Grey-prediction self-organizing fuzzy controller for robotic motion control

A self-organizing fuzzy controller (SOFC) under system control has online learning capabilities; nevertheless, the SOFC may excessively modify its fuzzy rules when its learning rate and weighting distribution are inappropriately selected. This results in oscillatory phenomena in the system during the control process. Moreover, the SOFC is mainly used to manipulate single-input single-output systems. When it is used to handle robotic systems, which are multiple-input multiple-output systems, the dynamic coupling effects between degrees of freedom (DOF) in the robotic systems are difficult to eliminate. To eliminate these problems, this study developed a grey-prediction self-organizing fuzzy controller (GPSOFC) for robotic systems. The GPSOFC introduces a grey-prediction algorithm into the SOFC to pre-correct fuzzy rules to reasonable ones for the control of robotic systems. This solves the problem caused by the inappropriate selection of parameters in the SOFC and compensates for the dynamic coupling effects between the DOFs in the robotic systems. To evaluate the feasibility of the proposed method, this study used the GPSOFC to manipulate a 6-DOF robot to determine its control performance. The GPSOFC yielded better control performance than the SOFC for robotic motion control, as shown in experimental results.