Fuzzy Inference Considering Data Aging in Smart Rules Engine
Anatolii Kargin, Tetyana Petrenko · 2020
When developing the Internet of Things and Smart Systems, framework based on pattern technology are used. The Rules Engine (RE) pattern implements the functionality of the system under development. A new model Smart RE (SRE) with a two-stage data processing is discussed. The Abstracting Engine (AE) on a first processing stage, are integrated with RE which implements traditional Fuzzy Logic System (FLS) on second stage. The AE maps the data from sensors into facts (words) that reveal the meaning of the data. The fact model takes into account various types of uncertainty, including those associated with data aging over time. The Certainty Factor (CF) as a numerical characteristic of a fact is an input variable of a fuzzy inference mechanism at the second stage of data processing in RE. This article discusses the requirements for the knowledge representation (rules and facts) and the features of the RE operating (fuzzy logic inference), allow using the quantitative estimates of data aging effect in FLS. Three FLS encapsulation options as RE in SRE on an example of application of a Smart Traffic Light system are considered.