A Gray-Box Neural Network-Based Model Identification and Fault Estimation Scheme for Nonlinear Dynamic Systems

Abstract

A novel gray-box neural network model (GBNNM), including multi-layer perception (MLP) neural network (NN) and integrators, is proposed for a model identification and fault estimation (MIFE) scheme. With the GBNNM, both the nonlinearity and dynamics of a class of nonlinear dynamic systems can be approximated. Unlike previous NN-based model identification methods, the GBNNM directly inherits system dynamics and separately models system nonlinearities. This model corresponds well with the object system and is easy to build. The GBNNM is embedded online as a normal model reference to obtain the quantitative residual between the object system output and the GBNNM output. This residual can accurately indicate the fault offset value, so it is suitable for differing fault severities. To further estimate the fault parameters (FPs), an improved extended state observer (ESO) using the same NNs (IESONN) from the GBNNM is proposed to avoid requiring the knowledge of ESO nonlinearity. Then, the proposed MIFE scheme is applied for reaction wheels (RW) in a satellite attitude control system (SACS). The scheme using the GBNNM is compared with other NNs in the same fault scenario, and several partial loss of effect (LOE) faults with different severities are considered to validate the effectiveness of the FP estimation and its superiority.

Clinical Neurophysiological and Automated EEG-Based Diagnosis of the Alzheimer's Disease

“…Earlier, Smith [58] reported that EEG signals are normal during the early stage of AD, but as the disease progresses, alpha waves disappear and slow waves become more apparent. In some cases, periodic sharp Bhat [48,64,65] have advanced the idea that adroit integration of three computing paradigms, time-frequency signal processing [66][67][68] , chaos theory and nonlinear methods [69,70] , and pattern recognition techniques such as artificial neural networks [71][72][73] is the best approach to analyze nonstationary and highly chaotic signals. Significant features can be extracted by nonlinear dynamics and classified using different data-mining techniques and neural networks [74] .…”

Neuronal Entropy-Rate Feature of Entopeduncular Nucleus in Rat Model of Parkinson’s Disease

“…The rate-irregularity relationship may be a relevant feature to identify, localize and quantify pathological electrophysiological activities of the BG in the context of movement disorders. Finally, integration of the entropy-rate feature into computational model 56, 57 and/or brain interface system 58 may help to detect misprocessing episodes of sensory motor information.…”

Computer-aided diagnosis of alcoholism-related EEG signals

“…This also applies to alcoholism and its impact on the human brain [39,91,92]. Nonlinear methods such as entropies, largest Lyapunov exponent (LLE) [93,94], fractal dimension (FD) [95], and Hurst exponent (H) help in the detection of minute variations in the EEG signals [45,46,96].…”