Robust motion planning for multi-robot systems against position deception attacks
Deep reinforcement learning (DRL) is widely applied in motion planning for multi-robot systems as DRL leverages the offline training process to improve the real-time computation efficiency. In DRL-based methods, the DRL models compute an action for a robot based on the states of its surrounding obst...
Saved in:
Main Authors: | , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2024
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/176231 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-176231 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1762312024-05-14T04:17:43Z Robust motion planning for multi-robot systems against position deception attacks Tang, Wenbing Zhou, Yuan Liu, Yang Ding, Zuohua Liu, Jing School of Computer Science and Engineering Computer and Information Science Deep reinforcement learning Denoising autoencoder Deep reinforcement learning (DRL) is widely applied in motion planning for multi-robot systems as DRL leverages the offline training process to improve the real-time computation efficiency. In DRL-based methods, the DRL models compute an action for a robot based on the states of its surrounding obstacles, including other robots in the system. They always assume that the number of obstacles is fixed and the obtained obstacles' states are reliable. However, in the real world, a multi-robot system may suffer from various attacks, such as remote control attacks and network attacks, that cause wrong positions of the surrounding obstacles received by a robot. In this paper, we propose a robust motion planning method DAE-Crit-LSTM, integrating a denoising autoencoder (DAE) with DRL models, to mitigate such position deception attacks in environments with a different number of obstacles. DAE-Crit-LSTM shows the following two advantages. First, DAE-Crit-LSTM can be applied in benign and attacked scenarios and thus does not require any detector. It learns an encoder and a decoder to approximate the accurate positions of the obstacles, no matter under attack or not. Second, DAE-Crit-LSTM applies an LSTM (Long Short-Term Memory)-based DRL model to deal with a variable number of obstacles in the environment. It is worth noting that DAE-Crit-LSTM is method-agnostic and can be easily implemented in state-of-the-art motion planning methods. Comprehensive experiments show that DAE-Crit-LSTM can mitigate position deception attacks and guarantee safe motion. We also demonstrate the effectiveness and generalization of DAE-Crit-LSTM. AI Singapore Ministry of Education (MOE) National Research Foundation (NRF) This work was supported in part by the National Key Research and Development Program of China under Grant 2022YFC3302600; in part by the National Natural Science Foundation of China under Grant 61972150 and Grant 62132014; in part by the Fundamental Research Funds for Central Universities of China; in part by the Zhejiang Provincial Key Research and Development Program of China under Grant 2022C01045; in part by the Academic Research Fund Tier 2 by the Ministry of Education in Singapore under Grant MOE-T2EP20120-0004; in part by the National Research Foundation, Singapore; in part by Defence Science Organisation (DSO) National Laboratories under the AI Singapore Program (AISG) under Award AISG2-GC-2023-008; and in part by the National Research Foundation (NRF) Investigatorship under Grant NRF-NRFI06-2020-0001. 2024-05-14T04:17:43Z 2024-05-14T04:17:43Z 2024 Journal Article Tang, W., Zhou, Y., Liu, Y., Ding, Z. & Liu, J. (2024). Robust motion planning for multi-robot systems against position deception attacks. IEEE Transactions On Information Forensics and Security, 19, 2157-2170. https://dx.doi.org/10.1109/TIFS.2023.3346647 1556-6013 https://hdl.handle.net/10356/176231 10.1109/TIFS.2023.3346647 2-s2.0-85181563211 19 2157 2170 en MOE-T2EP20120-0004 AISG2-GC-2023-008 NRF-NRFI06-2020-0001 IEEE Transactions on Information Forensics and Security © 2023 IEEE. All rights reserved. |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Computer and Information Science Deep reinforcement learning Denoising autoencoder |
spellingShingle |
Computer and Information Science Deep reinforcement learning Denoising autoencoder Tang, Wenbing Zhou, Yuan Liu, Yang Ding, Zuohua Liu, Jing Robust motion planning for multi-robot systems against position deception attacks |
description |
Deep reinforcement learning (DRL) is widely applied in motion planning for multi-robot systems as DRL leverages the offline training process to improve the real-time computation efficiency. In DRL-based methods, the DRL models compute an action for a robot based on the states of its surrounding obstacles, including other robots in the system. They always assume that the number of obstacles is fixed and the obtained obstacles' states are reliable. However, in the real world, a multi-robot system may suffer from various attacks, such as remote control attacks and network attacks, that cause wrong positions of the surrounding obstacles received by a robot. In this paper, we propose a robust motion planning method DAE-Crit-LSTM, integrating a denoising autoencoder (DAE) with DRL models, to mitigate such position deception attacks in environments with a different number of obstacles. DAE-Crit-LSTM shows the following two advantages. First, DAE-Crit-LSTM can be applied in benign and attacked scenarios and thus does not require any detector. It learns an encoder and a decoder to approximate the accurate positions of the obstacles, no matter under attack or not. Second, DAE-Crit-LSTM applies an LSTM (Long Short-Term Memory)-based DRL model to deal with a variable number of obstacles in the environment. It is worth noting that DAE-Crit-LSTM is method-agnostic and can be easily implemented in state-of-the-art motion planning methods. Comprehensive experiments show that DAE-Crit-LSTM can mitigate position deception attacks and guarantee safe motion. We also demonstrate the effectiveness and generalization of DAE-Crit-LSTM. |
author2 |
School of Computer Science and Engineering |
author_facet |
School of Computer Science and Engineering Tang, Wenbing Zhou, Yuan Liu, Yang Ding, Zuohua Liu, Jing |
format |
Article |
author |
Tang, Wenbing Zhou, Yuan Liu, Yang Ding, Zuohua Liu, Jing |
author_sort |
Tang, Wenbing |
title |
Robust motion planning for multi-robot systems against position deception attacks |
title_short |
Robust motion planning for multi-robot systems against position deception attacks |
title_full |
Robust motion planning for multi-robot systems against position deception attacks |
title_fullStr |
Robust motion planning for multi-robot systems against position deception attacks |
title_full_unstemmed |
Robust motion planning for multi-robot systems against position deception attacks |
title_sort |
robust motion planning for multi-robot systems against position deception attacks |
publishDate |
2024 |
url |
https://hdl.handle.net/10356/176231 |
_version_ |
1800916190125621248 |