Advances in Intelligent Health Reasoning and its Application to IBDM

We report on an experiment of using model-based diagnosis (MBD) for advanced health monitoring of the hard-docking system of the international berthing and docking mechanism (IBDM). The results of this experiment illustrate our approach in solving two major problems of MBD: modeling an artifact of non-trivial size and managing the representation complexity for finding a root cause of failure. The qualitative model of the hard-docking system has 425 variables (161 components) and 1328 clauses wherein the diagnostic reasoner is capable of determining a quadruple failure in less than 6 s on a midrange PC. Our conclusion is that the qualitative modeling language LYDIA and the automated reasoning framework UPTIME are capable of delivering of sound and accurate diagnosis imposing only a moderate computational burden, making this technology attractive for future health monitoring applications.