Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~34C)(~2
This inYenti~n relates to a method and a device for autom~tic positioning of
a working tDol in respect to a worl{piece (possibl~/ made of several parts) so a5
to ~ring a worl~ing direction characteri2ing the tool opposite at least one
material trace provided on such worl~piece at a predetermined location. The
5 object of this inven~ion i5 more particularly but not e~clusively the assembly
of structures of l rge dimensions such as for example aircraft structures ~wing
section, fuselage...) and the automatic positioning of at least one of the
worl~ing tools intervening in such an assembly before (for e:cample drilling
securemsnt holes~ or during the assembling proper of the structure or the
10 workpiece made of several parts to be assembled.
This invention more especially relates to riveted assemblies but can also
be applied to other types of assemblies such as screwed assemblies or else
assemblies by spot welding or T.I.G. (tun~sten inert gas)i this presupposes
preliminary superpositioning of elements to be assembled which generally are
15 in form of planar sheets or sheets that were formed (warped surfaces~.
Essentially, processes for assembling sheets b~/ superposition such as
encountered in the aeronautic industry have recourse most aften to methods of
boring, then riveting or bolting.
Automati2ation of such processes raise difficult problems due to the
20 required accuracy for positioning holes and the large dimensions of the
structures whether a fuselage or a wing section is dealt with. Providing an
accuracy of ~ O.l mm all along panels of a length of about lO meters depends on
technologies which are not yet operational in the robotic field.
Sophisticated systems of automatic assembling are in fact already l~nown,
25 comprising programmable robots and worl~ing tools having adaptive capacities
in regard to the surface of parts to be assembled, according for example to the
article by Mr.MORESSEE published in the maga2ine "MACHINE MODERNE" in
December 19~7 (pages 3 to 7), or GB patent N?0~:095.
~7~
~23~002
However, these s~/stems are e:~clusivel~/ capable of proceeding b~/ welding.
Thus, the first document relates tc, a spDt welding machine with digital cantrnl
for advanced derospace construction in which electrodes are positioned on
multiple occasions opposite the parts to be welded and are automatically
5 oriented perpendicular to the surface of such parts b~/ means of sensors
associated with these electrodes. The ver~ position of the welding spots is not
defined with great accuracy. As to the second document, it relates to
progressive welding along one edge of a part to be welded to another part of a
car body by means of a pivotally mounted welding rod which can follow up
10 accuratel~/ said edge despite any lateral drifts of a robot to which said rod is
attached. None of these documents provides for sensir,g punctual material
marl~s or traces such as holes and bringing opposite thereto with great
precision a characteristic direction of a worl~ing tool. Therefore the teachings
from these documents cannot serve for example for a riveting assembling in
15 which the locations of the holes are defined accuratel~/ and the worl~ing tools
must be disposed accuratel~ both as regards position and orientation.
If the technological process of this t~Jpe of assembly is anal~/2ed, it appears
that the virtual centerline through which a drilling bit for ma~ing holes should
pass, the centerline of a rivet or a securement screw and finally the centerline
20 of a riveting or screwing machine must be substantiall~ perpendicular to that
surface portion ~said surface being ver~ often warped) where a securement
point must be realized.
This double condition of accuracy and perpendicularit~/ is to be further
completed b~/ the requirements for placing opposite one another on either side
25 of the structure to be assembled both tools required for fi~ation, i.e. either a
riveting pacl~ and a (pneumatic) hammer, in the one case or a screwing device
and a nut carrier in the second case where moreover the precision of
coincidence of the screw with the nut must be perfectl~/ ensured.
12340~Z
Accordiny to the riveting art previously l{nown, the fitting and crushing of
rivets (mDre than 5 million rivets on a plane su.h as AIR13US) were essentially
effected until now manually with two worlsers, the one holding the riveting
hammer and the other the rivet pacl~, with the sequences of operatian being
reali~ed as follows ~after drilling):
- spotting the hole,
- inserting a rivet,
- checl~ing for centering and perpendicularity of the hammer (or gun) and the
pacl~,
- controlling the pplication of forces (the force e::erted upon the rivet
head by the gun should be higher than that exerted upon the paclO~ and
- synchronization of the riveting step proper.
The manual sequence therefore implied insertion of the rivet effected by
touch and visual centering; the bringing into perpendicularity was obtained
intuitively during hammering by interpretation sf the applied forces or else on
stopping by interpretation of the contact ~de on the rivet.
Apart from the other assernbling processes by bolting or spot welding, if
one summarizes all the difficulties connected with the riveting assembling
process which constitutes the essence of the worl( presentlr realized in
assembling stations, in particular, in the aeronautic field, it is to be noted
that the socalled "first generation" robots are unsuitable for the following
reasclns:
- insufficient worl~ing precision in view of the required tolerances,
- too great a number of fixation points ta be made thereby excluding
apprenticeship and point by point programmation,
- poor ascertainment and absence of repetitiousness of the absolute
position of the fixation points in the assembly.
The ob~iect of this invention is to remedy these disadvantages by
implementation of combined active and sensitive elements.
123~0(~2
The invention actually lies in the very general field of the positioning of a
worl~ing tool of any type whether it be adapted to be used alone or whether its
utili~ation requires the intervention of a complementar~/ tool.
Thus, this invention proposes a rnethod of automatic positioning a worl~ing
tool in respect to a warl~piece 50 as to bring a worl~ing direction characterizing
the tool DppDsite at least ane material trace previously farmed on such
worl~piece at a predetermined location the at least approximate cooroinates of
which on the worltpiece are l~nown, such method comprisin~ causing relative
motion between the tool and the worl~piece so as to bring approximately the
tool opposite the trace on the worl~piece, detecting the difference between the
tool worl~ing direction and the trace and servocontrolling the tool position so
as to cancel such difference, then also detecting the orientation o~ the
characteristic direction of the tool in respect to a normal to the surface of the
worl~piece at such trace and servocontrolling the tool orientation so as to mal~e
such direction coincide with such normal.
