Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

International Journal of Industrial Ergonomics 23 (1999) 473—487

Accessibility and ergonomic analysis of assembly product


and jig designs
Venkat N. Rajan*, Kadiresan Sivasubramanian, Jeffrey E. Fernandez
Department of Industrial and Manufacturing Engineering, Wichita State University, Wichita, KS 67260-0035, USA

Received 10 March 1997; received in revised form 10 September 1997; accepted 11 September 1997

Abstract

In the aircraft industry, the design of floor assembly jigs (FAJs) is an important activity that directly affects
productivity. It involves tool frame generation and locator and clamp placement to ensure that the assembly components
are held properly with respect to each other to meet the required tolerances. The tool designer also has to analyze the
design to ensure that the assembly process does not pose accessibility and ergonomics related problems. The current
approach is dependent on the experience of the tool designer and the limited visualization possible on commercial CAD
systems. This leads to extensive redesign when accessibility and ergonomic related problems are detected on the physical
prototype. In this research, an integrated Virtual Reality-based environment is being developed for the analysis of
assembly product and jig designs. CAD models of the assembly product and jig are imported into a Virtual Reality
(VR)-based visualization system for accessibility analysis. A motion tracking system is integrated to allow ergonomic
posture analysis. The combined VR and motion tracking system allows evaluation of alternate assembly sequences and
the jig design. In this paper, the theoretical basis for the analysis environment is presented along with details of the
prototype implementation of this system.
Relevance to industry
Floor assembly jigs are used extensively by the aircraft industry. Improvement of their design process will lead to
savings in better design and reduced development time and cost. Better designs will require fewer changes after the jigs
have been fabricated. The overall result will be a reduction in product realization time and cost and improved product
quality.  1999 Elsevier Science B.V. All rights reserved.

Keywords: Assembly planning; Ergonomics; Assembly jig design; Virtual reality; Anthropometry

1. Introduction downstream problems before the commitment of


physical resources. Many practical issues generally
Spatial simulation of reality aids the design pro- encountered during the manufacture of complex
cess by providing the ability to experience potential products can be avoided if the designer can predict,
analyze, and effectively resolve them at the design
stage. In assembly operations performed in the
* Corresponding author. aircraft industry, floor assembly jigs (FAJ) are used

0169-8141/99/$ — see front matter  1999 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 9 - 8 1 4 1 ( 9 8 ) 0 0 0 0 7 - 9
474 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

extensively to manually create large subassemblies. tooling. Therefore, there is a need for a system that
Aircraft sections like the fuselage, wing sections, is capable of allowing accessibility and ergonomic
upper tail cone, and lower tail cone are assembled analysis in a computer-generated environment.
separately and then joined together using these jigs. The objective for this research is to develop a VR
Such assembly tooling is generally of large dimen- based system to enable interactive analysis and
sions and complex design. The assembly activity in evaluation of assembly product and tooling for
aircraft manufacturing affects to a significant extent interference and reach. This paper describes the
the quality of the end product, and the productivity overall research framework from which this system
of the entire process. has evolved. An overview of the system implemen-
The assembly tool design activity initially in- tation and a case study demonstrating its capabili-
volves the generation of the FAJ frame. This is ties are also presented.
based on the general requirements of the subas-
sembly to be created. Then locators and clamps are
designed to ensure that assembly components are 2. Background
held properly, and assembly processes can be per-
formed. The FAJ should allow orientation, align- 2.1. Jig and fixture design
ment, and joining operations to be performed such
that required tolerances are met. Current tool de- Reliable assembly tooling must be able to hold
sign methods are based on the experience of the components and subassemblies in an accurate and
designer, coupled with the limited visualization repeatable position, prevent undesired motion of
and experiential design capabilities provided by components, and avoid posing interference prob-
commercial CAD systems. Problems that arise as lems in assembly tasks. Previous research has re-
a result of deficient or improper design of these sulted in several different approaches to certain
assembly tools usually result in expensive redesign tooling design problems. Computer-aided design
efforts. A host of associated problems also arise, techniques, heuristic algorithms, and Expert Sys-
including loss of accuracy, interference, and awk- tems have been developed and implemented to op-
ward postures. Conventional CAD and animation timize certain aspects of assembly tooling design.
systems usually fail to account for the actual behav- The use of software tools for the computer-aided
ioral and geometrical constraints of material, design and analysis of fixtures indicates that the
equipment, and people. Interference, an accessibil- trend is towards expediting product development
ity problem, refers to the unintended obstruction by concurrently engineering the manufacturing
caused by the presence of structural components of process with product design (Anon, 1995).
tooling in the assembly trajectory of components, For fabricating aircraft tubing assemblies, an
subassemblies, and hand tools. Problems also arise FMS has been developed (Kelley and Chu, 1987),
when assembly workers are forced to adopt ergo- that uses a robot-assisted flexible fixture assembly.
nomically stressful postures when assembling com- A CAD/CAM system has been incorporated in the
ponents on to the assembly jig. The effects of flexible fixture design process and in programming
awkward postures over extended periods lead to the robot and machine tool motions. A systems
musculoskeletal disorders. The cost of one such case approach to fixture planning and design integrates
can be up to $100 000 (CTD News, 1993) which can process planning information with results of geo-
drastically affect the financial stability of a company. metric and kinematic studies of tooling (Hargrove,
Systems available today do not allow the de- 1995). Computerized fixture design systems provide
signer to concurrently model the assembly task and sophisticated means of creating alternative fix-
perform several tool design iterations, to improve turing designs, but many of them overlook the
accessibility and ergonomic characteristics. Tem- integration of process planning information.
poral collocation of design and manufacturing/ A mathematical model has been developed (Prom-
assembly related issues would help identify prob- banpong et al., 1992) to integrate process planning
lems before constructing expensive equipment or in fixture design. This model claims to resolve the
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 475

