say people will feel the monetarily distribution of space and material is unjust. The private leaning state (US model) will fill those pushed out into the military or prison and those of the welfare leaning state (UK model) into healthcare and counseling and 'council housing'. That is, of course, what is already happening, but when the structure changes for more free/open elements it will be because enough people within a territory refuse closed/proprietary methods and demand more free/open options.
Mass production will lessen with more capable free/open personal robotics fabrication and free/open local materials channels with the reduction of market and state to follow. There is more than enough ability for all of us to live well without income (market) or taxation (state). To do so requires a change of value toward human life being of higher value than market exchange.
I live with a small group of friends on squatted green land in self-built structures. We live off the waste of West London. We could live off what grows wild around us, but there is no need to let tasty treats go to Colin Drummond's facilities.
Nathan
Nathan Cravens
Jul 28, 2011, 1:02:27 PM7/28/11
to Eric Hunting, Open Manufacturing, Abundance, Paul D. Fernhout, Kevin Carson, doug rushkoff, Chris Anderson, Samuel Rose, Smári McCarthy
Central vs. Distributed Device Control
Eric, you may be interested in Daniel Mellinger's description of the mix of internalized (distributed) and externalized (centralized) control systems based on his work with quadrotors. The interview begins 32 minutes into the podcast.
Real Spacial Internet Protocol (RSIP)
Each block of physical 3D space needs an Internet Protocol address. This would enhance GPS systems for real-time locating systems. Each space would be measured like an invisible cube with a code associated with each cube and objects within the grid. Each spacial IP address would represent latitude, longitude, distance above or below a surface, object features within the space, and other important information a retrieval device or person will need to know about the terrain and conditions of the area to deploy the right retrieval vehicle(s) for ground, air, or aquatic travel.
Gray Snappy Brambles
To aid imagination, the fire pit at my camp will have an RSIP of S1. When images of that site are captured by a device, 3D spacial information is represented by specific coordinates to visit and manipulate the location based on the simplified machine readable coordinates associated with S1. This assumes a fair amount of object recognition by the device, such as object avoidance, to go around a tree that may be in the direct path to an object location. RSIP would not be a conclusive AI package, but a universal spacial recognition standard to aid robot/materials ecologies and curious hands-on people.
The RSIP label enables a device to have spatial recognition by quickly locating specific places where objects are known to be found, like firewood for the fire pit at S1. Dead gray snappy brambles are good to boil the kettle, as they burn easily and for long enough to boil water. A small wheeled vehicle with manipulators able to grip various hand tools (or when wheels fail to meet the terrain) and a trailer hitch with a wheelbarrow behind it whizzes off to retrieve brambles. The device will know where to look for gray snappy brambles because it will have regularly scanned the campsite area. It will recall from its last scan the largest and closest areas for gray snappy brambles were S10 and S15 and goes immediately to those areas, snapping the brambles longer than the trailer into pieces, placing them into the trailer, and back to the fire pit for delivery.
Present methods of object location without the aid of RFID scan an entire space top-to-bottom much like a flatbed scanner until the right object is found. Such a process is computationally wasteful and slow. If the space itself has a code, including object code within the space, devices can learn common sense via a ranking system that associates paintings with being on walls, as RSIP code will need to account for walled surfaces. The spacial code in practice may be thousands of characters when spacial coordinates are needed for manipulating microscopic objects.
Nathan
Reply all
Reply to author
Forward
0 new messages