The invention i5 advantageously carried out where a tool i~ to be used for a
significant sequence of operations of the same nature under similar but not
strictly repetitious conditions. This can be an operation to be effected on a
large number of successive parts or a large nurnber of operations to be
effected on one and the same part or any one intermediary case between such
extreme cases.
For carm/ing the method mentioned above, the invention also proposes a
device for automatic positioning of a wor~ing tool in respect to a workpiece
adapted to bring a characteristic worl(ing direction of such tool ~uccessively
opposite a series of at least two material traces previously formed on such
worl~piece at predetermined locations having approximately l~nown coordinates,
comprising:
A
1234002
- a robot related to a reference structure and provide~ with ~. working head
admitting of at least one degree of freedom and carrying at least said working
to~l)
- a centering sensor carried b~ the worl~ing head and adapted to detect any
deviation between a tool working direction and a rnaterial trace on the
workpiece,
- a perpendicularit~/ sensor also carried by the working head and adapted to
detect orientation of the normal to the surface of the workpiece at the locationof a material trace ~P) in respect to the tool worl~ing direction, and
- a control and proce5sing system associated with the robDt, connected to
the centering sensor and the perpendicularity detector and adapted to
automatically bring the working head successivel~/ opposite appro~:imately each
trace and for each trace to servocontrol the working head position in respect tosaid worl~piece so as to bring the tool characteristic direction opposite such
trace, then to servocontrol the wDrking head orientation in respect to the
worl~piece 50 as to mal~e the chaFacteristic wor~ing direction coincide with said
normal.
It will be noted that neither the article by Mr.MOF~ESSEE nor the British
patent cited above describe or suggest the utili2ation of ooth punctual
2 0 material traces and of means for position centering of the worl~ing tool and
means for adjusting the orientation of the working tool.
The position sensor and the perpendicularit~/ detector are preferabl~/
substitutable for the worl~ing tcol considered, with the tool working direction
being defined in respect to the worl~ing head.
The material traces are generally holes and the invention can apply to the
drilling of fi~:ation holes in parts of worl~pieces to be assembled. It also
applies advantageousl~/ to the positioning of at least one of two
complementary assembly tools for example a positioning of a riveting pack
while a companion actuates a rivet hammer.
~34(~02
The invention hDwever provides for the above-mentioned positioning device
to ensure the positiDning of both complementary assembly tools or even
subsidiary wor~ing tQols for reali~ing preparatorr operations.
In practice, the available robots which meet the requirements of tolerances
specified for a given application admit of a limited deviation tor field of
action) generally lower than the worl~ing zone covered by the group of material
traces formed for e~:ample on a wing section or a fuselage. ~ccording to a
unique arrangement of this invention, it teaches in this case that the robot(s)
should be mounted to the carrier head of a wide motion m?nipulator placed
under the command of the robot control system so as to dispose successively
such robot opposite each of a series of trace sub-groups tscanning field)
d~limited in the worlCing ~one.
In practice, such manipulators of wide range of action are not very
accurate; this is not disadvantageous according to the invention since final
positioning of the tool opposite each trace i~ provided by servocontrol
depending on indications from a position sensor or even a perpendicularity
detector. It is to be noted that due to such capacity of the tool to be
positioned finally by servocontrol, there can be introduced into the robot
control system and any associated manipulator approximate coordinates of the
worl~ing points advantageously corresponding to a simplified version of the
actual meshing tor networlt) of the material traces formed on the worl~piece
considered. If need be, the memory capacities of the control system can be
reduced, which may present advantages as regards investment costs.
It is to be noted that the control system provides for:
- an open loop control of the robot and any associated manipulator, and
- a closed loop control of the robot after suitable processing of informatian
transmitted from the different sensors and detectors.
As was e~plained previously the wor~ing directions of two complementary
assembly tools must be both in coincidence with one another and with a
3L~34()02
fi~.~ation point (P) to be reali~ed, thereb~ leading to the definition of a virtual
centerline of preliminary detection before determining the actual positioning
of said worl~ing centerlines. In other terms~ sensors for defining such virtual
detection a~is must c,e associated with such tools to permit suitable
positioning and orientatisn of the latter relative to one another. It is to be
noted that the positioning of a rivet pacl~ might have to satisfy a less strict
tolerance than does a rivet h mmer in bringing the worl~ing a~es into
coincidence .
Such positioning and/or orientation sensors may be of different types in
particular pneumatic, with photoresistors, phototransistors, or inductive or
with optic fibers.
This invention moreover will be better understood hereinbelow by the
description of a non limil:ative e~:ample of a form Df embodiment which appears
to be the best solution in case of an automatic riveting device for aircraft
structure, with reference to the attached drawings in which:
- figure 1 is a schematic view showing the disposition in space of the
re~erence trihedrons of the constituent elements of an automatic assembly
device according to the inventioni
- figure 2 is a schematic view showing various characteristic sequences of
an automatic positioning method according to the inventioni
- figure 3 is a schematic view in perspective showing in an orthonormed
space the logic behaviour of a programmed robot according to the invention;
- figure 4 i5 a schematic elevational view of the mechanical elements
comprised in a device according to the inventioni
- figure 5 i5 a schematic detailed view thereof showing the disposition of
the outer robo-t in the position denoted by the arrow V of figure 4i
- figure ~ is another schematic detailed view thereof in the position
denoted b~/ the arrow VI of figure 4;
~23~
- 'igure 7 i5 a partial perspective view of the outer robot with its worl~ing
head and its assembl~/ tool5i
- figure ~A is a schematic view of part of a networl~ fixation point on the
structure to be assembledt and figures ~:B to ~E are partial sectional views of
one fi~ation point according to the steps of the assembling Df the structure to
be assembledi
- figure qA is a partial sectional view of two worl~ing heads during
positioning and orientation thereof, and
- figure qB is a partial top view of the inner worl~ing head showing the pacl~
centering means along line lX-IX of figure ~Ai
- figure 10 i5 a pariial perspective view of the inductive centering sensor
of figure 9Ai
- figure 11 is an electric diagram of the centering sensor of figure lOi
- figure 1~ is a partial schematic perspective view of the gauges of the
perpendicularit~/ detector of figure 9Ai
- figure i3 is a diagrammatic view of the logic control system of an
automatic assembly device according to the inventioni
- figures 14A to 14C are schematic elevational views of the outer worl~ing
head for successive positions of its sensors and worl~ing tools.
