Affective interactive agents with applications in control performance assessment

Based on Belief,Desire,Intention(BDI) model of intelligent agents,a kind of human-computer Affective Interaction model(AI-BDI) was proposed.According to affect intensity space representations of Hidden Markov Models(HMM) and frameworks of personal Ortony,Clore,Collins(OCC) models,as well as combining implications of a kind of human psychological stimulation model,a stimulated and transferring affect computing model was set up,and agent affect computation method was presented in detail.The model was further introduced to construct a human-computer interactive learning model.Human-computer interactive objectives achievement degree and the computation as well as learning algorithms were proposed.By taking the process control performance assessments as example,construction and relevant algorithms of human-computer interactive model were discussed.Two specified control loops were simulated and analyzedt to demonstrated the feasibility and effectiveness of proposed method.