contradictions between the computerized fixture place. Motorola, Caterpillar, John Deere, Genesis
design and the process plan. Systems Group, and Tecnomatix Technologies are
A Graphical User Interface-based interactive fix- some companies that report having benefitted from
ture design and assembly environment interacting using VR (Owen, 1994).
with a solid modeler and a fixture elements Repetitive strain injuries of the upper limbs and
database is described by Fuh et al. (1995). This was back persist as shown by the high number of re-
found to help designers select and verify fixture ported workplace illnesses and injuries. Research
elements from a database before assembling and clearly shows a relationship between awkward pos-
constructing them. While much work has been tures and the risk of injury (Tichauer, 1973; Arm-
done in machining fixture design, interactive and strong et al., 1993; Bridger, 1995). Muscle-loading
concurrent assembly tool design is yet to be ex- requirements of the forearm flexors have been
plored. This would shift the design from being found to change with posture (Moore et al., 1991),
purely based on experience and trial-and-error, to implying that certain postures are more favorable
one that is based on scientific principles. than others in keeping muscle exertion within safe
limits. The automobile industry is leading the effort
2.2. Virtual reality and ergonomics to take advantage of VR technology in addressing
ergonomic issues (Jackson, 1996). Virtual environ-
Virtual Environments (VE) are interactive vir- ments are now used to identify areas of safety and
tual image displays enhanced by special processing health concerns, review process changes, and evalu-
and nonvisual display modalities, such as auditory ate and rectify ergonomics-related problems in
and haptic, to present the effect of immersion in manufacturing before building physical prototypes
a synthetic space (Ellis, 1991). VE enables the (Holbrook, 1991; Hale, 1994; Deitz, 1995; Hancock
modeling of manufacturing systems and com- and Lai, 1995). Systems that allow the designer to
ponents using sensory features and feedback to evaluate the motions associated with an assembly
simulate or design a manufacturing environment. task while still in the design stage of the tool would
Boeing Computer Services reports two research prove useful in reducing product development time
projects in which aircraft CAD information is and therefore time to market, improving quality
loaded into a VE (Mizell, 1994). Applications in- and reliability, and lowering costs.
clude design activities in which engineers can view
and interact with CAD representations as if from
inside a physical prototype. The use of augmented 3. Assembly product and jig analysis framework
reality is also reported, wherein the user wears
a see-through head-mounted display with a Virtual A system for integrating assembly design, plann-
Reality (VR)-style position/orientation sensing sys- ing, and analysis with ergonomics, using VR as
tem for guidance in manufacturing. a medium is currently under development. Research
As opposed to conventional CAD systems which done as part of the Product Realization and Intelli-
have limited 3D visualization capabilities, a VE is gent Systems for Manufacture (PRISM) and the
proposed for simulation of an assembly model in Ergonomics groups at the Wichita State University
which a designer can visualize and manipulate the forms the basis for this effort, as shown in Fig. 1.
object using precise real time collision detection The JIGPRO system incorporates knowledge
(Kim et al., 1995). A virtual system has been de- available from past research and the functionality
veloped (Allen et al., 1995) for assessing manual of VR in providing the virtual environment for the
assembly forces applied and postures assumed by design of assembly tooling and its ergonomic anal-
human operators. A modified glove input device ysis. The PRISM research is integrated with the
and force triggered image capturing facility help in JIGPRO system in three ways. Assembly modeling
the assessment, yielding physiological information research has resulted in the development of assem-
for further analysis. This is seen to be useful in the bly representations and constraint propagation
prevention of repetitive strain injuries at the work- and verification methodologies (Lyons et al., 1997).
476 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

