Mutation-based simulation test data generation for testing complex real-time software

Simulation is proven to be an effective way for testing complex real-time hardware/software systems. However, simulation test data generation is a challenging issue which determines the efficiency in such testing. This paper proposed a data-driven mutation based approach for test data generation to address the unique requirements of realtime, input domain coverage, adaptability, and reliability. The architecture is designed to simulate the process for data sampling, transforming, packaging, and transmitting segments in an aerospace ground control system. Various mutant operators are defined for signal generation and data package generation to produce fault-sensitive inputs based on the system's fault pattern analysis. Mechanisms are defined to generate a large number of test cases by combination and composition of various data outputs from different dimensions to achieve a high coverage. A configuration mechanism is introduced to enable recomposition and re-combination of simulation components. The selective simulation testing method is discussed to improve test effectiveness. A case study shows that the proposed approach can achieve a high fault-coverage with a small number of effective test cases.