Cargando…
Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning
Visual navigation based on deep reinforcement learning requires a large amount of interaction with the environment, and due to the reward sparsity, it requires a large amount of training time and computational resources. In this paper, we focus on sample efficiency and navigation performance and pro...
Autores principales: | , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10378290/ https://www.ncbi.nlm.nih.gov/pubmed/37509957 http://dx.doi.org/10.3390/e25071007 |
_version_ | 1785079729066344448 |
---|---|
author | Zhang, Wenzhi He, Li Wang, Hongwei Yuan, Liang Xiao, Wendong |
author_facet | Zhang, Wenzhi He, Li Wang, Hongwei Yuan, Liang Xiao, Wendong |
author_sort | Zhang, Wenzhi |
collection | PubMed |
description | Visual navigation based on deep reinforcement learning requires a large amount of interaction with the environment, and due to the reward sparsity, it requires a large amount of training time and computational resources. In this paper, we focus on sample efficiency and navigation performance and propose a framework for visual navigation based on multiple self-supervised auxiliary tasks. Specifically, we present an LSTM-based dynamics model and an attention-based image-reconstruction model as auxiliary tasks. These self-supervised auxiliary tasks enable agents to learn navigation strategies directly from the original high-dimensional images without relying on ResNet features by constructing latent representation learning. Experimental results show that without manually designed features and prior demonstrations, our method significantly improves the training efficiency and outperforms the baseline algorithms on the simulator and real-world image datasets. |
format | Online Article Text |
id | pubmed-10378290 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-103782902023-07-29 Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning Zhang, Wenzhi He, Li Wang, Hongwei Yuan, Liang Xiao, Wendong Entropy (Basel) Article Visual navigation based on deep reinforcement learning requires a large amount of interaction with the environment, and due to the reward sparsity, it requires a large amount of training time and computational resources. In this paper, we focus on sample efficiency and navigation performance and propose a framework for visual navigation based on multiple self-supervised auxiliary tasks. Specifically, we present an LSTM-based dynamics model and an attention-based image-reconstruction model as auxiliary tasks. These self-supervised auxiliary tasks enable agents to learn navigation strategies directly from the original high-dimensional images without relying on ResNet features by constructing latent representation learning. Experimental results show that without manually designed features and prior demonstrations, our method significantly improves the training efficiency and outperforms the baseline algorithms on the simulator and real-world image datasets. MDPI 2023-06-30 /pmc/articles/PMC10378290/ /pubmed/37509957 http://dx.doi.org/10.3390/e25071007 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). |
spellingShingle | Article Zhang, Wenzhi He, Li Wang, Hongwei Yuan, Liang Xiao, Wendong Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title | Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title_full | Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title_fullStr | Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title_full_unstemmed | Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title_short | Multiple Self-Supervised Auxiliary Tasks for Target-Driven Visual Navigation Using Deep Reinforcement Learning |
title_sort | multiple self-supervised auxiliary tasks for target-driven visual navigation using deep reinforcement learning |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10378290/ https://www.ncbi.nlm.nih.gov/pubmed/37509957 http://dx.doi.org/10.3390/e25071007 |
work_keys_str_mv | AT zhangwenzhi multipleselfsupervisedauxiliarytasksfortargetdrivenvisualnavigationusingdeepreinforcementlearning AT heli multipleselfsupervisedauxiliarytasksfortargetdrivenvisualnavigationusingdeepreinforcementlearning AT wanghongwei multipleselfsupervisedauxiliarytasksfortargetdrivenvisualnavigationusingdeepreinforcementlearning AT yuanliang multipleselfsupervisedauxiliarytasksfortargetdrivenvisualnavigationusingdeepreinforcementlearning AT xiaowendong multipleselfsupervisedauxiliarytasksfortargetdrivenvisualnavigationusingdeepreinforcementlearning |