Results

Errors were not seeded in this exercise (we already know we can detect them from our original work) and were detected this time too. It’s also not an exact replication of the study due to the obvious fact that there is only D-RisQ doing this, so side-by-side comparisons were not possible, and we used different/new tools and there are now system requirements as well as code and more verification artefacts. In the original exercise, the savings were consistently 60-80% just for the time taken to verify the design against the requirements and were measured in labour hours.

Now we can check for example TA from the graphic above in around 6 minutes compared to the original effort of 3 hours 25mins...which in itself was already considerably faster than conventional (~7 hours) or other techniques (~8 hours). We also autocoded and verified in less than 20 mins, 15 minutes of which was setting up the autocoder.

Note that all the tools can be used to provide the requisite evidence needed to support claims against ISO26262.

*Mobilise is a grant funded programme run by the Advanced Propulsion Centre and Zenzic