On the dynamic neural network toolbox design for identification, estimation and control

Abstract

PurposeThere are common problems in the identification of uncertain nonlinear systems, nonparametric approximation, state estimation, and automatic control. Dynamic neural network (DNN) approximation can simplify the development of all the aforementioned problems in either continuous or discrete systems. A DNN is represented by a system of differential or recurrent equations defined in the space of vector activation functions with weights and offsets that are functionally associated with the input data.Design/methodology/approachThis study describes the version of the toolbox, that can be used to identify the dynamics of the black box and restore the laws underlying the system using known inputs and outputs. Depending on the completeness of the information, the toolbox allows users to change the DNN structure to suit specific tasks.FindingsThe toolbox consists of three main components: user layer, network manager, and network instance. The user layer provides high-level control and monitoring of system performance. The network manager serves as an intermediary between the user layer and the network instance, and allows the user layer to start and stop learning, providing an interface to indirectly access the internal data of the DNN.Research limitations/implicationsControl capability is limited to adjusting a small number of numerical parameters and selecting functional parameters from a predefined list.Originality/valueThe key feature of the toolbox is the possibility of developing an algorithmic semi-automatic selection of activation function parameters based on optimization problem solutions.

Guest editorial: CyberSystemic implications on the future of societies

“…Another level of standardisation is proposed by Chairez et al . (2023) in the field of AI modelling.…”

“…Another level of standardisation is proposed by Chairez et al (2023) in the field of AI modelling. The authors propose a version of the DNN toolbox that can be used to identify the dynamics of the black box and restore the laws underlying the system using known inputs and outputs, featuring the possibility of developing an algorithmic semi-automatic selection of activation function parameters based on optimisation problem solutions.…”

Learning Synchronous Control Based on Intelligent Methods for Ultra-Precision Stage Systems

“…Since MCM is based on linearization, it requires linearization when applied to nonlinear systems, which could result in significant residual error [23][24][25]. Iterative learning control (ILC) is adopted in industries to reduce dependencies on models, which is applicable for systems with repetitive movements [26][27]. For a wafer scanner, 6 iterations could reduce the tracking error to an accuracy of 100nm level through parameters update [26].…”

“…The advancement in computing and artificial intelligence enables extensive computation, thus supporting more complex strategies [19]. Numerous experimental studies have demonstrated the effectiveness of ICM in handling uncertainty and complexity [27,29]. Neural network control (NNC) is a typical example of ICM, which could adapt and compensate for non-linear errors very well.…”