| Space science satellites have produced major scientific breakthroughs and discoveries,and their achievements of high reliability and safety indicators for payload system rely on adequate and effective ground-based test validation and experimentation.Payload system Ground Integrated Test(GIT)is an important part for validating the design of payload system.To planning test cases for payload system is a complex and comprehensive designing task,which requires considering equipment/system level design,such as functional design,working mode design,remote/telemetry command design,and testing requirements comprehensively.With the development of space science and technology,and the deepening of exploration in the field of space science,the scientific tasks undertaken by payload system are increasingly complicated.In order to improve the efficiency of test case planning and the effectiveness of test cases,the method of planning test cases intelligently for payload system was studied thoroughly,a test case intelligent planning architecture based on knowledge-based payload system was designed,and a method based on combining ontology-based knowledge model and swarm intelligence algorithm was proposed.First of all,the knowledge model of the payload system was construted based on the analysis of the knowledge characteristics of the payload system and the test requirements of the GIT.In-depth study of payload system characteristics,the Hierarchical Knowledge Structure Payload System(HKSPS)was proposed,which divided the knowledge space of payload system into four levels: system-level knowledge,element-level knowledge,behavior-level knowledge and meta-level knowledge.The knowledge structure and formal model of each level are analyzed to help construct the domain ontology knowledge model of the payload system,and the knowledge of macro-behavior and micro-structure of payload system is fully described,which provides a knowledge base for test case planning.Secondly,according to the domain ontology of payload system,the intelligent planning method for functional sequence was designed.Since the traditional backtracking algorithm had low search efficiency when solving the functional sequence planning problem,an improved Co-evolution Genetic Algorithm(CGA)adopted "Worst Individual Mutation"(WIM)strategy was proposed and named WIM-CGA.WIM-CGA had the advantage of " Life Time Fitness Evaluation " of the original CGA,and employed a two-way evolutionary scheme,which was,"better individuals perform standard genetic processe,and worse individuals perform mutation operation",to improve the solution accuracy and search efficiency.Simulation results show that the proposed algorithm WIM-CGA has better search performance than CGA algorithm under all the problem size tested.Finally,the functional priority intelligent planning method was studied in order to improve the validity of test cases and optimize the test cases.Since the Ant-Q algorithm had low search efficiency and was easy to fall into local optimum when solving multi-objective priority problem,an improved algorithm MMAS-AQ algorithm is proposed.On one hand,the Max-Min Ant System(MMAS)was applied to improve the standard ant colony optimization algorithm,and the pheromone calculation method,boundary setting and transition probability were redefined.On the other hand,an adaptive small batch updating rule was proposed to improve pheromone updating strategy.Simulation results show that the proposed algorithm MMAS-AQ can effectively improve deficiency of local optimum and has higher search efficiency under all the problem size tested.The method proposed was verified by applying to a payload system of a certain type of scientific satellite.The experimental results show tha the knowledge model of the payload system can fully describe the domain knowledge and provide complete knowledge for test case planning,and the test case planning efficiency is higher.Planning a test case with a function scale of 25 takes about 3.20 second.Meanwhile,the APFD index of the test case is 2.88 percentage points higher than that of the test cases without optimizing,whith means the test case obtained by the planning method has higher fault detection ability.The intelligent test case planning method proposed in this paper meets the design requirements of test cases in the GIT of payload system.It can improve the efficiency of test case planning and the effectiveness of test cases,and provide theoretical basis and references for the intelligent development of the GIT technology of payload system. |
