Abstract

In this paper, we present a lifelong deep learning-based control of robotic manipulators with nonstandard adaptive laws using singular value decomposition (SVD) based direct tracking error driven (DTED) approach. Moreover, we incorporate concurrent learning (CL) to relax persistency of excitation condition and elastic weight consolidation (EWC) for lifelong learning on different tasks in the adaptive laws. Simulation results confirm theoretical conclusions.

Department(s)

Electrical and Computer Engineering

Comments

Office of Naval Research, Grant N00014-21-1-2232

Keywords and Phrases

concurrent learning; deep neural networks; elastic weight consolidation; lifelong learning; SVD

International Standard Serial Number (ISSN)

2405-8963

Document Type

Article - Conference proceedings

Document Version

Final Version

File Type

text

Language(s)

English

Rights

© 2023 The Authors, All rights reserved.

Creative Commons Licensing

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Publication Date

01 Jul 2022

Share

 
COinS