Design And Implementation Of Rule Based Uncertainty Reasoning For Smart House

The essence concept of smart home system is the integration of multiple domestic automation systems, namely passive response within a few simple remote control operations. With the development of artificial intelligence, ubiquitous computing and device miniaturization technologies, the concept of smart home evolved to the idea of smart home environments. A core technical challenge is to identify environmental conditions and user’s behaviors, and infers the results automatically. However, on account of the reliable data source, unstable transmission networks, and knowledge limited systems, etc, there is a lot of uncertainty in the smart home system, and therefore, how effective reasoning under uncertainty is a major technical challenge.In this thesis, taking into account rule-based knowledge representation model’s better readability and modularity(separation of decision logic and inference), this paper developed a uncertainty knowledge representation method. Rule-based knowledge representation model typically used to describe the upper process or the user’s intent, it has been widely applied in existing commercial reasoning and inference engines, on account of its intuitive knowledge representation structure. Most existing research on rule-based smart home does not take into account the beingness of uncertainties in knowledge representation methods. Therefore, as an important research emphasis in the thesis, the solution of dealing with the uncertainty in rule-based knowledge representation, and break the poor reasoning ability and other restrictions, e.g., the certainty factor is incompatible with the probability theory has become necessary.This thesis proposed a smart home uncertainty reasoning scenario, which use the Event-Condition-Action framework, and together with an efficient inference scheme which incorporates reasoning with uncertainty probabilities. In addition, the author construct inference networks abstracted from rules. On this basis, the author according to the inference networks exact and inexact reasoning algorithm. Finally, experimental results show that for smart homes applications of uncertainty reasoning method based on rules is full of effectiveness and practicality.