國立高雄大學圖資館 |

Language: English

Back

A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems.

Record Type:	Electronic resources : Monograph/item
Title/Author:	A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems.
Author:	Zakeri, Roja.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2020
Description:	193 p.
Notes:	Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
Notes:	Advisor: Shankar, Praveen.
Contained By:	Dissertations Abstracts International82-07B.
Subject:	Applied mathematics.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28258723
ISBN:	9798557030328

A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems.
Zakeri, Roja.

A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 193 p.

Source: Dissertations Abstracts International, Volume: 82-07, Section: B.

Thesis (Ph.D.)--The Claremont Graduate University, 2020.

This item must not be sold to any third party vendors.

The objective of this research is to develop a new methodology by combining Artificial Neural Networks and Bayesian approach which utilizes kinematic quantities of a nonlinear dynamic system to estimate uncertain and unknown parameters more accurately with reduced estimation error and using fewer iterations. Kinematics pertains to the motion of bodies in the robotic mechanism without regard to the forces or torques that cause the motion. In this study, a new methodology which is the combination of a heuristic method (Neural Network) and Bayesian Approach (Particle Markov Chain Monte Carlo) is developed to determine and estimate the unknown system parameters with high accuracy, in more efficient way with fewer iteration number. The new methodology can reduce the iteration number in Bayesian samplers’ algorithms and maintains the estimation accuracy, therefore it could make the algorithm less computationally expensive and demanding. At the accuracy level of 0.002, the result showed the average of 33.64% improvement in proposed method compared to the regular PMH sampler, At the accuracy level of 0.001, the average of improvement was 34.99% in proposed methods compared to the regular PMH sampler and finally and finally, at the accuracy level of 0.0005, the result shows the average of 32.34% improvement in proposed method compared to the regular PMH sampler.

ISBN: 9798557030328Subjects--Topical Terms:

377601
Applied mathematics.
Subjects--Index Terms:

Neural network

A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems.
LDR:02572nmm a2200361 4500 001 616390
005 20220513114324.5
008 220920s2020 ||||||||||||||||| ||eng d
020 $a 9798557030328
035 $a (MiAaPQ)AAI28258723
035 $a AAI28258723
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zakeri, Roja. $3 915558
245 1 0 $a A Neural Network-Augmented Bayesian Approach to Uncertain Parameter Estimation in Nonlinear Dynamic Systems.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 193 p.
500 $a Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
500 $a Advisor: Shankar, Praveen.
502 $a Thesis (Ph.D.)--The Claremont Graduate University, 2020.
506 $a This item must not be sold to any third party vendors.
520 $a The objective of this research is to develop a new methodology by combining Artificial Neural Networks and Bayesian approach which utilizes kinematic quantities of a nonlinear dynamic system to estimate uncertain and unknown parameters more accurately with reduced estimation error and using fewer iterations. Kinematics pertains to the motion of bodies in the robotic mechanism without regard to the forces or torques that cause the motion. In this study, a new methodology which is the combination of a heuristic method (Neural Network) and Bayesian Approach (Particle Markov Chain Monte Carlo) is developed to determine and estimate the unknown system parameters with high accuracy, in more efficient way with fewer iteration number. The new methodology can reduce the iteration number in Bayesian samplers’ algorithms and maintains the estimation accuracy, therefore it could make the algorithm less computationally expensive and demanding. At the accuracy level of 0.002, the result showed the average of 33.64% improvement in proposed method compared to the regular PMH sampler, At the accuracy level of 0.001, the average of improvement was 34.99% in proposed methods compared to the regular PMH sampler and finally and finally, at the accuracy level of 0.0005, the result shows the average of 32.34% improvement in proposed method compared to the regular PMH sampler.
590 $a School code: 0047.
650 4 $a Applied mathematics. $3 377601
650 4 $a Engineering. $3 210888
650 4 $a Robotics. $3 181952
653 $a Neural network
653 $a Bayesian approach
653 $a Uncertain parameter estimation
653 $a Nonlinear dynamic system
690 $a 0364
690 $a 0537
690 $a 0771
710 2 $a The Claremont Graduate University. $b Institute of Mathematical Sciences. $3 915559
773 0 $t Dissertations Abstracts International $g 82-07B.
790 $a 0047
791 $a Ph.D.
792 $a 2020
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28258723