The prototype assembly modeling system is dir- sign and analysis of fixturing setups. These method-
ectly interfaced with the JIGPRO system to pro- ologies form the basis for the assembly tool design
vide the assembly model. The modular fixturing system that provides feasible jig designs to the
research (Rajan et al., 1997a) has resulted in the JIGPRO environment. Virtual Reality has been
development of methodologies for geometric de- used to design factory layouts (Rajan et al., 1997b).
This knowledge is applied to the assembly tool
evaluation process to create the JIGPRO system.
Ergonomics research has been focused in various
areas, such as psychophysics, cumulative trauma
disorders, and anthropometry (Fernandez, 1995).
The effects of various ergonomic risk factors such
as force, posture, duration, and repetition on hu-
man capacity have been extensively studied. Psycho-
physical studies have revealed that deviated wrist
postures affects maximum acceptable frequency of
upper extremity exertions in some assembly tasks
(Fernandez et al., 1995). The JIGPRO system is
designed to integrate this knowledge to evaluate
the ergonomic aspects of a given FAJ.
Fig. 2 shows the architecture of the JIGPRO
Fig. 1. Structure of overall research effort. system. Currently available systems allow certain

Fig. 2. JIGPRO system architecture.


V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 477

limited analyses of assembly jigs and permit the The user is forced to follow this trajectory when the
analysis of simple tool designs. More attention component nears its final position and orientation.
needs to be given to features that would enable Once the component is located in its final position
designers to create virtual mock-ups of typical and orientation, the hand tools are used to perform
assembly tasks on a given tool, manipulate and the appropriate drilling and riveting operations.
‘operate’ hand tools as in manual assembly and Accessibility problems arise when components in-
determine the characteristics of the tooling with terfere with any jig elements during the course of
regard to accessibility and ergonomics. For accessi- locating them. Such problems arise when the drill-
bility analysis, the JIGPRO system derives its be- ing or riveting work area cannot be reached using
havior from assembly modeling and assembly tool the hand tool because of obstruction by jig
design considerations. In the ergonomic analysis, elements. Ergonomics related problems also arise
ergonomic principles are used to determine the during component assembly and drilling/riveting
appropriateness of postures and work heights. operations.
In the following sections the nature and content
of the assembly product model, process constraints,
4. Accessibility and ergonomics analysis and FAJ design will be detailed. The procedure for
methodology determining feasible assembly components for
a given assembly state will also be discussed. The
Simulation of the process to create an assembly geometric representations for the gross motion and
product using the JIGPRO system consists of the fine motion trajectories and their generation are
following steps: presented. Methods and tools used to identify ac-
1. Import the assembly product CAD model and cessibility and ergonomics related problems are
process constraints. also presented.
2. Import he FAJ CAD model.
3. Import hand tool CAD models. 4.1. Assembly product and process information
4. Generate the gross motion trajectory for a feas-
ible assembly component by manipulating it from In order to analyze the jig design in a VE, the
its pick up point to near its assembly location. assembly process needs to be simulated. Such
5. Assemble the component into the current partial a simulation requires information on the assembly
assembly state by following the fine motion tra- product model and any constraints that affect the
jectory. assembly process. As shown in Fig. 2, the assembly
6. Manipulate hand tools to perform the assembly product model is generated by the prototype As-
operation. sembly Modeling System (AMS) and the process
7. If any additional components remain to be as- constraints are generated by the Assembly Plan-
sembled, go to step 4, else stop. ning System (APS) and exported to the JIGPRO
The assembly product model and process con- system. The nature and content of the assembly
straints are obtained from the Assembly Modeling product model and process constraints are dis-
and Assembly Planning Systems respectively. The cussed below.
FAJ and hand tools (drill and rivet gun) have been
modeled using a geometric modeling system and 4.1.1. Assembly product model
are directly imported into the JIGPRO system. The The assembly product model needs to include
gross motion trajectory is obtained by capturing the models of the various components comprising
the motion of the hand within the VE as the user the assembly and the relationships that exist be-
moves components from their initial to their final tween them. A traditional geometric modeling sys-
positions. A feasible assembly component is deter- tem will yield component geometric models and
mined as one whose assembly will not violate any their mating constraints. While the component geo-
process constraints. The fine motion trajectories metric models are useful for determining interferen-
are obtained from the assembly product model. ces, the assembly mating constraints are of little
478 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

