The System Identification of Nonlinear Dynamic System Using Universal Learning Network with Multi-branches

摘要

When we identify the nonlinear system, traditional method needs much information of the system, but the control system is more and more complicated, it is difficult to get much information of the system. The artificial neural network has the ability to approximate the nonlinear function in any extent, so it provides a new method of system identification. However, the present neural network has shortcomings. To solve the problem, this paper proposes the Universal Learning Network (ULN). This network has characteristic as follows: (1) all of the nodes are connected each other; (2) there are multiple branches between every two nodes; (3) arbitrary time delay can be set on every branch. Using ULN to identify the nonlinear dynamical system, it can be proved that this network has excellent learning and general ability.