Base Truss Crawler
Crawler
Created 2020-01-03 Friday
Review 2022-11-24
– AssemblyCrawler (redirects here)
Definition
A crawler or "assembly crawler" is a robotic pick and place manipulation platform
permanently attached to a RepRec Base Truss such that
– it has boundless (or at least sufficient) mobility on that RepRec Base Truss in at least one dimension
– it can boundless (or at least sufficient) extend the basetruss
A crawler may be dysfunctional without a RepRec Base Truss because
the motion across the basetruss may be depended on for positioning parts in pick and place operations.
Why crawlers on base-truss rather than self-contained replicators?
The same reason as for the RepRec Base Truss.
Crawlers and base-truss go together, they are mutually dependent for self-replicative capability.
In essence: To have an extensible system that as a whole remains in GMT/MachinePhase.
Disjoint self-replicating units (crawling or swimming around) would need some fancy things
(like possibly even image recognition) to seek out and dock with other disconnected parts of the system
in order to make larger scale products.
Replication is only a means to the target of productivity not "the end goal".
At the nanoscale one can go for self-assembly. But that's only useful for early foldamer based systems. Very useful there though.
Crawlers on base-truss match quite well as a RepRec Second Assembly Level in the gemstone metamaterial on chip factory concept.
Self contained replicators match the (now obsolete) Molecular Assembler
Example approach (Moses2014)
In Moses2014 the crawlers have a Cartesian gantry crane geometry and
move on a 2D square grid of racetracks ( a 2D BaseTruss )
All crawlers are restricted to motion along a linear track in one dimension only.
To get 2D extendability crawlers assemble further crawlers in a 90° rotated state on 90° rotated track seeds.
Designing RepRec Crawlers for motion along a 1D track can simplify the design complexity notably.
As demonstrated in Moses2014.
Restriction to one DOF only, and how to deal with it
In the case of 3D RepRec base truss, 2D-tracks are replaced by 3D channels.
For reducing system complexity crawlers can be restricted to just one single base-truss channel.
Just like in the 2D example case in (Moses2014) RepRec Crawlers
restricted to a single base-truss channel must be able to assemble
"offspring" replicas that feature rotated orientations, such that
all the six orientations of the Octet truss channels (not counting sign thus not 12) are accessible
either via one or several stage rotations via replication.
More formally: Replication operation rotations need to span the whole mathematical rotation group.
See: RepRec crawler replication rotation group
Mass
When putting more and more complexity from the systems products, the BaseTruss and other static parts
into the mobile #Crawler then mobility may be come harder and harder to justify.
Possible alternative:
Having the manipulator static and moving an assembly strip (like a conveyor belt) back and forth within the BaseGrid
But assembling a replication of the assembly mechanism on that moving strip what one gets essentially is a mobile
manipulator. One is back at the RepRec crawler
Alternative approaches
Design space parameters:
★ Complexity balance …
– Crawler can replicate in multiple orientations(places/spaces) but cannot go anywhere — vs —
– Crawler can go anywhere but replicate only in one orientation(place/space)
★ Locomotion … rotative — vs — translative
★ Location of the motors … RepRec motor backpack on base or on tip of RepRec positioner
★ Crawler interdependence … one crawler or more than one crawler that only can replicate as a complete set (borromean crawlers)
Yes, there are also other less motion limiting approaches.
– One of them is turning on the tracks (nontrivial)
– One of them is rotation around truss beams and moving brachiatingly through the base-truss.
More on that below ...
Brachiating
Issue: Locomotion by brachiating through the base-truss gets more impractical with bigger more sophisticated crawlers,
as likely for crawlers in nanoscale systems with Mechanical through joint threading
Also at the macroscale gravitative mass becomes a problem leading to:
– no longer ignoreable downward deflections (aka sagging) needing nontrivial software compensation
– need for high forces due to high mechanical (dis)advantage => stronger heavier motors, a vicious cycle
– potentially damaging load on the truss
=> thus this approach will not be taken here (for macroscale prototypes)
There are examples of existing prototype systems taking this approach (The Center For Bits And Atoms is working on this)
Note that the complex crawlers there are fully pre-produced. They do not assemble other crawlers from more fundamental base-parts.
Maybe (likely?) suitable for space-tech context, but not suitable for nanoscale-tech.
Toothed racks –vs– brachiating
Racks are a much faster mode of motion than brachiating. A RepRec base trusses
3D printed racks are probably less accurate than properly designed 3D printed rotative bearings (??).
But given that the accuracy requirements here are not at the level of a 3D printer it should be ok.
Motion by racks either locks RepRec crawlers into a specific motion axis (e.g. a specific octet truss channel)
or require complex means to switch between them (does not seem worth the effort ??)
(In an octet truss there are three channels in a truss layer and six channels in full 3D space).
More complex math:
A brachiating mode of locomotion changes assembly coordinates in quite complicated ways
if assembly is meant to be allowed not just in a few special positions but in all positions .
And the double usage of the locomotion DOF as assembly DOF is desirable.
For continuous racks the nodes of the RepRec base truss are in the way.
In an octet truss rack crossings are probably worse than in a square grid like the one in Moses2014.
Alternative names
– Crawler … troublesome and misleading bio-assozialtin (crawling proliferating out of control bug infestation)
– IncompRec … making it more clear that this is not a complete RepRec without its [RepRec base truss]]
– MobileUnit … bland & inconvenient & too unspecific
Note: RepRec poser and repRec positioner are
not applicable as names because these are
sub and sub-sub systems respectively.
Ideas
Macroscale: If rotations of big/heavy motors are involved:
Maybe its possible to rotate motors almost around their center of gravity ??
Regarding driving a crawler (all scales):
RepRec mechanical demultiplexing could be done to use the same motors
for the motion of the RepRec crawler crawler as are used for the RepRec poser.
Not advisable though if the motion of the RR/Crawler serves as one of the axes of the RR/Poser as in Moses2013.
Regarding energy supply for a crawler (macroscale specific)
The energy supply for the RepRec crawler could be a strapped on battery and data can be transmitted wirelessly.
This way there is no requirement for …
★ either threading mechanical motion across all of the BaseTruss, See: Mechanical through joint threading
★ or cheating with …
– either free floating flexing electrical coil-spring cables or
– or electrical sliding contacts
Regarding energy supply for the driving of a crawler (nanoscale specific)
★ If motors are kept static in the base-truss then Mechanical through joint threading
becomes even more of a headache. Requiring base-truss to crawler threading and long range transmission along
some changing length of base-truss struts/tracks.
★ If motors get integrated into crawlers (may make sense for big crawlers!)
– Wired electric drive: Graphene sliding or rolling contacts.
– Elastomechanical battery drive: Recharging internal springs (how long can that last when operating efficiently?)
– Chemomechanical battery drive: Gathering and dumping bond-zipper energy storage capsules (complex & very big?)