by analyzing the degree to which an assembly com-


ponent is constrained. A component is said to be
completely constrained if it cannot be disassembled
from the final assembly state (Rajan and Nof, 1996).
Clearly, for such components, one or more of the
other assembly components needs to be disassem-
bled before they can be removed from the assembly.
For a completely constrained component, the min-
imal precedence constraint (MPC) is defined as the
alternate sets of the other assembly components
Fig. 3. Insert and mate assembly constraints in a peg-in-hole whose presence in any partial assembly state will
assembly. prevent its assembly (Rajan and Nof, 1996). The
notion of minimal is obtained by the fact that each
use. These mating constraints are used to specify alternate set is by itself necessary and sufficient to
the relative location and orientation of compo- prevent the assembly of the completely constrained
nents. However, mating constraints that specify component. An example of the MPC is shown in
contact relationships between components Fig. 7. Components A and C are completely con-
have significant impact on their degrees of freedom strained if it is assumed that the only permissible
(Rajan et al., 1997c). For example, consider the assembly/disassembly motion is a single translation
peg-in-hole assembly shown in Fig. 3. from outside the assembly envelope. The MPCs for
The insert and mate constraints specified between these two components are given as
the mating surfaces of the two components con-
strain each component to translate along and ro- MPC(A)"(BCD)R(BDE)
tate about the axis of the cylindrical surfaces. The MPC(C)"(ABE)
impact of these surface mating constraints on the
component degrees of freedom are not captured by The MPCs are computed by the APS and exported
traditional CAD systems. to the JIGPRO system.
In prior research, representations for capturing
joint information and embedding it within a hier- 4.2. FAJ design information
archical (bill of materials) assembly structure have
been developed (Rajan et al., 1997d). This repres- The jig design information consists of the geo-
entation is shown in Fig. 4. The joint information metric model of the tool and associated non-geo-
includes surface mating constraints between com- metric information such as material type. The tool
ponents, joint type (rigid or kinematic), and joint geometric model is in fact an assembly of compon-
attributes. Component degrees of freedom are clas- ent models. The base model is the tool frame that
sified into 11 categories, as shown in Fig. 5, and provides strength to support the assembly compo-
vary from unconstrained to completely constrained. nents and stability during the assembly processes.
As components are added to an assembly and mat- The strength and stability of the frame have to be
ing constraints are specified, the component de- such that no significant deflections occur when
grees of freedom are computed, as shown in Fig. 6. components are loaded and processes are per-
The assembly representation is then exported by formed.
the AMS for use in the JIGPRO system. Other geometric models within the tool assembly
model are locators and clamps that are used to
4.1.2. Assembly process constraints accurately position and constrain the assembly
During the assembly process, precedence con- components. As components are mounted on the
straints exist and they dictate the order in which assembly jig, the locators are used to correctly
components can be assembled, and cannot be viol- position components relative to each other. Addi-
ated. Such precedence constraints can be computed tional temporary locators are used to correctly
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 479

Fig. 4. Representation of kinematic and surface mating constraints, and joint attributes.

