Hardware redundancy to increase coverage of dangerous failures

Hardware redundancy to increase coverage of dangerous failures 

To gain in safety level, the coverage rate of critical faults must be increased. 

The two processing units have the possibility to put the system in a safe state via two independent (distinct) channels. 

It is important for this type of safety based on self-tests, to have an application that is as robust as possible regarding transient faults that are not covered by periodic self-tests. This is the case for the majority of current automotive applications. For example, data stored in memory are refreshed before their impact becomes critical. 

The test intervals on the main unit must be defined to allow the second processing unit to put the system in safe condition before the spread of a defect has become critical. 

---