Adaptive static neural network control of robots
Shuzhi Sam Ge, Zhanlin Wang, Chen Zong-ji · 2002
In the paper, a novel neural network model of robots is presented. Its structural properties, such as the linear-in-the-parameters dynamics, are investigated to facilitate controller design. Since the neural networks are used to model the inertia matrix D(g) and gravitational potential energy P(q) only, they are static networks and the size of the resulting model is much smaller than the dynamic ones. Subsequently, a general controller based on the resulting neural network model is discussed. It can be shown that all the closed-loop signals are bounded and tracking error goes to zero.>