Improvement of Large Data Acquisition Method without the Interference on the CPU Load for Automotive Software Testing

Some black-box testing methods utilize memory update frequency to solve the difficulty of fault localization during system integration testing for automotive software. They need memory information, but the strict test policies and environments don't allow to use existing debugging tools. The proposed data cascading algorithm could deliver large memory information within the limited bandwidth. However, it caused a performance issue that took a long time to get a large volume of data while trying to leave the bandwidth. Moreover, it also caused abnormal latencies of the system when the processor load of System-Under-Test (SUT) is high. Then we propose an improved test agent to address these issues. It consists of traffic-load model construction, a variable test agent that manages its processor load by adjusting its traffic, compensation algorithm and standard traffic determination algorithm. We conducted a series of experiments to validate the proposed method with a simulated Body Control Module (BCM) in real hardware configuration. As a result, our method can deliver memory information completely and reduces 30% of the total testing time.