A Fault Diagnosis Method for Smart Home Services

Smart home service has been an emerging and profitable new business for IT and telecomm corporations. The nature of integrated heterogeneous gadgets and devices makes smart home system builders face unprecedented challenges in maintenance tasks. In this paper, we propose a symptom-problem correlation model to locate system faults in smart home services. First, we investigate the tree-like structure of a smart home system's objects which reveal their dependency among one another and deduce instinctively a coarse object state to a problem correlation model that relates a set of symptoms to a particular problem. Then the proposed algorithm is applied to verify and reduce the size of this model so that the model can accurately identify the defined problems using minimum number of observed objects' states. The simplified model can be utilized for automatic fault location, optimizing Home Gateway configuration, and boosting system performance.