Bounded control of robotic manipulators via fuzzy logic systems

In this paper, a fuzzy neural network controller is proposed for a class of robotic manipulators with unknown dynamics and input constraints. First, to cope with the unknown dynamics, fuzzy logic systems are designed to approximate it. Second, considering the case where an aggressive control input may lead to the poor performance, even resulting in instability in practice, the controller is designed with the function tanh(·), and the function has the ability to make the controller keep in the predefined range. At last, evidence based on the Lyapunov theory proves the system errors converge to a small range near zero. Simulations are carried out to show the effectiveness of the designed control.