[Robotics] Blender Free and Open Source Robot Worker (Will Jackson)

[robotuser118txu]

Will,

The answer is very simple - because it is more difficult to implement
a diverse set of robot workers instead of a singular (humanoid type)
one - it is more development-concentric, quite simply easier,
especially for a free and open source "project". We don't need a
humanoid robot worker that can make facial expressions (what japanese
engineers would lead us to believe). What is especially bad is to
believe that a mechanical / artificial representation of humanity
should be created - this is very bad because it is very unethical. A
machine is a machine, a machine that has been created by us to serve
our and the Earth's needs - that is it. Remember what I wrote about
the challenges and difficulties to open source development - we need
the most effective design, leave the niches to particular developers.

I am not suggesting that humans are adept at working at all tasks, but
their body is the most adept at a wide range of tasks. You could have
robots that instead of two legs hovered / flew through the air, but
where is the technology to implement it? That is why I made that
guideline - it is the most simple, effective, and most elegant of
solutions, if I may state that.

Like-wise, you could have three arms instead of two. But what would
that solve? Two mechanical and well-implemented arms will be
omni-directional (moving both forward and backwards - when needed, the
robot will quite simply reposition itself as effectively as possible
so as to allow it to achieve what it wants), so a third is not only
useless (for most tasks), but also ugly to look at.

Again, I am not foregoing the need for more specialized robots. But
for a free and open source "project", the simplest, most time-proven
and most effective / generalistic design, is a humanoid one.

