Yaoxin Duan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yaoxin Duan has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Wendi Nie
Key Laboratory of Industrial Internet of Things and Networked Control, Ministry of Education; the School of Automation/School of Industrial Internet, Chongqing University of Posts and Telecommunications, Chongqing, China
11 shared publications - Kai Liu
Chongqing University
10 shared publications - Victor Chung Sing Lee
The University of Hong Kong
6 shared publications - Chun Jason Xue
Mohamed bin Zayed University of Artificial Intelligence
4 shared publications - Kam-Yiu Lam
Computer Science Department, City University of Hong Kong, Hong Kong
4 shared publications - Desheng Zheng
School of Computer and Software, Southwest Petroleum University, Chengdu, China and Intelligent Manufacturing Institute, Kash Institute of Electronics and Information Industry, Kash, China
3 shared publications
Latest Publications
- 29 October 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 334
Improving adversarial transferability with pre-deep search of initial momentum
- 29 October 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 334
Improving adversarial transferability with pre-deep search of initial momentum
- 24 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Towards Pervasive WLAN Localization Leveraging Collaborative Mobile Sites: A Multi-Agent Deep Reinforcement Learning Approach
Indoor location-based services (LBS) have witnessed rapid growth in applications such as user tracking, healthcare monitoring, and smart facility management, driving the critical need for efficient and pervasive indoor localization. Traditional WiFi fingerprinting methods face significant challenges: multi-site localization (MSL) relies on densely deployed static WiFi sites, incurring high infrastructure costs and conflicting with the Integrated Sensing and Communication (ISAC) paradigm; single-site localization (SSL) requires complex hardware; and single mobile site localization (SMSL) suffers from poor real-time performance due to long traversal paths. To address these limitations, this paper proposes a Multi-Agent Deep Reinforcement Learning-based Collaborative Indoor Localization (MADRL-CIL) framework. MADRL-CIL leverages multiple collaborative mobile sites to dynamically acquire Received Signal Strength (RSS) fingerprints. By modeling each mobile site as an agent, the framework formulates the path selection and fingerprint acquisition task as a Multi-Agent Deep Reinforcement Learning (MADRL) problem under a Centralized Training with Decentralized Execution (CTDE) paradigm, facilitating effective collaboration among multiple mobile sites to optimize localization accuracy while minimizing localization time. Additionally, a Multi-Site Fingerprint Matching (MS-FM) model is specifically designed to process collaboratively collected RSS fingerprints, enabling fine-grained localization accuracy. Experimental evaluations in a real-world indoor environment demonstrate that MADRL-CIL achieves localization accuracy comparable to dense multi-static site deployments while provides good real-time performance.
- 24 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Towards Pervasive WLAN Localization Leveraging Collaborative Mobile Sites: A Multi-Agent Deep Reinforcement Learning Approach
Indoor location-based services (LBS) have witnessed rapid growth in applications such as user tracking, healthcare monitoring, and smart facility management, driving the critical need for efficient and pervasive indoor localization. Traditional WiFi fingerprinting methods face significant challenges: multi-site localization (MSL) relies on densely deployed static WiFi sites, incurring high infrastructure costs and conflicting with the Integrated Sensing and Communication (ISAC) paradigm; single-site localization (SSL) requires complex hardware; and single mobile site localization (SMSL) suffers from poor real-time performance due to long traversal paths. To address these limitations, this paper proposes a Multi-Agent Deep Reinforcement Learning-based Collaborative Indoor Localization (MADRL-CIL) framework. MADRL-CIL leverages multiple collaborative mobile sites to dynamically acquire Received Signal Strength (RSS) fingerprints. By modeling each mobile site as an agent, the framework formulates the path selection and fingerprint acquisition task as a Multi-Agent Deep Reinforcement Learning (MADRL) problem under a Centralized Training with Decentralized Execution (CTDE) paradigm, facilitating effective collaboration among multiple mobile sites to optimize localization accuracy while minimizing localization time. Additionally, a Multi-Site Fingerprint Matching (MS-FM) model is specifically designed to process collaboratively collected RSS fingerprints, enabling fine-grained localization accuracy. Experimental evaluations in a real-world indoor environment demonstrate that MADRL-CIL achieves localization accuracy comparable to dense multi-static site deployments while provides good real-time performance.