The following description i5 given by war of example in respect to the
assembly of an aircraft structure of large dimensions namely, a fuselage
according to fixation point networl~s of any form ~point lines or mo5t uftent
rectangulart square or quincunx meshes). The elements in the structure (S) to
be assembled are positioned relative to a reference structure R with which a
fixed reference trihedron (O(R)t Xt Yt Z) i5 associated . Each of the two
complementary assembly tools is mounted to a working head which is
designated in view of the closed configuration of the structure to be
assembled by internal or external worl~ing head depending on its position
relative to such structurei each wor~ing head is associated with a reference
~234002
trihedrDn denoted ~O~TT), X, Y, Z), O(TT) being identified by O(TTi) or O~TTe)
depending on whether the internæl or the external head is dealt with. Each
worl~ing head i5 connected with a robot in reference to which it has at least one
degree of freedom (generally five degrees~. Each of the~e robots is mounted to
the carrier head of a manipulator 50 as to have at least one degree of free~om
available relative to the reference structure ~for example four degreesj; each
carrier head i5 associated with a reference trihedron ~OSSP), X, Y, Z), the
designation SP bein3 followed with the character "i" or "e" dependins on
whether the carrier head is inside or outside the structure S. In a simplified
msdified form of embodiment, not shown, bDth robots are mounted on one and
the same manipulator when the geometry of the structure to be assembled 50
permit~.
Figure 1 illustrates very schematically a portion of structure to be
assembled (S), a reference structure ~R). a carrier head (SP) and a first
external worl~ing head (TTe) as well as a second internal wor~ing head (TTi) theassociated carrier head of which i5 not shown. The reference trihedrons
associated with heads TTe and SP are connected through connecting means
designated by A.
By conventiont axes OtTTe) -Z and O~TTi)-Z designate those axes according
to which the complementar~f assembly tool5 comprised by the heads are adapted
to work.
With reference to figure 1, it can be noted that the problem to be solved in
an automatic assembl~/ device according to the invention consists, before any
assembl~/ operation:
- of bringing into coincidence the axis O(TTe)-Z of the external worl~ing
head (TTe) and the axis O~TTi)-Z of the internal working head ~TTi) according
to a common axis Az,
- of providing the orientation of such axis Az according to a normal to the
surface of the structure to be assembled (S), and
~234002
- of centering such axiC A2 to a an assembly point ~P) of said structure.
Complementaril~/, translations along such axes are required to permit
disengagement of the worl~ing tools or sensc,rs.
Assuming now that the assemblies in question must be effected according
S to lines (ly~ and ~lx) on the surface ~S~, which corresponds to the meshing of the
structural rivetin4s or boltings, it can be seen im nediately that where said
structures are of a very high dimensions (for example, fuselage~, direct
positioning of A2 to successive points ~P~ of fixation cannot be ensured with
the required accuracy through such devices as the presently l~nown
manipulators or robots.
In the general case, to provide the re~uired motions for orientation and
positioning of the axis ~A2), the worl~ing heads must be permitted to move and
rotate according to five degrees of freedom about O(TTe) and about O(TTi), i.e.
three translations about the axes and rotations about two axes preferably X
and Y.
It is to be noted that if such positioning and arientations are perfectly
reali2able nowadays with the rec~uired accuracies for little structures, they are
unreali2able with huge structures encountered in airspace industries for
exarnple, with airplane fuselages.
2 0 This is why according to a preferred mode of embodiment of the invention,
there are associated an accurate robot and a heavy little accurate manipulator
having however a ver~/ high dimen~ion and there is added to the so reali2ed
assembly logic 5ystems of positioning relative to traces ~or tracers~ for
detection of the fixation points to be reali2ed as well as appropriate locl~ing
mean5 for the robot and manipulator.
An axis A materiali~es in figure 1 the coupling mode between O~TTe) and
O~SP) i the trihedron associated with the latter is in the general case driven
with four degrees of freedom ~translations according to axes X, Y and Z, plus
one rotation according to Y).
lC
l`~.
~23~L002
The logic positioning system preferably controls the wor~ing heads
according to the sequence of steps indicated by the diagram of figure 2.
The function of the carrier head (SP) is to supply motion of the associated
robot of a scanning field (D1~ to another (D~...) on the structure (S) whereas the
worldng head associated with such robot effects within said scanning fields d
programmed motion from one approach zone ~za) to another, with each approach
zone ~eing centered on a fi:~ation point (P). Sensors as defined hereunder
thereafter provide in each approach zones (za) for perfect centering of the
worl~ing head on each point (P).
Thus, when the heavy manipulator and the robot have brought according to
F1 the head ~TTe) within ~zal), in a point such as (P'1) for e:~ample, and when
their locl~ing has been effected, the sensors carried by the head autcmatically
provide for correct positioning of the latter on (Pi) thereby permitting after
proper retraction the utilization of suitable worl~ing tools.
