The CMAC Neurocontroller for efficient learning in visual servoing

Abstract

The use of vision has increased the capabilities of manipulator robots, making the problem of control more complex. Aimaing to pathway a target with a robot arm in Cartesien space involves to use precise commands. We propose a neural controller based on CMAC (Cerebellar Model Articulation Controller) networks in a visual servoing. This new structure splits the robot's workspace and assigns different CMAC controllers imposing thus specialized region CMAC. Then, the neurocontroller's sensitivity and precision is increased compared to a single CMAC with the same number of weights. Consequently, Robot positioning and target tracking with visual feedback can be done with a better precision.