Knowledge Acquisition Method for Spacecraft Anomaly Detection Expert System Based on Finite State Machine
A"data-driven and domain knowledge"fusion acquisition method based on the principle of finite state machine(FSM)is proposed for the rapid and accurate acquisition of knowledge in spacecraft anomaly detection expert systems.Firstly,based on structured data such as telecommands,telemetry,and normal threshold in the database,establish the state set,input set,and output set.Secondly,considering the changes in equipment operating conditions as state variables and using injection,telecommands and other events as input signals,a mapping relationship between telemetry changes and state transfer is established,which generates expert system knowledge by solving state transition functions.Finally,using historical telemetry data to drive finite state machines for state migration,a set of state transition functions will be gained,which can generates actionable knowledge.Taking the telemetry of TT&C of China Space Station as an example for method experiments,the experimental results show that the anomaly detection knowledge obtained by the proposed method has strong interpretability and engineering applicability,which can provide reference for the acquisition of detection knowledge for subsequent satellite engineering tasks.
NETL
NSTL
万方数据