] 5 Through programmation of the robot, (TTe) moves thereafter to the second
approach zone (za2) according to (F2) to get to (P'~) where the centering occursthereafter as described herein~bove in (P1).
The sequence is continued for zones (zaâ), (za4~ where the head (TTe) coming
in the respective points (P'3), ~P'4) will be centered on ~P3) and (P4) according
to paths (F3) and (F4).
In the e~:ample considered, such reduced programmation i5 based on
previous l~nowledge of pitches in a meshing representative of the regular or
irregular networl~ of the assembl~/ points (P) according to lines (ly) and (lx).When in the scanning field (D1) the working head reaches its limit of motion
for example on the position (P4) of the field (Dl), it is neessary to produce achanye in the position of the robot and the carrier head (SP).
To this end, in the field (D1) the worl~ing head remains centered on the trace
(P4); upon programmed incremental displacement of the carrier head to reach
~2~002
$he field ~D~), the wor~Sing head remains servocontrolled to said point (P4) by
means of its sensors.
The assembl~ operations in position ~P4) will generall~/ be effected on field
(D~) which actually will happen to be common with (D~) according to ~la) which
therefore contains at least one point.
After processing ~P4~ it is shifted to the approach ~one ~a5) and this occurs
as previousl~/ described in regard tu the occurrence of the fùllowing
operations.
Figure ~ shows schematic~ the "travel" of the worl~ing head (TTe)
according to Fl, then F2, F3, and 50 on.. , as well as the "traveli' of the
wor~ing head ~TTi) according to F'l, then F'?, F'3, and so on..., for assembling
metal sheets 5a and 5b. Various obstacles 5c uch as pinning clamps or else
ribs are advantageousl~/ avoided due to suitable detectors which are being
dealt with in the continued description.
In this description the following definitions are given merel~ by
convenience to the various elements used according to the invention:
- LUCIFON ~or Logic used in cDnversational for integration of evolutive
functions to a robot) for the robot control software,
- a heav~ programmable manipulator for manipulator 1, the worl~ing space of
20 which represented in mixed lines in figure 4 is defined as a function of the si~e
of the elements to be assembled. It can be of an~/ nature in form of a gantr~/...
- tools for the active devices mounted to the wor~ing head whether the~/ are
drilling, riveting, mounting tools, and/or dismounting tools for pinning clamps,
spot welding taols, screwing tools... and
- structure for the whole of the elements to be assembled.
It is to be noted that the fixation points (P) must be understood as
positioning marl~s for the assembl~/ operations; they can be simple traces
materiali~ed on the structure to be assembled i the points 0 designate the
origin points of the mDtions and rotations in space.
12
lZ34~02
Returning to the fundamental f~mction resulting from the utili~ation of a
device according to the invention, in an ~ssembling operation it is to be
considered that all the assemblies of the riveted or bolted type require a
series of simple operations that can be summari2ed as follows ~while remindin
that such operations were until now reali~ed manually):
- spotting the location of the assembly paintt
- drilling an oriented hole completed or not by countersinl~ingt
- oriented positionins in such hole Df a pinning clamp, a rivet or a bolt, and
- jointed oriented applications of the clamp pinning/removing or riveting
gun and of the pacl~. followed by riveting proper or oriented presentation of
the nut opposite the bolt, then screwing of such nut with the predetermined
torque (it being noted that a washer can be interposed therebetween).
The complexity of the operations tc be reali~ed on large structures require
definition of the specific functions in each case which must ta~e into account
various parameters such as precision, quicl~ness of execution, easy
implementation
Merely b~ way of non restrictive example and in order to illustrate any
possible conditions of application of the invention, figures 4 and following
show how a meshed rivetin~ on a section of heavy carrier airplane fuselage is
practically carried out.
The whole of the device substantially comprises ~figure 4) an external
structure 5, fixed relative the ground, on which the riveted assembly points
must be reali2ed, a programmable external heavy manipulator i provided with a
carrier head SPe carrying a rc,bot ~ with its wor~ing head and which for exampleis caused to move longitudinally on rails 7, a prograrnmable internal heavy
manipulator 3 also movable longitudinally and provided with an internal carrier
head SPi which carries one or more rob~ts 4 with the worl~ing head(s) ~: thereof.
In the form of embodiment shown the worl~ing head carries a riveting pacl~
and the manipulator ~ is an adapted mounting to permit for example by
13
~2340()2
securement to the rails of the fuselage floor the bringing into coincidence of
said pacl~ opposite each riveting hole.
In figures 5 and 6 the robot 2 associated with the programmable heavy
manipulator I is in the summit position and the low position respectively.
Examination of these figures shows that the programmable heavy manipulator
1 is used for providing displacement of point 0 ~SP) ~with a low accuracy) on a
traiectory f which must remain at a substaLntially constant distance c relati~/e t~
the surface of the structure ~S)i point 9 ~SP) is representative of the a~is of
connection with point 0 (TTe) representative of the point of movement in space
of the wor~ing head ~. It is specified that such programmable heavy
manipulator must be lQd~able when the external worl<iny head ~ lies in a given
situation.
E:~amination of figures 5 and ~ also shows that it is possible to ensure
motion thereof in all cases and to permit the axis 0 (TTe) -~ (see figure 7) to
lS be perpendicular to the normal to surface (S).
In figure 7 a conformation of the robot (2) i5 shown in perspective thereby
illustrating the possibilit~f of the five degrees of freedom about O~TTe), it
being noted that displacement according to O~TTe)-Y results from rotation
about O~SP)-Z. Since the structure of this robot does not itself belong to the
inventiont it will not be described in detail.
