RE: manufacturing/robotics in construction

From: Makai Smith (
Date: Tue Jan 14 2003 - 17:27:53 EET

Rupert & Scott,

Thank you for engaging in this conversation publicly, especially for taking
the time to post such informative messages.

After poking around online, I see that these projects like T-up and ABCS are
innovations in technique for sequencing and construction on (mostly) tall
structures; their development being driven primarily from the construction
end. It if fascinating what a large degree of resources are spent by Asian
construction firms, such as Ohbayashi, on R & D. I wonder if the American
architecture and construction industry is throttled by a lack of basic
research, if not vision.

The other techniques of which you speak, the family of additive techniques
resembling RP, offer a completely different set of potentials to rearrange
the industry, both in regards to what is built and who participates in that
process. You hint at this in your post by referring to the "capture of the
architectural and structural engineering expertise within a computer
program", implying the eventual shakeup of what our professions can call
"their work". (Architects have yet to fully embrace the computer as a
design tool let alone as an intellectual studio partner).

You are quite right in saying that when most architects think of solid
freeform fabrication they are thinking of a modeling process, that is, for
the scale representation of a form. That other form, the 'real' structure,
is likely to be constructed using conventional construction techniques. So
the possibility of form is limited not by the modeling techniques but by the
construction method and therefore RP is overkill given most of what is being
built. RP is capable of achieving fantastic complexity usually beyond what
is necessary for an architectural model.

When construction techniques change, RP will be an accurate analogue to the
construction process, much like modeling with basswood and chipboard present
a similar material reality to stick frame and drywall.

I am excited by the research you are doing into functionally graded
construction. I am reminded of the marrow in our bones. We are made
lighter by having a hollow structural member, while using that cavity for
essential functions (blood production). The implication of this, with
respect to the organization of our various professions, is that architects
cannot be isolated as those who 'design' buildings, assisted by consulting
engineers, to be realized by yet another trade. Integration of the spatial
and mechanical systems will require an accompanying integration of the
architectural and engineering aspects of the building design. (Although who
knows how collaborative social structures via the web or other such system
may allow the disciplines to work).

Another important implication, for me, is that when buildings are
constructed directly by machines from electronic files, the design
profession will be required to rethink the 'drawing' tools it currently
uses. Honestly, I will get a good laugh when AutoDesk tries to get me to
buy "Voxel Desktop" to breathe new life into its legacyware yet again.

I am left with questions such as which side of the equation will need to
change first and how? Will RP and 3D in the architectural studio will be
passed on to the job site, or instead it will wait until contractors begin
demanding 3D waypoints for positioning due to jobsite automation. (Maybe
I'm too impatient, but can anyone say what's the darn hold up?)

|\/| /\ |< /\ |
O. Makai Smith

-----Original Message-----
From: Rupert Soar []
Sent: Friday, January 10, 2003 7:56 AM
Subject: manufacturing/robotics in construction

Hi Scott,

Please pass on my greetings to Ian. I have been following the
development of the projects you list in your e-mail for some time
(including ABCS, Big Canopy, T-up etc) and I am firm supporter of these
construction techniques as solutions for tall structures and
skyscrapers. These techniques have been developed to partly overcome
the very real problems associated with employing skilled steel workers
at great heights. I think I am correct in saying that these processes,
essentially, take the worker to the location where they are needed and
supply a safe/stable environment in which to fabricate.

Relating these techniques, and the very real benefits you mention
(incidentally I would very much like to know the status of current
research in this field), to what I discussed in my RPML posting. As you
rightly distinguish, the systems you are working with are essentially
'dry' due to the problems associated with control of 'wet' materials,
but for me these differences vary little from what the rest of the
construction industry is moving towards (i.e. 'dry' modularisation and
sub-assemblies both on and off site) and does not break away from the
key issue which is the problem of organising and utilising large numbers
of skilled workers that are hard to synchronise and hard to find
(particularly in the UK).