locate holes relative to each other as they are dril- 4.3. Feasible assembly components
led. After the holes have been drilled, the temporary
locators are replaced by rivets sequentially. To ana- The evaluation of the assembly design involves
lyze accessibility and anthropometric character- determining the ease with which the assembly can
istics, it is essential to simulate the loading and be put together. The assemblability of a set of
unloading activities and one assembly process, components is a function of the assembly sequence.
drilling or riveting. Only one of the four activities However, a particular sequence can and should be
associated with each assembly operation (drilling, selected only by considering the current state of
placing the temporary locator, removing the tem- a given assembly system (Rajan and Nof, 1996).
porary locator, and riveting) needs to be evaluated This ensures that the components can be assembled
because limited accessibility and violation of ergo- in the best possible manner consistent with the
nomics principles for any one activity will apply to capabilities of the assembly system. Because the
all of them. human operator and the assembly tool represent
480 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

Fig. 5. Joint definitions and associated degrees of freedom.

the assembly system, the most effective assembly partial assembly state is +B, D,. Of the remaining
sequence that complements the tool design and the assembly components A, C, and E, only A is a feas-
capabilities of the human operator must be ible assembly component because assembly of C or
selected. Because the proposed approach also E will violate MPC(A).
integrates the evaluation of the tool and assembly As the operator evolves the current partial as-
designs, it provides an excellent framework to allow sembly state from completely unassembled to fully
the human operator to evaluate and determine tool assembled, the appropriate feasible assembly
designs as well as assembly designs/sequences that components are identified by comparing the as-
best fit his/her capabilities. sembly state with the MPCs of unassembled
In order to allow the operator to experiment completely constrained components. By identi-
with alternate assembly sequences, the methodo- fying all feasible assembly components for each
logy only enforces assembly process constraints. state and allowing the operator to pick any one of
Thus, depending on the current partial assembly them, he/she can explore all feasible assembly se-
state, feasible assembly components are identified. quences. By constraining the selection to only feas-
An assembly component is said to be feasible if its ible assembly components, it is ensured that only
addition to the current partial assembly state will feasible assembly sequences are explored. There-
not violate the MPC of any unassembled, com- fore, the efficiency and quality of the assembly
pletely constrained component (Rajan and Nof, product and tool evaluation activities are im-
1996). Consider the example shown in Fig. 8. The proved.
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 481

Fig. 6. Computation of component degrees of freedom in the Assembly Modeling System.

4.4. Gross and fine motion trajectory representations to the final assembly position. This is determined
by the assumption used to generate the MPCs. The
The trajectory used to assemble a component interactive capability of the VE allows us to exploit
can be divided into two segments: gross motion and the ability of humans to quickly identify good feas-
fine motion. The gross motion trajectory is between ible gross motion trajectories. The general trajec-
the initial location of the component to a position tory of an object handled by the human hand does
near its final location. The fine motion trajectory is not follow a well-defined geometric path such as
used to maneuver the component from that posi- a circle or a parabola. The motion could be charac-
tion to its final location to accurately mate the terized by a freeform curve. Because the position of
component with the other assembly components. the object is sampled at particular time instants, the
The generation of the gross and fine motion tra- curve has to fit multiple points in a smooth fashion.
jectories is equivalent to the general 3D findpath The more closely the points are sampled, the more
problem which is known to be intractable accurately will the curve represent the actual tra-
(Ahrikencheikh and Seireg, 1994). jectory. Because of the need to fit through multiple
The fine motion trajectory is a single transla- points, a Bezier curve formulation (Mortenson,
tional motion from outside the assembly envelope 1985) is adopted to represent the gross motion
482 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

Fig. 9. Gross and fine motion trajectories.

translational motion assumption may be somewhat


restrictive, it provides a quick and simple approxi-
mation for fine locational motions without the as-
sociated accurate positioning problems faced by
Fig. 7. Completely constrained components in an assembly. human operators in computer-generated environ-
ments.

