IJCA Vol 4 i1 2025 webmag - Flipbook - Page 31
2025 | Volume 4, Issue 1
When comparing the performance
of different tags on different
materials, it has been found that
the Dogbone tag requires the
highest power for activation,
especially on thermoset/
thermoplastic materials where the
activation power is at 30 dBm.
In the case of metal, only the OnMetal tag (11dBm) and the Flag
tag (24 dBm) worked. Neither the
Miniweb tag nor the Dogbone
tag could be detected on the
metal component in any position,
regardless of the transmission
power, indicating that there was
interference due to the physical
properties of the material.
Component Class 3 and
Component Class 4 tags required
less than 15 dBm on average for
activation. The results show that
higher received signal strength
indication (RSSI) values can
be achieved with increasing
transmission power. The PoC has
shown that RFID technology is
applicable to CoP components,
and there are specific tags for each
component class that enable a
100% pass rate. Thus, it is proven
that RFID tags are suitable for
digitizing all component classes.
The next step is to evaluate the
results under real conditions.
Deployment Under Real Assembly
Conditions:
An evaluation of the results from
the ideal experiment is performed
directly under real assembly
conditions. The execution is carried
out based on the defined boundary
conditions in Section 5.1.2. The
tags are evaluated according to the
defined component classes (see
Section 4.2) at the production line.
Under real conditions, both
the On-Metal tag and the
Dogbone tag require increased
activation power for thermoset/
thermoplastic materials.
For Component Class 1 (metal),
similar results were obtained as
under ideal conditions. Neither the
Miniweb tag nor the Dogbone
tag could be detected on the
metal component in any position,
regardless of the transmission
power. In general, all other
component classes consistently
require higher activation energy
under real production conditions.
Component Class 3 and
Component Class 4 consistently
require higher activation energies
under real production conditions
(averaging above 15 dBm).
Overall, satisfactory results could
only be achieved at the highest
power level of 30 dBm. The
results from the ideal PoC could
not be replicated in the tests
conducted in the real production
environment. It can be concluded
that for all tags and materials in
the real production environment,
higher transmission power
leads to better and more reliable
results. The PoC for transmitterreceiver systems has confirmed
that a suitable tag is available
for each component class.
6 Applicability of the Detection
Technologies to the CoP
Components
Section 6 analyzes the applicability
of the evaluated technologies
to determine which technology
are best suited for specific CoP
components.
The results of the PoCs (see
Section 5.2) indicate that both
transmitter-receiver systems and
optoelectronic systems can be
applied for the digitalization of the
homologation process.
For transmitter-receiver systems,
appropriate tags were identified
31
for each component class. In the
case of optoelectronic systems,
OCR models demonstrated strong
performance—achieving over 98%
success rates for easily legible
components, such as labels.
However, for other component
classes, frequent identification
deviations were observed. In
comparison, GPT-4v achieved
an even higher percentage of
successfully tested components
(see Section 5.2).
The results from the PoCs—
specifically those that reached
100% correct detection—have
been mapped to the component
classification classes defined in
Section 4.2. A sample of these
results is illustrated in Figure 14.
Figure 14 assigns the previously
de昀椀ned component classes
(see Section 4.2) to the most
promising detection technologies
that achieved 100% detection
accuracy. The right-hand column,
"Technology Classi昀椀cation,"
identi昀椀es which detection
technologies can be used for CID
veri昀椀cation of the component ID
(homologation label). This 昀椀gure
shows an excerpt from Component
Class 1 (metal, molding), which
has been assigned to the most
effective evaluated detection
technologies.
7 Summary of the Results and
Outlook
The increasing diversity of vehicle
variants and the tightening legal
requirements in the automotive
industry continue to pose
signi昀椀cant challenges for ensuring
production conformity (Sabadka
et al., 2019; Sonya Gospodinova;
Federica Miccoli, 2020). Previous
studies (Sturm, 2023) have shown
that component IDs (homologation
labels) do not always meet legal
standards, in part because manual
