Epistemic Uncertainty Quantification in State-Space LPV Model Identification Using Bayesian Neural Networks

Bao, Yajie; Velni, Javad Mohammadpour; Shahbakhti, Mahdi

doi:10.1109/lcsys.2020.3005429

Cited by 22 publications

(18 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Class 1 weight, a system of protection against acoustic interference and vibrations, a voltmeter to control the voltage stability at the input of the ultrasonic emitter, and a thermometer to control the change in water temperature were introduced into the structural and functional diagram of the standard. Formulas (1) and ( 3) are approximate, they are used only for working measurements, therefore, (2) -( 6) was taken as a refined mathematical model of the measurement process (2)- (6). c) the procedure for assessing uncertainty using the improved PUMA method.…”

Section: Resultsmentioning

confidence: 99%

“…where u A -the standard uncertainty, calculated by type A; u B -standard uncertainty, which is calculated according to type B: uncorrelated (r=0) components are added geometrically, and strongly correlated (r=+1 or r=-1) components are added arithmetically: (6) where u k (y) -the contribution of the k-th input quantity to the combined standard uncertainty; m -the number of uncorrelated sources of uncertainty; u q -the sum of highly correlated uncertainty components: (7) where q -the number of strongly correlated components.…”

Section: Methodsmentioning

confidence: 99%

“…The important parameters that determine the efficiency of neural network training are the duration and error [5]. For a given training sample, in order to reduce the error and the training duration of the neural network, most of the known studies suggest using various procedures for pre-processing the training example parameters [6]. The implementation of the procedures is reduced to centring, normalisation, scaling of the input and output parameters [7].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Methods for Improving the Efficiency of Diagnostic Systems in the Neural Network-based Sound Analysis

Panaskin

Bilokon

Piddiachyi

2021

International Journal of Biology and Biomedical Engineering

View full text Add to dashboard Cite

The study of the possibilities of modelling the use of neural networks while increasing the efficiency of diagnostic systems consists of creating a standard that would satisfy the conditions for maintaining the sound quality. At the same time, the effectiveness of diagnostic systems can be considered when applied both in a technological environment and in a virtual space. The relevance of the study is determined by the possibilities of using the reference sound, which forms and uses the basis of the neural networks. The scientific novelty of the study is determined by the fact that an adaptive method for creating standards of units of measurable quantities with specified accuracy characteristics is proposed, subject to limited resources. The first version of the mathematical model of the measurement procedure is formed during reproduction, storage and transmission of a unit of measurement developed on the basis of a physical model, which, in turn, is built in accordance with a priori information on the principle of reproduction (storage and transmission) of a unit of measurement, a list of informative parameters and influential quantities when measuring. The authors have developed the necessary accuracy characteristics, specified by the technical specifications and determined the resources allocated for the creation of the standard. The practical significance of the study lies in the establishment of distributed networks for sound quality measurement, mainly within the structures of the study of sound transmission between high-tech devices.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Methods for Improving the Efficiency of Diagnostic Systems in the Neural Network-based Sound Analysis

Panaskin

Bilokon

Piddiachyi

2021

International Journal of Biology and Biomedical Engineering

View full text Add to dashboard Cite

show abstract

“…Several ML-based data-driven modeling techniques have been used, including Artificial Neural Network (ANN) [197,203], Extreme Learning Machine (EL)M [198,204], Bayesian Neural Network (BNN) [205], and Least-Square Support Vector Machine (LS-SVM) [21,22] to provide a predictive model of sufficient accuracy for control of ICEs. MPC and ML integration in ICEs applications is depicted in Figure 16.…”

Section: Ai and Mpc Integrationmentioning

confidence: 99%

“…Combustion type RCCI [21,22,112,129,130,197,205] CI [206,207] SI [24,203,206,208] HCCI [198,204] ML Modeling methods BNN [205] ANN [24,112,197,203,208] ELM [198,204] LS-SVM [21,22,129,130] Control objectives IMEP & CA50 [21,22,112,129,130,197,198,204,205] Speed [206] Airpath [208] Air-fuel ratio [24,203] EGR rate [207] MPRR [22] Model structure for control Nonlinear [130,198,203,206,208] state-space LPV [21,22,…”

