IJCA Vol 4 i1 2025 webmag - Flipbook - Page 32
32
The International Journal of Conformity Assessment
The primary objective of this study
was to determine the appropriate
detection technology for each
CoP component. A detailed
analysis was carried out to identify
the most effective solution for each
component class.
The results of this contribution
have far-reaching implications for
the entire automotive industry.
With the increasing diversity of
vehicle variants and the tightening
of regulatory requirements,
automating the homologation
process is crucial to ensure
production compliance and
minimize recalls. The identi昀椀cation
and evaluation of suitable
automation solutions, as presented
in this contribution, provide
valuable insights for automotive
manufacturers.
Figure 14: Illustration of the assignment of the detection technologies to the
homologation components (excerpt)
sampling inspections cover only a
limited portion of the components.
As a result, automotive
manufacturers face recurring
recalls and must ensure the safety
and quality of their products in
accordance with legal regulations
(Bratzel, 2021).
One of the main objectives of this
contribution was to identify and
evaluate a suitable automation
solution for the component
identi昀椀cation process.
The results show that both
transmitter-receiver systems (e.g.,
RFID) and optoelectronic systems
(e.g., ICR/OCR) are viable options
for automating the homologation
process. However, as current
regulatory requirements mandate
that component labeling be visible
and not encoded (Certi昀椀cation and
Accreditation Administration of the
People’s Republic of China (CNCA),
2020), transmitter-receiver systems
such as RFID are not compliant at
this time. Nevertheless, they were
included to account for potential
future changes in regulatory
standards.
In the evaluation of optoelectronic
systems, GPT-4v demonstrated
particularly strong performance,
correctly detecting a high
percentage of components under
relevant conditions. Hardware and
software improvements enabled
100% accurate detection for
more than half of the evaluated
components (see Figure 12).
By implementing the recommended
detection technologies, such
as optoelectronic systems like
GPT-4v, the manual component
identi昀椀cation process can be
transformed into a digitized
veri昀椀cation process. This
transformation improves e昀케ciency
and accuracy in the 昀椀eld of
homologation assurance. The
昀椀ndings can be also applied in
other quality assurance processes.
Further research is needed to
de昀椀ne potential process steps
for detecting component IDs
(homologation labels) within the
CID work昀氀ow. An overall process
concept should be developed for
all 300 components (BMW Group,
2021) in the product development
process, incorporating the
recommended detection
Implementing these technologies
would transition the manual CID
process into a digitized veri昀椀cation
process. This transformation
will require employees at BMW
Group and its suppliers to adapt
