Explainable and Safety Aware Deep Reinforcement Learning-Based Control of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition

Behzad Farzanegan; Sarangapani Jagannathan

doi:10.1109/TASE.2025.3554431

Explainable and Safety Aware Deep Reinforcement Learning-Based Control of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition

Behzad Farzanegan
, Sarangapani Jagannathan

Missouri University of Science and Technology

Research output: Contribution to journal › Article › peer-review

Original language	English
Pages (from-to)	13556-13568
Number of pages	13
Journal	IEEE Transactions on Automation Science and Engineering
Volume	22
DOIs	https://doi.org/10.1109/TASE.2025.3554431
State	Published - 2025

ASJC Scopus Subject Areas

Control and Systems Engineering
Electrical and Electronic Engineering

Keywords

Deep reinforcement learning
approximate dynamic programming
explainability
online safety-aware control
singular value decomposition-based weight tuning

Access to Document

10.1109/TASE.2025.3554431

Cite this

@article{be2b42b3fbfb4a9b829043e59beff6f1,

title = "Explainable and Safety Aware Deep Reinforcement Learning-Based Control of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition",

keywords = "Deep reinforcement learning, approximate dynamic programming, explainability, online safety-aware control, singular value decomposition-based weight tuning",

author = "Behzad Farzanegan and Sarangapani Jagannathan",

note = "Publisher Copyright: {\textcopyright} 2004-2012 IEEE.",

year = "2025",

doi = "10.1109/TASE.2025.3554431",

language = "English",

volume = "22",

pages = "13556--13568",

journal = "IEEE Transactions on Automation Science and Engineering",

issn = "1545-5955",

}

TY - JOUR

T1 - Explainable and Safety Aware Deep Reinforcement Learning-Based Control of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition

AU - Farzanegan, Behzad

AU - Jagannathan, Sarangapani

PY - 2025

Y1 - 2025

KW - Deep reinforcement learning

KW - approximate dynamic programming

KW - explainability

KW - online safety-aware control

KW - singular value decomposition-based weight tuning

UR - https://www.scopus.com/pages/publications/105003088666

UR - https://www.scopus.com/pages/publications/105003088666#tab=citedBy

U2 - 10.1109/TASE.2025.3554431

DO - 10.1109/TASE.2025.3554431

M3 - Article

AN - SCOPUS:105003088666

SN - 1545-5955

VL - 22

SP - 13556

EP - 13568

JO - IEEE Transactions on Automation Science and Engineering

JF - IEEE Transactions on Automation Science and Engineering

ER -

Explainable and Safety Aware Deep Reinforcement Learning-Based Control of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition

ASJC Scopus Subject Areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this