Trustworthiness Model Representation
recalculates the Trust Score for each
workstation and those Scores can be
combined to produce an integral Trust Score
for the end-to-end production process.
Business goals ● Monitor and measure manufacturing quality metrics
● Manage labor costs
Core Requirements ● Remotely manage factory operations (smart factory as a
service)
● Capture and analyze video and sensor data for the
manufacturing process
● Protection of the video data to comply with a variety of
legal restrictions on employee video in various countries
Key Trustworthiness
Characteristics ● Reliability
● Security
● Safety
Attributes ●
●
●
●
●
Video data
Sensor data
Network connectivity
Worker identification
System uptime
Table 1: Practical application of Trustworthiness in a Smart, Connected Factory
data streams reliably. The manufacturing
team was able to diagnose the issue based
on the immediate notification provided by
the decrease in the overall score
supplemented by the detail of the
component scores. This example also
illustrates that the team has weighted the
attribute of Reliability more heavily than the
attributes of Security or Safety.
If any single input changes (for example, a
video stream is missing or sensor data is
interrupted, or the wrong worker is at the
workstation), this may affect not just that
workstation but the entire production line
and the end-to-end process. The operator
decides if the change in input was due to a
known (trusted) event or not. Combining this
kind of environmental information with the
dynamic Trust Score calculation is important
for the proper functioning of the Trust Score
system.
Figure 8 illustrates how the trust model
measured the decrease in system
trustworthiness when network connectivity
was insufficient to transmit all the required
- 39 -
IIC Journal of Innovation