The robot 2 comprises apart from the mechanisms and jacl~s required for its
orientation the worl~ing head ~ itself carr~/ing a sensor 9 for detecting the
fixation points to be executed which will be described in detail hereinbelow as
well as worlting tools which can be a drilling machine and/or a riveting gun
shown schematically by the volumes 10 and 11 respectively.
Upon sequential motions implied by the assembly the sensor 9 is retracted
by lateral translation as well as the successive tools such that point ~P)
always lies opposite the member for executing the respective phase ~see
figures 14A to 14C). Thus, upon a locating phase the sensor 9 will be centered
1~
~L2;~4a~02
on the trace ~P) which ma~/ be any marl~ reali~ed by paint loaded with magnetic
powder or any other tracer. In the drilling phase said sensor 9 moves off
laterall~ leaving room for the drilling machine ll which thereafter will drill
the hole at ~P). After the drilling step said drilling machine will retract at its
turn leaving room for the riveting gun lO which introduces the rivet thereinto
and then crushes it against the pacl~ which will be opposite thereto.
In a modified form of embodiment, the gun can be replaced with a pin fitting
or removing device or a screwing machine or else 2 bolting machine.
Figure ~A shows in a top view the meshing of the assembly to be realized
10 and figures B to ~E shows different steps of realization of such assembly:
- ~B, spotting the point (P),
- æ~, ~after drilling the holel possibly with countersinl~ingi insertion of the
rivet,
- ~D, mounting the rivet, or
- ~Et fitting a bolt or a screw in.
The meshing shown is for example substantially of a pitch according to X, of
40 mm, and of a pitch according to Y, also of 40 mm, and the metal sheets to be
assembled which generally are made of aluminum alloy are substantially of a
thicl~ness of 2 mm and the holes of a dian~eter in the order of 4 mm. Thus, the
20 dimension of the fields ~l;l) and (D2)... illustrated in figure ~A is substantially
in the order of 430 >: 4~;0 mm, which includes 130 points (P).
Apart from the centering on point (P), which requires mDtions according to
O(TTe)-X and O(TTe)Y, and noting again that this setond type of motion is
provided by the rotation about O(SP)-Z ~by suitable iacl~s not shownJ, the
25 question is raised of bringing the worl~ing head perpendicularly to the
assembly surface which results from the rotations about O(TTe)-X and
O(TTe)-Y.
These rotations are ensured by means of mechanisms (schematized at 12 in
figure 7) connecting the worl~ing head ~ to the robot 2.
1234~)0~
The motions and rotati~ns of the head ~ controlled from sensor ~ with
which a perpendicularit~ detector 13 is associated will be better explained
hereinafter in reference to fi~ure qA.
Similarly, the worl~ing head :: containing the riveting pacl~ carries for
reasons mentioned above a centering device 14 associated with a centering
reference 34 on the head ~ as well as advantageously obstacle detectors 15, as
shown generally b~ the schematic view of figure ~A.
With reference again to figure ~C, it can be readily seen that if the drilling
phase requires onl~/ a centering and then correct perpendicular positioning of
the drilling machine onto the assembly surface opposite pDint IP)t which was
reali~ed manually in the prior art, the same does not apply to the positioning
of the rivet, the screw or the bolt and their assemblings which must be
effected accurately in automati2ation. Studies for automati2ed insertions of a
cylindric part into a hole such as positioning of a rivet or a bolt are dealt with
in many worl~s, in p~rticular, b~/ J.NEVINS and D.WHITNEY "Categorization and
status of assembl~/ research", Report N P 33û Lab. STARK l:~ROOPER 197~, and
bring to light that the misalignment to prevent blo~l~ing results anyhow from
the dimensioning and the assembl~/ tolerances and that "the insertion funnel"
determines the space geometry of the centering and the perpendicularity of the
insertion to be effected.
Thus, according to figure ~C:, the radial tolerance of presentation of the
rivet ~R) on the axis of its hole for insertion thereof shall be lower than 0.1
mm after said rivet was in the approach 20ne of the hole ~za), which is in the
order of + 3 mm, on either side of the a>:is in the proposed e>:ample.
In other terms, and summari2ing the different positioning sequences. it can
be mentioned that:
- the programmable heavy manipulator 1 provides for motiQns of robot ~
according to O(SP)-X, -Y and -Z, plus one rotation about Y to permit surface
follow up of the structure to be assembled. Its positioning accuracy is in the
16
~23~002
order of 1 mm i its carrying head is locl~able to a given positic,n and it is
programmed by a logic to permit translation from one field ~ to another,
- the robot 2 provides for rnotions of the worl~ing head according to
O(TTe~-X, -Y and -Z to ensure centering and perpendicularity on point (P). It
5 also provides for two rotations about X and Y. Its positioning accuracy is in
the order of 0.01 mm ano it is proorammed through a LUCIFON logic,
- the worl~ing head 6 causes lateral retraction of sensor 9 carryin3 the
perpendicularity detector 13 to leave room for the drillin~ machine 11 or the
riveting/screwing device 10. These operations are programmed,
- the programmable heavy internal manipulator 3 with its pac!~ carrier robot
4 and its worl~ing head or pacl~ ~: provides for motions opposite the worl~ing
head 6, which requires three movements according to O(TTi)-X, -Y and -Z, plus
two rotations according to O~TTi)-X, and -Y. Its positioning accurac~/ is lower
than 1 mm, its robot can be servocontrDlled to robot 2 and can be provided with
an obstacle detector in view of the many elements existing for e~:ample within
a fuselage.
In figure qA all sensors and detectors transmit information to control and
processing units denoted by:
- 16, for the reference ~4 of the pacl~ centering means 14 (optic fiber
2 0 transmitter)~
- 17, for the perpendicularity detector 23 (with stress gauges~,
, for the centering sensor 9 (with induction coils),
- 19, for the pacl~ centering means 14 (optic fiber receiver and
ph otodete ctors),
- 20, for the obstacle detector 15 (with magnetoresistors).