4.5. Accessibility evaluation

Accessibility problems could arise when assem-


bly components cannot be located in their final
assembly position, when the completed assembly
cannot be unloaded from the tool, or when hand
tools cannot access the work area. Such problems
are created when interference between the assembly
component loading/unloading and hand tool tra-
jectories, and the jig elements such as the frame, or
locators and clamps. An example of interference
between the assembly component loading traject-
Fig. 8. Partial assembly state +B, D,. ory and the jig frame is shown in Fig. 10.
In other cases, while the trajectory may be feas-
trajectory. An example of the gross and fine motion ible, the operation cannot be performed because of
trajectories is shown in Fig. 9. interference between the human arm or body with
The capture of the gross motion trajectory from the jig elements, lack of sufficient space for the arm,
the actual motion adopted by the human operator or lack of proper visibility of the work area. In all
eliminates the mathematical complexity associated these cases, the possible remedies are to redesign
with the findpath problem. It also incorporates the the assembly tool, modify the assembly sequence,
specific motion capabilities of the human operator modify the trajectory, or change the assembly de-
performing the simulation. By using a well-defined sign.
fine motion trajectory (straight-line motion), the The accessibility problems are detected by com-
problems associated with accurately locating com- puting the intersection of the swept volume of the
ponents in a VE are eliminated. While the single moving objects (assembly components, human
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 483

Fig. 10. Example for interference.


Fig. 11. Neutral posture and joint degrees of freedom.

hand, hand tool) at certain time intervals with the and does not capture the particular attributes (i.e.,
volume occupied by the stationary objects (jig anthropometry) of the operator using the tool. In
elements, previously assembled components). Any this research, the VE is combined with a 2D motion
intersection between the rest of the body of the tracking methodology to capture the motion of the
human operator and the jig elements is not detec- actual human operator. This ensures that the tool
ted because only the hand model exists in the VE. design is customized to the particular capabilities of
An extension to this work will create a world model the operator.
of the work environment independent of the VR The motion tracking methodology is used to
system. Such a model will allow detection of inter- obtain human posture information. Only 2D analy-
ference between the entire human body and the jig sis is utilized for the rotational joints at the hip,
design using a combination of the VE with a real- shoulder, elbow, and wrist. The hip, shoulder and
time 3D motion tracking system. elbow joints provide gross motion for the hand
while the joint at the wrist provides orientation
4.6. Ergonomics evaluation capabilities. The neutral posture is shown in
Fig. 11, while an extended reach posture is shown
Ergonomics related problems arise because of in Fig. 12. The base coordinate system is located at
extended reach of work areas of stressful posture the hip. Motion tracking provides the location and
adopted during assembly operations by the oper- orientation of each of the joints in Cartesian space.
ator. While some of these problems may be created The location and orientation of the hand joint are
because of the lack of operator awareness, usually used to identify extended reach and incorrect pos-
they are due to improper design of the assembly ture problems. The inverse kinematics problem is
product/sequence or the jig. solved for each of the intermediate positions to
The VE only contains a model of the human determine the individual joint angles. These values
hand. Because the local coordinate system of the are used to determine whether the problems are
human hand is referenced with respect to the global created because of motion of the shoulder and
coordinate system of the VE, no information can be elbow within the sagittal plane.
obtained about the posture of the human operator. Future extensions to the system include main-
One potential solution to the problem is to create taining an independent world model of the system
a model of the entire human body within the VE. to allow accessibility and ergonomic evaluations of
This, however, tends to be a generic human model the entire body as described in the previous section.
484 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

The motion of a user’s hand during the assembly


of a component is captured by the WTK system as
a series of points in 3D space. A Bezier curve is
fitted to the gross motion trajectory points for each
assembly component. This path can be saved for
further analysis and played back with the user’s
hand constrained to follow this predefined trajec-
tory towards the mating location of the component.
The arm and hand motions of the user during
virtual assembly are videotaped. This video is ana-
lyzed for angles between body segments and hance
the upper extremity postures of the user, on the
Peak5 2D motion measurement system. The output
of this analysis serves as feedback to indicate areas
of improvement in the ergonomics of the assembly
process.
Fig. 12. Marker placement and joint angles.
5.2. Case study
An extension to 3D tracking will allow modeling of
joints that move the arm out of the plane to identify A simple jig model imported into the VE was
shoulder, elbow, and wrist abduction problems. evaluated for interference and accessibility. The
setup and interaction of the operator with the sys-
tem are shown in Fig. 13. The loading of a compo-
5. System implementation and case study nent, and gross and fine motion trajectories are
shown in Fig. 14. Fig. 14a shows the initial state
The JIGPRO system is implemented on a Silicon with the component at its pick location. Fig. 14b
Graphics Inc., workstation using the WorldTool- shows the gross motion trajectory as a set of points.
Kit (WTK) VR modeling software from Sense8 These points are fitted into a Bezier curve and
Corporation. Models of assembly tooling (jig and stored for future playback. Fig. 14c shows the
hand tools) are created using PTC’s Pro/Engineer straight line fine motion trajectory obtained from
CAD software package. VR hardware includes the
Power Glove, VR glasses, and associated peri-
pherals. For ergonomic analysis of the gross
motion assembly trajectory, the Peak5 2D motion
measurement system is used.