The manufacturing industry has been through various stages of
standardisation and modularisation but in the end the technology allowed
us to move beyond this until today we can consider concepts such as Mass
Customisation. The construction industry is in exactly the same
position. In order to synchronise the construction process the only
option has been to look at standardisation and then modularisation as it
ensures consistency away from the hands of the site staff who cannot be
automated. The only real problem with modularisation and lean
construction is that once a 4 tonne modules has been made up it must be
assembled to tolerances of tenths of a millimetre, by cranes (and
operatives) working with tolerances of a millimetre. The other key
drawback is that modularisation results in a descretisation of the
building design into cubes and destroys any chance of design freedom.

So as I stated, a solution is to drive a machine on site (much like a
boom pump) or set up a gantry system and shutters and start building
that structure using the principles found in techniques such as Behrokh
Koshnevis's Contour Crafting process or a large scale 3DP process (see
Generis) which are essentially modified and controlled cure 'wet'
systems that will allow all the design freedom and little human
interaction as well as all the other benefits I mentioned in the RPML
posting. In that posting I also inferred that the major advantage of a
wet system is that you can modify its properties at the point it is
deposited/extruded. This does not necessarily imply you need "hitech"
materials (though they will be important). Just by blowing gas in to
say an extrudate, as it leaves a nozzle, during deposition, may be
"hitech" enough. Hope this answers a bit. I'll keep you posted.

I've taken the liberty of posting this back on the RPML as there has
been a good response to the first posting.

Dr Rupert Soar
Wolfson School of Mechanical and Manufacturing Engineering
Loughborough University
LE11 3TU
tel: +44(0) 1509 227637
fax: +44(0) 1509 227549

-----Original Message-----
From: A. Scott Howe []
Sent: 10 January 2003 08:18
To: Rupert Soar
Cc: Ian Gibson; Prof. Thomas Kvan;
Subject: manufacturing/robotics in construction

Rupert, I just got your email forwarded to me from Dr. Ian Gibson of the
Department of Mechanical Engineering at University of Hong Kong. The
email discusses robotics and automated manufacturing in architecture. I
am very much interested in this topic and would like to join any
discussions that you might have along these lines.

I am currently an assistant professor at HKU Department of Architecture.
Before this time I worked for Kajima Corporation in Tokyo where we
developed the AMURAD lift-up construction system, and I am continuing my
involvement as team member with Kajima, Shimizu, and other Japanese
firms on the next generation robotic construction systems. Our approach
has been to develop design grammars for flexible kit-of-parts systems
that result in demountable, rearrangeable structures. The construction
hardware can become extensions of the kit-of-parts system, and building
components become extensions to the construction hardware (eliminating
heavy lift hardware and assisting in material handling). Using this
technology, buildings become dynamic systems that can be rearranged by
the users, rather than static one-off artifacts that are more often than
not obsolete as soon as the occupancy begins. Though much of our work is
still conceptual, several concepts such as AMURAD or Shimizu's SMART
have resulted in prototypes that have constructed several high-rise

As architecture and the construction industry attempts to rise above
dated construction processes and enter the information age, we are often
self-conscious about wet trades (in-situ concrete, plaster, masonry,
etc) that have been so typical in our field and bring so much chaos to
the construction site. For this reason we have avoided wet trades in our
research in favor of "hitech" materials and processes, except that which
has been completely necessary (I don't include precast concrete in this
observation). However, some of the items in your email hint at digitally
controlled real-time forming and layering of materials that may be
considered wet trade. I would be very much interested in having some
discussions along these lines.


A. Scott Howe, Ph.D.
University of Hong Kong Department of Architecture
AIAA Design Engineering Technical Committee (DETC)
AIAA DETC Aerospace Architecture Subcommittee
Plug-in Creations Architecture, LLC

This archive was generated by hypermail 2.1.7 : Sat Jan 17 2004 - 15:17:01 EET