Fuzzy tuning of Brain Emotional Learning Based Intelligent Controllers

This paper presents a fuzzy parameter assignment method for Brain Emotional Learning Based Intelligent Controller (BELBIC). In the proposed methodology, a Sugeno fuzzy inference system (FIS) is applied to tune the parameters dynamically during the control procedure considering main characteristics of plant like error and its derivative. Human knowledge and experiences is used to extract fuzzy rules. These rules determine when and how much change (or even none) should take place for each parameter. It has applied to control a 2-DOF rehabilitation robot while tracking different reference trajectories. It has concluded that changeable parameters provide better performance in different conditions of a particular control trend in comparison to rigid setting of BELBIC parameters. Some approaches have introduced to discuss system stability. Computer simulations are performed to verify analytical results.