It is to be noted that the sensors and detectors of which the operation will
now be described are those which were retained as being entirely satisfactory
but it will be understood in this respect that any other type of appropriate
sensor copuld also be us@d.
17
~L23~00Z
The centering sensor q is in the e~:ample shown of the inductive eàd~/
current t~pe. It consists of ~our flat coils 21a to ~ld with axe~ parallel to
O(TTe~-Z such a5 illustrated in figures 9A and 10 and disposed in a square
symmetrically relative to the axis; the straight lines extending through the
symmetrical centerline of sensor 9 and the centers of each pair of opposite
coils constitute both sensitivity axes of the sensor. A hole 22 formed in the
axis permits an optic fiber 34 which is a reference of the pacl~ centering rneans
14 to extend therethrough.
As shown by the theoretical diagrammatic view of figure 11 two coils 21a
and ?lc (or 21b and 21d) constitute by pairs the diagonal of an alternative
current bridge cornprisiny two resistors R1 Fe2 an amplifier AMt an oscillator
OSCL and a synchronous demodulator I}EMSYN. When the symmetrical axis of
the coils is offset by a length a x relative to the a:xis of the hole ti-ere
occurs an unbalance f~Z in the impedance of the coils due to dissymmetry of
eddy currents induced into that portion of the surface 5 of the structure to be
assembled opposite the axis of the coils. The output voltage Vs i5 then
proportional to the offset relative to the axis in a given linearity range.
A model suitable for the ex~mple given in figure ~:A permits to point out
the following characteristics:
- sensor with two uncoupled axes and coils of ~00 spires of thread of 0.0-:
mm with adjustable magnetic core
- + 3 mm of linearity range
- resolution to 15 micrometers
- overall position error lower than 150 micrometers.
Each coil lies at a small distance from the portion of the surface of the
assembly 5 and is clamped in a box 22A.
The perpendicularity detector 13 associated with the sensor 9 comprises
three strain gauges mounted to metal blades as shown in figures 9 and 1 .
1~
1234~)02
It may be reminded here that force sensors permit determination of
perpendicularity through identification of a complete or p~rtial twister
relative to an orthonormed referential.
The perpendicularity detector with deformations recommended within the
scope of this invention for reasons of simplicityt accurateness high
sensitivity and little bul~ comprises three L-shaped metallic beams denoted
23a ~3b and 23c receiving by glueing strain gauges J11 and J1? J2i and J22t
J31 and J32 J41 and J42 J51 and J52 and J~l and J~.
The three beams are mounted sandwiched between cylindric platforms one
of which denoted ~4 is made integral with the slide mentioned hereinafter
while the other denoted 25 is a spacer connected to the inductive sensor 9 ~see
figure 12). The platforms and spacers are considered to be endlessly rigid
with sunl~en type connections.
The principle of operation of such gauges resorts to the elasticity theory
by which one can deduce through measurement of elongations by means of the
gauges the cDrresponding strain twisters relative to an orthonormed
referential denoted RJAU in figure 12.
The gauges used are of a current type associated with conditioners and
mounted in a l~nown manner.
The question i5 to measure a group of microdeformations ~ i in certain
points of the three bearns disposed at i?O4 and to which said gauges are ylued.
These beams essentially worl~ on bending and the recorded defarmation is a
linear function system of the forces applied thereto.
According to the arrangement of the gauges relative to the orthonormed
referential RJAU:
- (jil ji~) i = 1 to 3 give the following lateral forces F:~ and Fy by
measurement of the bending moments.
- (jil ji2) i = 4 to 6 permit measurement of the force Fz and the moments
Mx and My.
19
~34002
Without entering intc~ mathe natical details of force modeli~ation, it may be
mentioned that the model is of the following type:
Fz U2
Mx = M, U3
U4
My U5
U6
where Ui are the responses from the six gauge bridges used.
Calibration of M was obtained by application of pure forces l~nown in
direction to points with l~nown coordinates in referential NRJAU. Terms of the
matrix M were extracted from the group of linear equations obtained. Such a
calibration permits cancellation of anr ambiguitr as to the position of the
referential where the forces are e~:pressed and moreover mal~es it pDssible to
tal~e into account certain indeterminate elements (glueing of the gauges.. ~.
Information on measurement of the forces is processed by microprocessor
provided with a device for obtaining the six voltages in analog form from the
conditioners of strain gauges.
She centering device for the pacl~ carrier 14 comprises according to figure
9A an assembly permitting accurate centering in the order of O.i mm to provide
correct riveting of the rivet. Such assembl~f comprises a light source
incorporated in blocl~ l6 with optic fibers 34 fixed to the centering device 9 and
perpendicularity detector 13 assembly as well as a light receiving system 35
which is mounted to the pacl~.
The principle of operation of the sensor is such that when the sensor 9 in
contact with the metal sheet is centered to the hole a cold light beam 35 is
transmitted from l~ passing through the hole whereas on the Dther side a
receiver system 35 (see figure 9E3) composed of four optic fibers 36a to 36d is
connected to photodetectors 37.
1234()02
~he lo~ating method i5 based first on the balance of the output signal frorn
fiber 3~a and that from fiber 3~c (in opposition) thereby to have the pacls 14
situated at the abscissa of the hole, then by balancing out the output signal
from the fibers 3~b and that from fiber 36d the pacl~ is centered strictly on
th at hole .
She principle of operation results frorn a transmission linl~ between the
photodetectors 37, the signal processing circuitt the ccmputer and the
actu tors which are all located in blocl~ 19.