Section: And Mpc Integration In Icesmentioning

confidence: 99%

Model Predictive Control of Internal Combustion Engines: A Review and Future Directions

et al. 2021

Self Cite

View full text Add to dashboard Cite

An internal combustion engine (ICE) is a highly nonlinear dynamic and complex engineering system whose operation is constrained by operational limits, including emissions, noise, peak in-cylinder pressure, combustion stability, and actuator constraints. To optimize today’s ICEs, seven to ten control actuators and 10–20 feedback sensors are often used, depending on the engine applications and target emission regulations. This requires extensive engine experimentation to calibrate the engine control module (ECM), which is both cumbersome and costly. Despite these efforts, optimal operation, particularly during engine transients and to meet real driving emission (RDE) targets for broad engine speed and load conditions, has still not been obtained. Methods of model predictive control (MPC) have shown promising results for real-time multi-objective optimal control of constrained multi-variable nonlinear systems, including ICEs. This paper reviews the application of MPC for ICEs and analyzes the recent developments in MPC that can be utilized in ECMs. ICE control and calibration can be enhanced by taking advantage of the recent developments in the field of Artificial Intelligence (AI) in applying Machine Learning (ML) to large-scale engine data. Recent developments in the field of ML-MPC are investigated, and promising methods for ICE control applications are identified in this paper.

show abstract

Learning‐based adaptive‐scenario‐tree model predictive control with improved probabilistic safety using robust Bayesian neural networks

Bao

Chan

Mesbah

et al. 2022

Intl J Robust & Nonlinear

Self Cite

View full text Add to dashboard Cite

Scenario-based model predictive control (MPC) methods can mitigate the conservativeness inherent to open-loop robust MPC. Yet, the scenarios are often generated offline based on worst-case uncertainty descriptions obtained a priori, which can in turn limit the improvements in the robust control performance.To this end, this paper presents a learning-based, adaptive-scenario-tree model predictive control approach for uncertain nonlinear systems with time-varying and/or hard-to-model dynamics. Bayesian neural networks (BNNs) are used to learn a state-and input-dependent description of model uncertainty, namely the mismatch between a nominal (physics-based or data-driven) model of a system and its actual dynamics. We first present a new approach for training robust BNNs (RBNNs) using probabilistic Lipschitz bounds to provide a less conservative uncertainty quantification. Then, we present an approach to evaluate the credible intervals of RBNN predictions and determine the number of samples required for estimating the credible intervals given a credible level. The performance of RBNNs is evaluated with respect to that of standard BNNs and Gaussian process (GP) as a basis of comparison. The RBNN description of plant-model mismatch with verified accurate credible intervals is employed to generate adaptive scenarios online for scenario-based MPC (sMPC). The proposed sMPC approach with adaptive scenario tree can improve the robust control performance with respect to sMPC with a fixed, worst-case scenario tree and with respect to an adaptive-scenario-based MPC (asMPC) using GP regression on a cold atmospheric plasma system. Furthermore, closed-loop simulation results illustrate that robust model uncertainty learning via RBNNs can enhance the probability of constraint satisfaction of asMPC.

show abstract

Epistemic Uncertainty Quantification in State-Space LPV Model Identification Using Bayesian Neural Networks

Cited by 22 publications

References 9 publications

Methods for Improving the Efficiency of Diagnostic Systems in the Neural Network-based Sound Analysis

Methods for Improving the Efficiency of Diagnostic Systems in the Neural Network-based Sound Analysis

Model Predictive Control of Internal Combustion Engines: A Review and Future Directions

Learning‐based adaptive‐scenario‐tree model predictive control with improved probabilistic safety using robust Bayesian neural networks

Contact Info

Product

Resources

About