SAMPE Journal - November/December 2019 - 41

such a system could provide. Feedback from these
trials was generally positive, which is encouraging
given that it is still at the prototype stage. The key
feedback lies in the simplicity of use and clarity of
the instruction steps. Almost every person who had
seen the system demonstrated or took part in the
trials could see the potential usefulness of a system
like this, albeit with further development, to reduce
some current operational inefficiencies.
Other useful feedback from the trials and
demonstrations of the initial software were its
limitations. Due to the equipment being low-cost
and requiring a complicated setup and calibration
the system is an order of magnitude less precise
than laser ply projection. This can be mitigated
somewhat using higher resolution camera and
projector systems, but this will also dramatically
increase the system cost. The Kinect™ also has
some accuracy issues, and some users struggled
with the interactive controls (mitigated in later
trials with backup keyboard controls).
In particular trials at the National Composites
Centre, which consisted of both graduates/year in
industry students and technical staff, gave some
interesting insights into both the hand layup process
and the use of AR instruction. The trial consisted of
a simple comparison between no instruction and
LayupRITE instruction. Both information types
were filmed, and the videos analysed post-trials, to
capture the route used to lay down the composite
ply as well as the time taken to achieve layup to a
typical acceptable standard. Table 1 shows a clear
improvement between no instructions and using
LayupRITE, as would be expected. However the
data also presents interesting outcomes in terms
of the reduction in the range of time taken to lay
down a ply. This reduction in spread could be
used to give better estimates of manufacturing
times and improve layup standardisation. This
increased standardisation was verified by where
the laminators first put pressure on the ply. This
point of first pressure effectively represents the
start point for the layup. In the uninstructed
control group, the variation in starting point was
as shown in Figure 7. Current instructions will not
allow the user to advance without completing the
current step, thus the starting points were all in the
same area, for all participants.

Figure 5. Projected user interface, as seem by the
laminator in normal lighting conditions.

Figure 6. The system being trialled at Weston College.

FUTURE SOFTWARE DEVELOPMENTS
There remains several potential avenues for
development of the current system, in order to
develop the system into a useful product. The
chief area remains in terms of feedback - using
the system to not only capture data about the
laminator, but also carry out in-process inspection.
This can potentially be achieved using the current
hardware with limited accuracy or with additional
equipment, but this would increase the cost of
the system. Secondly - there is a need to develop
training content for the system to be a viable
training methodology. Integration into an existing
training program would be a good method to test
its fitness-for-purpose in a relevant training setting
in the short-term, with a view to overhauling
composites training for a more interactive,
digitally-enabled era.

Table 1. Comparison of layup timings between uninstructed and LayupRITE instructed participants.
Maximum
Minimum
Average
Range

w w w. s a m p e . o r g

Control (s)
662
188
361
474

LayupRITE (s)
474
176
326
298

N OV E M B E R / D E C E M B E R 2 0 1 9

% Improvement
28.4
6.4
9.9
37.1

SAMPE JOURNAL

http://www.sampe.org

SAMPE Journal - November/December 2019

Table of Contents for the Digital Edition of SAMPE Journal - November/December 2019