Approximation and Inverse Control of Nonlinear System using Standard Continuous Piecewise Linear Neural Networks

摘要: The main difficulty for standard continuous piecewise linear neural networks (SCPLNN) approximation is how to partition the definitional domain into several simplices, which called a triangulation. In this paper, we firstly propose method of triangulation perform SCPLNN approximation. Our scheme starts with an initial, coarse given data and subdivides simplex until error smaller than some tolerance. Then based on identified. proposed involving identification shown be useful in approximating nonlinear systems. addition, each simplex, local inverse model can easily calculated linear. From control perspective, exploit advantage property design controllers approximate model. validity using tested by NARX