Thomas

On Fri, Apr 30, 2010 at 2:07 PM, <robotics-request at blender.org> wrote:
>
>
>
> Message: 3
> Date: Fri, 30 Apr 2010 14:06:57 +0100
> From: Will Jackson <will at engineeredarts.co.uk>
> Subject: Re: [Robotics] Blender Free and Open Source Robot Worker
> To: Blender and Robotics <robotics at blender.org>
> Message-ID: <4BDAD5F1.9000705 at engineeredarts.co.uk>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> The notion of humanoid robot as worker is an odd one - there really is
> no good reason to do it - we are already surrounded by highly
> optimised 'robot' workers.
>
> There are however some reasons to build a humanoid robot that IMHO are
> valid - interaction with humans, entertainment and education.
> These are the areas where a human body model makes practical sense.
>
> There is an often repeated assertion that in order to operate in a human
> centric world humanoid robots are necessary - this is simply not true
> and is proven by all the functional robots
> already in use.
>
> Humans are actually poorly adapted to many manual tasks - which is why
> we have developed a large range of power tools to assist us, the essence
> of good design is suitability for purpose why start out with a design
> model that is unfit?
> I acknowledge a long standing and passionate desire to create mechanical
> / artificial  representations of humanity - but why?
>
>
> On 30/04/2010 13:49, robotuser118txu wrote:
> > Hello,
> >
> > I sent this email to Ton, and he directed me to this mailing list. My idea
> > is simple - to create a Blender robotics side of Blender. I say this because
> > I find Blender to complement this notion that I had very well - both appeal
> > to the same things, and both suit and complement each other perfectly. I
> > cannot call this a project because it would be too simplistic to call it so.
> > It's thus better to call it a notion - because that is what it is, a
> > movement, in its most basic (and only) form.
> >
> > Dear Ton,
> >
> > All in all I believe this to be very exciting. The beauty of it is that it
> > goes hand in hand with what you are trying to do. Believe it or not, this
> > has the potential of changing our world into one which does not need to
> > thrive in capitalism in order to survive. This complements, if not augments
> > and extends, what you are trying to do with the free and open source model,
> > which it will help immensely, as you can imagine why.
> >
> > You have seen the power that free and open source development has, with
> > Blender gaining strides with each day that passes.
> >
> > It only takes good will to do this, of which you have plenty. I developed
> > this notion because I believe that the free and open source model relates
> > significantly with a society that does not need to depend on money in order
> > to survive. I contacted you because I am impressed with what you have been
> > able to do with Blender, given too that I am a user and also because I see a
> > relation between Blender and the notion of a free and open source robot
> > worker. Please give it serious thought. In order to thrive, this project or
> > ambition needs a centralized direction, and you are the one that is best
> > capable of providing that, given how successfully, enthusiastically and
> > transparently you have managed Blender.
> >
> > There are many of us that wish that technology be used for the good of all,
> > and this along with Blender are going to be the most important developements
> > in human history, if I may take it lightly. Please tell me what you think of
> > it.
> >
> > Thank you for your help and consideration,
> >
> > - Thomas
> >
> > Free and Open Source General Robot Worker
> >
> > Layout and guideline
> >
> > Two principal tenets of one fundamental principal, movement:
> >
> > 1) Actuators
> >
> > 2) Programming
> >
> > 1) Image recognition
> >   2) Sound waves - sonar, echolocation
> > Objective is to minimize recognition errors in relation to 3d model, in
> > order to aid its movements as well as possible.
> >
> > - Create and recreate 3d map / model continously.
> >
> > Additional characteristics:
> >
> > - Light weight (aluminium, carbon fibre), soft body, moderate carrying and
> > handling loads. All three will help in minimizing unwanted damage due to
> > error. The objective is to ideally make it as least dangerous as possible.
> >
> > - Programmed through Opencl - parallel programming and computing.
> >
> > Challenges that have prevented robots from gaining foothold and that will
> > need to be solved or are in the process of being solved:
> >
> > 1) Actuators - robot needs to be bipedal (with subset of robots being of
> > other types) and with two arms and two hands, which is the most effective
> > and natural of types in relation to movement and nature of work. Actuators
> > need to be gravity-based - need to work with gravity (i.e., unlock/dislodge
> > and lock) and not against it with regards to all of its movements, which is
> > again the most normal and effective way.
> >
> > Point will be reached soon (through technological subsidies - i.e. moore's
> > law generalized for entire branch of technology) whereby actuators will be
> > good enough in relation to the necessities and nature of work, i.e. "good
> > enough". This point can be reached more quickly through an effective open
> > source technology/design.
> >
> > 2) Battery - assume limitless, point will be reached whereby batteries will
> > be sophisticated and powerful enough to power robot for many hours. Can be
> > solved partially by providing a remote source of energy, i.e. remote vehicle
> > with power generator (through gas, oil) or battery tank which can serve as a
> > recharging base for robot worker.
> >
> > 3) Programming - both dynamic (one branch will be everchanging without
> > regards to improvement, the other continously updating/improving, noting
> > that the two are not mutually exclusive) and static (which is not mutually
> > exclusive with dynamic, assuming that a part of the dynamic side can be
> > considered "almost static", i.e. various degrees of dynamic / static
> > interoperability). There are only two true static parts that are static laws
> > / definitions that do not and cannot be changed or updated by the robot:
> >   1) Never put in danger other living beings.
> >
> > 2) Always create and recreate 3d map / model so that it is possible to work
> > in the environment in relation to the first law and in relation to not
> > damaging the environment, which is defined as causing any change that is not
> > wanted when performing the task at hand, which ultimately always relates to
> > the first.
> >
> > 4) Computing. The advent of parallel computing through cGPUs has brought the
> > possibility of using many teraflops of computing power without a
> > proportional increase in power usage, heat, space requirements, and cost.
> >
> > Objectives:
> >
> > To replace primary / raw labour
> >
> > 1) Construction. Houses, buildings, public necessities, and other subsets of
> > construction, some of which may or may not be directly considered to be
> > related to "construction" in the way that it is primarily defined.
> >
> > 2) Raw Material aggregation - mining, extraction, farming.
> >
> > 3) Agricultural.
> >
> > 4) Recycling.
> >
> > Eventually, to replace all of labour in public service.
> >
> >
> > Challenges to Open source model:
> >
> > - Management. General management and direction through internet and website
> > is needed in order to direct and motivate developers. Blender is an example
> > of successful management.
> >
> > Advantages of Open source model:
> >
> > - Unrivalled values and unrivalled motivation. The best of ideas and the
> > best of dedication come from the purest and most virtuous of ideals.
> >
> > - Number of workers working on a unified (when properly managed)
> > development.
> >
> >
> > Disadvantages:
> >
> > - Difficulty of proper execution, which can be solved through proper and
> > indirectly centralized management - again, Blender is a successful example
> > of proper execution.
> >
> >
> > Types of work needed and available:
> >
> > - Programming
> >
> > - Engineering / Design
> >
> > --
> >
> > Hello Ton,
> >
> > Thanks for your reply :) Managing Blender must indeed be a very time
> > consuming work, and I understand why you would like to schedule less work so
> > that you can rest more.
> >
> > My idea is very simple - creating a free and open source robot worker is not
> > that hard, it only takes good will and motivated people, who are guided by
> > noble virtues and sentiments.
> >
> > The problem, as I explained, is that for a project like this to be
> > successful, it needs to be directed as well as possible. This could
> > apparently go against the notion of "free and open source", but it does not,
> > because 1) it prevents the project from being fragmented or destroyed (thus,
> > a momentum type of problem, where proper management makes the project gain
> > momentum as greatly and as well as possible) 2) the direction of the project
> > will still be democratic as long as the centralized management is as
> > transparent in what they do as possible.
> >
> > Blender is as I said, the primary example of proper management. The idea
> > behind this project is to be synergistic with Blender, in all of the
> > intended forms that Blender was conceived to be, which are 1) the program
> > itself 2) the will and community behind it, that are able to make Blender as
> > good as possible (advantages that commercial applications by their nature
> > cannot have). Thus, this project needs to be a subset of what Blender
> > actually is, not merely/only a 3d program, but an all-important device that
> > is able to turn the most noble and thoughtful of dreams into reality.
> >
> > I see that Blender also hosts a robotics related mailing list. That's great
> > to know :) I'll register and send them them this.
> >
> > Regards,
> >
> > - Thomas
> >
> >
> >
> >
> > _______________________________________________
> > Robotics mailing list
> > Robotics at blender.org
> > http://lists.blender.org/mailman/listinfo/robotics
> >
>
> --
> Will Jackson
>
> Director
>
> Engineered Arts Limited
> Unit 11 Kernick Business Park
> Annear Rd
> Penryn
> Cornwall
> TR10 9EW
>
> T. +44 (0)1326 378129
> F. +44 (0)1326 375752
> M. 07765 885545
>
>
> ------------------------------
>
> _______________________________________________
> Robotics mailing list
> Robotics at blender.org
> http://lists.blender.org/mailman/listinfo/robotics
>
>
> End of Robotics Digest, Vol 12, Issue 5
> ***************************************