Detection of obstacles is effected in this example by means of an obstacle
detector 15 with magnetoresistors ComprisinQ a stationary section 39, a
movable section 40t magnetoresistors 43, magnets 4i and drawbacl~ springs 42.
These elements being disposed in the radial and axial interface of the fixed
and movable pDrtions in opposition, detection is produced by current variation
in a WHEATSTONE bridge when the pacl~ 14 meets an obstacle 44. In such case
induction varies and an electric signal occurs and interpretation of the
imbalance in the bridge permits to locate the direction of the obstacle and
produce the required reversal cammands. The obstacle detection signal is
furthermore processed in blocl~ 20.
The sensor q and perpendicularity detector i3 being positioned on surface
5, as appears from figure 9A, the assembling phases result from the following
ordinogram:
. RESEARCH AND POSITIONING PHASE:
. Preposition to a point by using the programmable manipulator as
remote control manipulator,
. Position after e~:trapolation,
. Plane on plane contact of sensor locating the point.
. Proceed with centering of sensor to the point and recording the
corrections to be made to provide perpendicularity relative to the metal sheet.
~L~3~C)O;;:
, Move off sensor. I proceed with compensatiDn
. Approach sensor. I for any plays.
. Store this new position and e~:trapolaîion.
. Center the pacl~.
. Move off the senor.
. Position tools ~uch as drilling, riveting machines
. Approach tools.
RI VE Tl N G PHASE
. Suppl~/ rivet.
. Force on gun.
. Force on pac~.
. Strike.
. Stop stril~e.
. Stop force exerted on pack.
. Stop force e:~erted on gun.
. Move off gun.
. Position locating system.
. End of cycle.
-
Several remarks can be IT~ade:
- the rivet5 are aligned and the pitch is roughly constant,
- for the riveting, holes are supposed to be drilled without diameter error.
The information processing means required for the control algorithm of
robot 2 ~figure 13) substantially comprises:
- a microprocessor ?6 adapted to pilot the step-by-step motors of the
2 5 robot,
22
~'~340~;~
- a microprocessor 27 adapted to control the orientation motQrs for head
digitili~e analog data from the inductive sensor ~ and the strain gauges of the
perpendicularity detectclr 13~ to control the u5ual electric cantrot ~alves ~nd
finally to stop the manipulators when a strain sensor~ not shownt is
overloaded, and
- a computer 2~: which incorporates th3se generic logics which permitted to
obt~in human being and system interactivity of the required level such as
cnntrol of manipulators, procæssing of forces.
Finally, the automatic device according to the invention is completed by a
manual control console 45, graphic consoles 46, control unit 47 and operator 4~:.
All these operations have been organi~ed relative to a group of
orthonormed referentials constituting an arborescent graph for five degrees of
freedom controlled a~out O~TTe)-X and -Y and according to O~TTe)-X, -Y and -Z.
With reference to figure 3 the orthonormed referentials are denoted ~s
follows in the LUCIFON software logic:
- NRJAO robot end reference
- NRJAU n-_asurement force referential
- NRRIV inductive sensor position
- NRCO ~magnetic sensor referential
2 C - NRMAG l ~plane on plane contact)
- NRE JTA robot referential
- NCONTn actual position of point ~P)
- NCONTn + 1 estimated position of point ~P)
- NRTAO ground reference
- NRETA worl~ing head referential
- NRCCO position contact nominal value for
servocontrol
DEP dimension vector ~ e~pressing translation
and rotation of NTROU/NRTAO.
~34002
Servocontrolling by force and~or position permits modification of the
position of the socalled "contact referential NR~O" relative to the socalled
"contact nominal value referential NRECG" as a function Df either the relative
position IIRCCO~NRCO ar measured forces. The movement of NRCO relative to
NRCCO is defined axis by axis in NRCO.
To establish plane on plane contact (NRMAG applied onto the metal sheet),
the servocontrol has the following parameters:
- the translational nominal values according to axes X and Y (origin of
NRCO on a:~is Z in NkCCO~
- the force nominal value F2 according to axis z tc, be satisfied except if
the origin of NRCO i5 situ~ted on the plane X and Y of NRC(::O,
- the speed nominal value of rotation about aY~es X and Y of NRCO if the
measured torques expressed in NRCD exceed a certain threshold.
For learning the position of the point (P~ or a hole, the position of NRCC~
i5 piloted relative to NRCO from manual control of translational motian of the
robot, thereby to permit thereafter to obtain the plane on plane contact as
soon as a vertical force ~which will remain limited) will appear.
As soon a5 point (P) is situated in the capture field or approach zone ~2a)
determined b~/ the inductive sensor 9, measurement of the variation permits
positioning of the marlt NTROU, the estimated position of point (P).
The extrapolation method applied is the following:
- let be NTROUn the group oi locations previously spotted and DEPn the
position of such marl~s relative to the referential NRTAO, there will be the
following:
DEP = Vl, V2 ' V3 V4 V5 V6
The extrapolated position NCONTn + 1 is calculated from displacement
DEPn + 1, each component Vn + I of which is obtained by the extrapolation
method described hereinafter:
- linear extrapolation (2~n~3):
~4
~234~0~
one lays down: Vn ~ 1 = a + bn
it must be verified, if m6n
Vm = a + b.m
rhe coefficients a and b intervene linearly in the previous equation and are
identifiable by the use of the matricial pseudo-irverse notion:
- quadratic e~ trapqlation ~n~4)
one lays down Vn + l = a ~ b.n + c.n and the preceding method i5 applied
theretD.
In this extrapolation mode which can more efficiently provide the follow up
of a curved line of points, only the four l ~st positions of the spotted points
will be used.
l hus, by application of su h a method, a line of points ~P) can be readily
marl~ed by teaching the system only the position of the first two points.