5.1. System operation

In the prototype JIGPRO implementation, users


load in a Pro/Engineer CAD model of an assembly
tool into the WorldToolKit VE. The design can be
examined from all directions for a preliminary idea
of assembly locations and distances. Components
are loaded into the environment and moved to-
wards their assembly locations. Once the object is
in a predetermined region near the point of assem-
bly, it automatically moves along the fine motion Fig. 13. Simulation of assembly operation on the JIGPRO
trajectory and reaches its mating location. system.
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 485

Fig. 14. Depiction of loading and gross and fine motion trajectories on the JIGPRO system.

the assembly model and Fig. 14d shows the com- information does not reveal any problems in
ponent in the final assembly state. Because the reachability during the assembly task. If such prob-
loading operation involves mating the peg with the lems existed, they could be used to improve the jig
hole and no jig elements interfere with the fine design by eliminating or reducing the need for
motion trajectory, no accessibility problems are working in areas outside the maximum reach envel-
detected. ope of the individual performing the assembly task.
The joint positions and trajectories as recorded
and analyzed by the motion tracking system are
shown in Figs. 15 and 16. The risk of injury to the 6. Concluding remarks
upper extremities increases with increased devi-
ation of body segments and joints from their neu- A prototype VR-based system that allows com-
tral position. The variation in elbow angle, seen in bined evaluation of assembly product and jig
Fig. 15, reveals that certain sections of the jig may designs has been developed. The methodology
have to be redesigned because of deviation from the involves importing the assembly product model
neutral posture. Other information obtained from and process constraints, jig design, and hand tool
this analysis includes linear displacement of the CAD models into the VE, followed by simulation of
upper extremity segments shown in Fig. 16. This the virtual assembly activity. Gross and fine motion
486 V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487

Fig. 15. Angular displacement of upper extremities.

trajectories generated by the user during assembly


of feasible assembly components are captured. Anal-
ysis of accessibility and ergonomics during the as-
sembly process is carried out. In this paper, the
analysis methodology, system implementation, and
a case study outlining its operation have been de-
scribed. This methodology offers a proactive ap-
proach to reducing the risk and associated costs
of musculoskeletal injuries during the assembly Fig. 16. Linear displacement of upper extremities.
process.

Anon, 1995. Computer-aided fixture design. Manufacturing En-


References gineering 114 (6), 4.
Armstrong, T.J., Buckle, P., Fine, L.J., Hagberg, M., Jonsson, B.,
Ahrikencheikh, C., Seireg, A., 1994. Optimized-Motion Plann- Kilbon, A., Kuorinka, I.A.A., Silverstein, B.A., Sjogaard, G.,
ing: Theory and Implementation. Wiley, New York. Viikari-Juntura, E.R.A., 1993. A conceptual model for
Allen, C., Karam, K.Z., Le, C.P., Hill, M., Tindle, T., 1995. work-related neck and upper-limb musculoskeletal disorder.
Application of virtual reality devices to the quantitative Scandinavian Journal of Work, Environment, and Health
assessment of manual assembly forces in a factory environ- 19, 915—933.
ment. Proceedings of the 1995 IEEE International Bridger, R.S., 1995. Introduction to Ergonomics. McGraw-Hill,
Conference on Industrial Electronics, Control, and Instru- New York.
mentation, IEEE, Los Alamitos, CA, U.S.A. CTD News, 1993. Tallying the True Costs of On-the-Job CTDs.
V.N. Rajan et al. / International Journal of Industrial Ergonomics 23 (1999) 473—487 487

