The Research On Parameter Adaptive Fuzzy Control Strategies For Robotic Visual Servo System

In this paper, aiming at the task that the three-joint eye-in-hand robotic visual servo system picks a fixed object, some self-adaptive fuzzy control method such as fuzzy control method, parameter self-adaptive fuzzy control, fuzzy neutral network control are studied to conquer the influences of nonlinearity , coupling and uncertainty existed in robotic visual servo system.Fuzzy control is designed directly by human specialist's control experience, its design doesn't depend on the model of controlled plant, therefore, it can effectively control the complex, nonlinear, coupling and severely uncertain system. In this paper, human specialist's control rules for robotic visual servo system are summarized, and a fuzzy robotic visual servo controller is designed.Aiming at the limitation that ordinary fuzzy controller's parameter doesn't adjust online under the condition of system functioning and its self-adaptive capability is very bad. Two parameter self-adaptive methods are designed in order to control the robotic visual servo control system. The one is subsection parameter adjustment which need to carry through subsection adjustment for quantification factor and proportion factor; the other is function mode parameter adjustment which adjusts online quantification factor and proportion factor by Gauss function.By merging fuzzy control and neutral network technologies together, and realizing fuzzy reasoning by neutral network, a kind of fuzzy neutral network robotic visual servo controller is proposed in this paper. In this network, Gauss function is used as fuzzy membership function. This controller not only has fuzzy control's characters that it doesn't depend on the controlled object's characters, such as accurate mathematic model, strong Robust ,simple and pellucid arithmetic, etc, but also has neutral network's learning ability, so it can effectively solve the robotic visual servo system's uncertainty question.For each proposed method, simulation experiments are carried out to prove its efficiency. Results of experiments show that each proposed self-adaptive fuzzy control model has good control characteristic, and can also gain good control effect when each model is applied to robotic visual servo system.