Visual Servo Control of a Two-axis Turntable Based on Fuzzy Cerebellar Model Articulation Controller

In visual servo control system, the design of visual loop is a key link to precise tracking. To improve tracking accuracy, we discuss a FCMAC-based (fuzzy cerebellar model articulation controller) control strategy. The controller includes a basic proportional regulator and a FCMAC controller. The two blocks work together to approximate the visual-mapping model through an adaptive online learning law to adjust the weights of the network. By mapping the 2-D image characteristic space to the 3-D kinematic space, we can achieve the control task of a turntable directly through visual information. Applying the proposed method to a two-axis turntable to track a dynamic object, experimental result indicates that FCMAC is capable of accurately approximating visual-mapping model. The visual system shows a good performance in both accuracy and robustness.