Abstract
Conventional model-based computed torque control fails to produce good trajectory tracking performance in the presence of payload uncertainty and modeling error. The problem is how to provide accurate dynamics information to the controller. A new control architecture that incorporates a neural network, fuzzy logic and a simple proportional-derivative (PD) controller is proposed to control an articulated robot carrying a variable payload. A feedforward (multilayer) neural network is trained off-line to capture the nonlinear inverse dynamics of the system. The network is placed in the feedforward path to minimize tracking error. The network receives the same input signals as conventional computed torque as well as the payload mass estimate, which comes from a fuzzy logic mass estimator. The fuzzy logic, trained off-line to optimize the membership function, is developed to estimate the changing payload mass. The fuzzy logic estimator is based on joint acceleration error to improve the speed of detection and estimation of payload mass change. The effectiveness of the proposed architecture is demonstrated by experiment on a two-link planar manipulator with changing payload mass. Experiment results show that this control architecture achieves excellent tracking performance in the presence of payload uncertainty. The results of the control architecture are also compared with those of a model-based control architecture. This approach can be employed in any nonlinear mechanical system with a sudden change in a parameter.
Design of fuzzy logic controllers for Takagi–Sugeno fuzzy model based system with guaranteed performance
