TY - JOUR

T1 - Universal learning network and its application to robust control

AU - Hirasawa, Kotaro

AU - Murata, Junichi

AU - Hu, Jinglu

AU - Jin, Chunzhi

PY - 2000/12/3

Y1 - 2000/12/3

N2 - Universal learning networks (ULN's) and robust control system design are discussed. ULN's provide a generalized framework to model and control complex systems. They consist of a number of inter-connected nodes where the nodes may have any continuously differentiable nonlinear functions in them and each pair of nodes can be connected by multiple branches with arbitrary time delays. Therefore, physical systems which can be described by differential or difference equations and also their controllers can be modeled in a unified way. So, ULN's constitute a superset of neural networks or fuzzy neural networks. In order to optimize the systems, a generalized learning algorithm is derived for the ULN's, in which both the first order derivatives (gradients) and the higher order derivatives are incorporated. The derivatives are calculated by using forward or backward propagation schemes. These algorithms for calculating the derivatives are extended versions of back propagation through time (BPTT) and real time recurrent learning (RTRL) by Williams in the sense that generalized nonlinear functions and higher order derivatives are dealt with. As an application of ULN's, the higher order derivative, one of the distinguished features of ULN's, is applied to realizing a robust control system in this paper. In addition, it is shown that the higher order derivatives are effective tools to realize sophisticated control of nonlinear systems. Other features of ULN's such as multiple branches with arbitrary time delays and using a priori information will be discussed in other papers.

AB - Universal learning networks (ULN's) and robust control system design are discussed. ULN's provide a generalized framework to model and control complex systems. They consist of a number of inter-connected nodes where the nodes may have any continuously differentiable nonlinear functions in them and each pair of nodes can be connected by multiple branches with arbitrary time delays. Therefore, physical systems which can be described by differential or difference equations and also their controllers can be modeled in a unified way. So, ULN's constitute a superset of neural networks or fuzzy neural networks. In order to optimize the systems, a generalized learning algorithm is derived for the ULN's, in which both the first order derivatives (gradients) and the higher order derivatives are incorporated. The derivatives are calculated by using forward or backward propagation schemes. These algorithms for calculating the derivatives are extended versions of back propagation through time (BPTT) and real time recurrent learning (RTRL) by Williams in the sense that generalized nonlinear functions and higher order derivatives are dealt with. As an application of ULN's, the higher order derivative, one of the distinguished features of ULN's, is applied to realizing a robust control system in this paper. In addition, it is shown that the higher order derivatives are effective tools to realize sophisticated control of nonlinear systems. Other features of ULN's such as multiple branches with arbitrary time delays and using a priori information will be discussed in other papers.

UR - http://www.scopus.com/inward/record.url?scp=0033723746&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033723746&partnerID=8YFLogxK

U2 - 10.1109/3477.846231

DO - 10.1109/3477.846231

M3 - Article

C2 - 18252374

AN - SCOPUS:0033723746

VL - 30

SP - 419

EP - 430

JO - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

JF - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

SN - 1083-4419

IS - 3

ER -