The K-filter: A New Model Of Non-linear Systems With Memory

Abstract

This paper deals with the problem of modelling and identifying non-linear systems (NLSs). Although the final architecture applies to NLSs with memory, it is also important to bear in mind that the proposed filter consists in a generalisation of a previous scheme which was designed to model memoryless N U S . In consequence, the simplicity of the memoryless filter is useful to intuitively understand the final architecture and how it can be improved when dealing with specific problems. On the other hand, the filter with memory shows the complexity and computational load that non-linear theory usually involves. But at the same time. it allows a large number of different possibilities depending on the concrete application and it also opens new ways of future work.