It can be seen in this manner that the method shown schematically in figure
2 and illustrated by way of e~ample in figure ~A i5 reali~able withaut
difficulty within a given field (Dl).
Soon after scanning of this given field, i.e. as soon as the last point tP4)
was scanned, the robot ~ remains controlled on said last point, while the
programmable heavy manipulator l causes displacement there~f such that said
robot 2 i5 again in a position to scan the following point ~P5) of the followingfield (~2) and 50 on
As a matter of fact, the robot 2 provides for progran~med displacement of
the wor~ing head 6 within a given scanning field while the programmable heavy
manipulator causes said robot 2 to move from one scanning field to another and
this constitutes one of the preferred characteristics of this invention.
As was already mentioned above, it is actually not possible to design at
present a single robot presenting the same characteristics as those perrnitted
1~3~002
by application of this invention comprising an accurate programmed robot and a
programmable heav~/ manipulator of large amplitude servocontrc,lled thereto.
Servocontrolling the programmable heavy manipulator 1 only requires little
accurate po~itioning for e~:ample in the order of + 1 mm which is sufficient to
ensure the locætion o~/ the centering sensor S in the approach 20ne which is in
the order of + 3 mm, whereas the accurate centering will be effected
thereafter with an accuracy better than 0.1 mm, and this, in as large a range asdesired such as that encountered for example on a fuselage or an aircraft wing
section.
The assembly operations proper such as appearing in figure ~ comprise
several steps impl~/ing the spotting of point (P~, drilling of the hole at that
point, introductiDn of the rivet, the screw or the bolt, then riveting or
screwing. In a modified form of embodiment, a pin clamp can be inserted
temporarily into the hole~
This implies the positioning of different members on the same side, which
will now be explained in reference to figures 14A to 14C.
In these schematic figures as well as also in figure 7, a working head has
been shDwn carrying the instruments in their different operation sequences:
- in 14At the inductive sensor 9 with the perpendicularity detector 13 are
positioned on the surface 5 of the structure to be assembled opposite point
(P),
- in 14B, the sensor was retræcted to the left-hænd side to leave room for
the drilling machine 11,
- in 14C, the drilling machine moves to the right-hand side to leave room
for the riveting or screwing device 10, with conjoint positioning of the pack
carrier (or nut carrier) 14. Automatic supply of rivets 32, screws or bolts, of
known type (not shown) can be added thereto.
26
t~i~
~23~002
The wor~ing head ~ (accor~ing to figure 7 and figures 14A to 14C) i5
~ubstantially L-shaped one outer side of which is connected to its own
orientation mechanisms with jac~s 12 mentioned acove Inot ~hown~.
The substantiall~ vertical inner face 29 carries a longitudinal slideway 30
in which a slidable part made integral with parts 10 and 11 move.
The substantially horizontal inner face is in the form Df two slideways 29A
slidingly enclosing a slide carrying the sensor q and the detector 1~ as well asparts iO and li.
The terminal portion of these slideways is inclined upwardly so that the
sensor q and the detector 13 in their displacement lie remote from structure 5
upwardly .
A iacl~ 31 causes the assem~ly to move in its slideways 29A and 30 while a
second iacl~ 33 detern~ines the relative position between the sensor q and the
elements 10 and 11.
In the detection phase comprising the centering on point (P) and evaluation
of perpendicularity the sensor 9 must be applied to the assembly according to
figure 14A. As soon as the centering and the perpendicularity have been
obtained the worl~ing head 6 stops and the sensor 9 is moved off laterally and
releas~d by the jacl~ 31 to leave roo for the drilling machine 11 on the point (P)
as results from figure 1413.
After drilling the drilling device is transferred by the iac~ 31 to leave
room to the riveting or screwing device lO which implies conioint application ofthe pacl~ carrier 14 opposite thereto. The last phase can then occur after this
last operation.
~3~002
It is to ~e noted that the order of the drilling and riveting phases can be
changed so that a complete row of holes can be first drilled and then fitted
with rivets, screws ur bolts.
Obviously, the technology of the sliding, drilling, riveting and screwing
systems, with or without automatic supply of rivets, screws and/or bolts, and
ccntrol systems through hydraulic, or pneumatic or electric jac~s are
sufficientl~ nown in themselves 50 that it is not useful to descricbe them.
Also, it will be understc,od that the riveting device 10 and the pac~ 14 can
be replaced withDut any particular difficulty with a point welding assembly or
else T.I.G.(tungsten inert gas)~ one of the tools cbeing then a welding torch.
lt will be noted that this description was only given by way of not
limitative ex~mple to better disclose the possibilities offerecl by the
invention.
In this respect~ a processing image camera could complete the sensitive
system in particular to facilitate learning processes and also obstacle
detections. The numbers of degrees of freedom of the different heads (SP) Dr
(TT) can be reduced for example up to 1, depending on the simplicity of the
assembly operation (~or example, a planar assembly surface without any
perpendicularity problems and in this case no perpendicularity oetector is
2 0 required).
In such simple cases a single manipulator can carry both worl~ing heads
through tWD robots coupled to one another and in this case only one centering
sensor and one perpendicularity detector are sufficient for both heads. As to
the meshing or networl~ of the fixation points, it can be reduced to one line oron the contrary~ correspond to meshes of very complex shapes. When the
fixation paints (P) are not formed by holes, they are advantageously marl~ed by
an electromagnetic field disturbing material such as paint loaded with powder.
Furthermore, it will be understood that the programmable heavy
manipulators can be remote control manipulators with mechanical hinged
28
~;~340~Z
~ystems with six axes mounted on rails. One heavy manipulator can also be
mounted as a gantry enclcsing the ~tructure.
It will also be L!nderstood that many further variations can be proposed cy
the man cf the art without however departincJ from the ~cope of the invention.
29
.~