Deitz, D., 1995. Human integrated design. Mechanical Engin- Mizell, D.W., 1994. Virtual reality and augmented reality in
eering 117 (8), 92—96. aircraft design and manufacturing. Proceedings of the 1994
Ellis, S.R., 1991. Nature and origins of virtual environments: Wescon Conference, IEEE Wescon, p. 91.
Immersed in a synthetic space. Computer Systems in Engin- Moore, A.E., Wells, R., Ranney, D.A., 1991. Quantifying expo-
eering 2 (4), 321—347. sure in occupational manual tasks with cumulative trauma
Fernandez, J.E., 1995. Ergonomics in the workplace. Facilities disorder potential. Ergonomics 34, 1433—1453.
13 (4), 20—27. Mortenson, M.E., 1985. Geometric Modeling. Wiley, New York.
Fernandez, J.E., Fredericks, T.K., Marley, R.J., 1995. The psy- Owen, J.V., 1994. Making virtual manufacturing real. Manufac-
chophysical approach in upper extremities work. Contem- turing Engineering 33—37.
porary Ergonomics. Prombanpong, S., Lewis, R.L., Bishop, A.B., 1992. Computer-
Fuh, J.Y.H., Nee, A.Y.C., Senthil, K.A., Teo, J.C.S., 1995. IFDA: aided fixture design system with process planning integra-
An interactive fixture design and assembly environment. tion for prismatic parts manufactured on CNC machining
International Journal of Computer Applications in Techno- centers. Computers in Engineering — Proceedings of the 1992
logy 8 (1-2), 30—40. ASME International Computers in Engineering Conference
Hale, J.P., 1994. Applied virtual reality in aerospace design. and Exposition, ASME, New York, NY, U.S.A, pp. 369—380.
Proceedings of the 1994 Wescon Conference, IEEE Wescon, Rajan, V.N., Lyons, K.W., Sreerangam, R., 1997c. Generation of
Los Angeles, CA, USA, pp. 378—383. component degrees of freedom from assembly surface mating
Hancock, P.A., Lai, J., 1995. Virtual reality technology to em- constraints. Submitted to the 9th International ASME Con-
brace information gap in industrial hygiene. Occupational ference on Design Theory and Methodology, Sacramento, CA.
Health and Safety, January, 59—81. Rajan, V.N., Lyons, K.W., Sreerangam, R., 1997d. Identification
Hargrove, S.K., 1995. Systems approach to fixture planning and of requirements and representations for assembly modeling.
design. International Journal of Advanced Manufacturing Submitted to the 1997 IEEE International Symposium on
Technology 10 (3), 169—182. Assembly and Task Planning, Marina del Ray, CA.
Holbrook, D., 1991. Stereo viewing, looking ‘into’ manufactur- Rajan, V.N., Nof, S.Y., 1996. Minimal precedence constraints for
ing. Manufacturing Systems 9 (1), 30—31. integrated assembly and execution planning. IEEE Transac-
Jackson, K., 1996. Virtual Humanity Helps Boost Quality, Cut tions on Robotics and Automation 12 (2), 175—186.
Costs. Crain Communications. Rajan, V.N., Prabhakar, S., Dhananjaya, K., 1997a. A correct
Kelley, D.G., Chu, K.-R., 1987. CAD/CAM Design of a Flexible and complete algorithm for geometric design and analysis of
Fixturing System for Aircraft Tubing Assemblies. American modular fixtures. Transactions of the NAMRI/SME (in
Society of Mechanical Engineers, New York, p. 7. press).
Kim, H.S., Ko, H., Lee, K., Lee, C., 1995. Collision detection Rajan, V.N., Yashvant M.B., Sreerangam, R., 1997b. Virtual
method for real-time assembly simulation. Proceedings of environment for design and analysis of factory layouts. Ac-
the IEEE International Symposium on Assembly and Task cepted for Presentation at the 1997 SAE General, Corporate
Planning, IEEE, Pittsburgh, PA, USA, pp. 387—392. & Regional Aviation Meeting and Exposition (GCRAM),
Lyons, K.W., Rajan, V.N., Sreerangam, R., 1997. Representa- Wichita, KS.
tions and methodologies for assembly modeling. Draft NIST Tichauer, E.R., 1973. Ergonomic Aspects of Biomechanics. In-
Internal Report, National Institute of Standards and Tech- dustrial Environment — Its Evaluation and Control. NIOSH,
nology, Gaithersburg, MD, 20899. Washington, D.C.

You might also like