The effect of time-between-events for sequence interaction testing of a real-time system

Automotive electronic control systems are expected to respond to input demands in real-time (circa: milliseconds) to ensure occupant and road user safety and comfort. System complexity and real-time computing requirements create significant challenges in proving the robustness of control systems; here robustness is the degree to which a system can function correctly in the presence of unexpected inputs. Evidence shows that faults still escape to customers incurring large warranty costs. Existing test methods can be ineffective in testing robustness with the primary focus being on requirements validation. Evidence from other industries such as IT and medical suggests faults that are difficult to find, manifest due to complex interactions and sequences of events. Whilst model based approaches are becoming more prevalent, they are still an abstraction of the real world and deal with 'time' at this conceptual level. This can leave some doubt on their validity for representing real-time testing. However, there is little evidence within the literature pointing to effective sequence interaction testing (SIT) within real-time test environments. The novelty of the proposed approach is a methodology for creating and running t-way input sequence interaction test suites in real-time with the time between input events (TBE) considered as a critical test parameter. The effect of TBE on the triggering of faults for a safety monitoring system deployed in a prototype embedded electric machine control unit is presented to demonstrate the approach. The system is tested with 2688 3-way event sequence interaction tests across a TBE range from 1 millisecond to 500 milliseconds validating the effectiveness of the approach in eliciting faults