Feeling AI (FAI) servicing the robots as a kind of Hybrid AI based on cognitive models borrowed from living beings is discussed. Since the FAI knowledge base presented by a set of independent Knowledge Granules (KGs) communicating by messages may be obtained from the original presentation by the response prototypes, the Event-Driven Architecture (EDA) is most suitable to carry out the processing. The major problem of applying the EDA is the necessity of real-time response-making by FAI's components on the base of sensory data obtained at different times and having different characteristics of aging. A new cognitive model of data aging, which supports the gradual forgetting over time of the events and assesses its current relevance by fuzzy Certainty Factor (CF), is introduced to overcome the above issue. Implementation of the EDA as underpinning of the pipeline processing significantly simplifies real-time FAI computing. Computer experiments with a robot at the intersection show how the FAI response-making engine with the model of CF, taking into account data aging, overcomes the problem by aligning the CFs to the current point of time.