25.07.2014 Views

pdf download - Software and Computer Technology - TU Delft

pdf download - Software and Computer Technology - TU Delft

pdf download - Software and Computer Technology - TU Delft

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Chapter 6<br />

Conclusions <strong>and</strong> Recommendations<br />

The complexity crisis is most visible in complex, multi-disciplinary embedded systems. Although<br />

individual components might work well separately, the system as a whole may exhibit unexpected<br />

faults. More <strong>and</strong> more components are bought third party <strong>and</strong> might not be used as the supplier<br />

expected when developing them. There is a complex dynamic interaction between subsystems, <strong>and</strong><br />

this is what makes fault diagnosis difficult <strong>and</strong> time consuming. The Philips Cardio-Vascular X-<br />

Ray system is a complex, multi-disciplinary embedded system. The current efforts for doing fault<br />

diagnosis at PMS is increasing; the portion of the PMS employees working on service instead of<br />

development grows. This thesis presents possible improvement when using a model-based approach<br />

to fault diagnosis.<br />

6.1 Conclusions<br />

In this thesis, it is shown that the current practice to fault diagnosis at PMS is not optimal with<br />

respect to the items that an ideal fault diagnosis technique would have; accuracy, speed of diagnosis,<br />

uncertainty of diagnosis, context independency, development costs, runtime costs, explanation<br />

facility, <strong>and</strong> adaptability. Furthermore, model-based diagnosis is chosen as the automated solution<br />

that could improve fault diagnosis, because this technique to automated fault diagnosis is likely to<br />

possess these items that, when present in fault diagnosis, improve dependability of the target system.<br />

Unfortunately, it is not possible to conclude that MBD improves fault diagnosis with respect to<br />

higher dependability of Philips Cardio-Vascular X-Ray Systems. In order to so, it must be possible<br />

to determine the accuracy of both the current practice to fault diagnosis, <strong>and</strong> the proposed technique<br />

MBD. This cannot be done. The reason for this is that it is impossible to find out what faults caused a<br />

certain failure. Although, in many cases, it is possible to determine the observations at the moment<br />

of failure (the error detection mechanisms in the Philips Cardio-Vascular X-Ray System produce<br />

clear log entries), it is impossible to determine the actual fault that caused the specific observable<br />

values. Without this information, the accuracy of the current practice, <strong>and</strong> a MBD implementation,<br />

cannot be quantified. In other words, the significant problem is the inability to determine what<br />

components recovered a certain failure.<br />

Entropy is a valuable metric to show the improvement of different MBD implementations. In the<br />

case study, the diagnostic performance of three MBD implementations that can diagnose the beam<br />

propeller of the frontal st<strong>and</strong> is estimated. The entropy gain of implementation MBD-2 compared to<br />

MBD-1 is zero, thus implementing MBD-2 is not useful. Adding extra sensors to the target system<br />

69

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!