Towards Complete State Machine Traversal via Pseudo Transitions in Automated Lab Verification

The complexity of modern automotive smart power devices increases due to their growing amount of functionality. The resulting extensive internal state machine requires more verification time especially in the mixed-signal post-silicon domain, which still relies on manually intensive directed testing. A state transition based methodology is proposed, that automatically finds a verification path through the complete state machine and ensures that each transition, and hence each state, is traversed at least once. By extending the device's state machine with pseudo transitions, that include information about the required lab setup configuration, a correct measurement of each requirement in question is ensured. Advanced graph theoretic algorithms can now be applied to cover all requirements that are associated to the state machine. In an industrial case study with an automotive smart motor controller it is shown, that the methodology finds the optimal traversal through all 82 transitions of the device's state machine. The automated adaption of the measurement setup, with regard to the load circuit and adequate measurement equipment selection, enables a correct verification of the requirements. This approach allows a lab-setup-agnostic automated formal verification of over 80% of the functional requirements and hence drastically reduces the time required for the formal mixed-signal post-silicon verification.