- Open Access
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Signal Processing
- Vol. 2026 (1)
BP Neural Network–Based Kalman Filtering Method Under Multiple Cyberattacks
This paper proposes a Kalman‐gain‐driven neural Kalman filtering (KF) defense framework, termed KFDBP, for secure state estimation in cyber–physical systems (CPSs) under denial‐of‐service (DoS), spoofing, and replay attacks. Unlike end‐to‐end neural filtering approaches such as KalmanNet that directly learn state estimators or implicitly approximate the Kalman gain using deep recurrent architectures, the proposed method employs a lightweight back‐propagation (BP) neural network to adaptively regulate the Kalman gain online, while strictly preserving the classical Kalman filter prediction–correction recursion. By formulating an innovation‐oriented Kalman gain learning objective, KFDBP explicitly addresses attack‐induced observation uncertainty and non‐Gaussian measurement corruption without requiring prior knowledge of attack timing, attack type, or attack probability during online estimation. The bounded gain regulation mechanism enhances estimation stability and interpretability, which are critical for safety‐sensitive CPS applications, while significantly reducing computational complexity compared with deep neural network–based filters. Extensive Monte Carlo simulations under single and hybrid attack scenarios demonstrate that KFDBP consistently achieves lower estimation error and improved robustness than the conventional Kalman filter and KalmanNet under different attack probabilities, making it suitable for real‐time and resource‐constrained CPS applications.
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Signal Processing
- Vol. 2026 (1)
Open AccessBP Neural Network–Based Kalman Filtering Method Under Multiple Cyberattacks
This paper proposes a Kalman‐gain‐driven neural Kalman filtering (KF) defense framework, termed KFDBP, for secure state estimation in cyber–physical systems (CPSs) under denial‐of‐service (DoS), spoofing, and replay attacks. Unlike end‐to‐end neural filtering approaches such as KalmanNet that directly learn state estimators or implicitly approximate the Kalman gain using deep recurrent architectures, the proposed method employs a lightweight back‐propagation (BP) neural network to adaptively regulate the Kalman gain online, while strictly preserving the classical Kalman filter prediction–correction recursion. By formulating an innovation‐oriented Kalman gain learning objective, KFDBP explicitly addresses attack‐induced observation uncertainty and non‐Gaussian measurement corruption without requiring prior knowledge of attack timing, attack type, or attack probability during online estimation. The bounded gain regulation mechanism enhances estimation stability and interpretability, which are critical for safety‐sensitive CPS applications, while significantly reducing computational complexity compared with deep neural network–based filters. Extensive Monte Carlo simulations under single and hybrid attack scenarios demonstrate that KFDBP consistently achieves lower estimation error and improved robustness than the conventional Kalman filter and KalmanNet under different attack probabilities, making it suitable for real‐time and resource‐constrained CPS applications.
- 6 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 188
Enhancing adversarial transferability by resolving gradient conflicts in multi-input transformations
- 6 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 188
Enhancing adversarial transferability by resolving gradient conflicts in multi-input transformations
- 1 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (18) , 37912-37927
A Multiagent DRL-Based Method for Cooperatively Determining Coordination and Lane Change of Vehicles at Signal-Free Intersections With Free-Direction Lanes
Owing to the growing population and rapid urbanization, intersections, where traffic converges from various directions, have become major bottlenecks for road capacity due to frequent congestion. Recent advances in connected and autonomous vehicle (CAV) technology enable signal-free intersections, where CAVs collaborate to cross intersections without collisions. Most existing signal-free intersection control methods focus on accommodating conflicts among vehicles inside the intersection and fixed-direction lanes are commonly adopted. However, the use of fixed-direction lanes is a legacy from conventional signalized intersections, where turning lanes are predetermined and fixed, so as to direct vehicles with different turning intentions to different lanes and avoid collisions. In this article, we aim to make full utilization of the capacity of signal-free intersections by making use of free-direction lanes, which allow vehicles to make right, straight or left turns from any lane. To this end, we propose a cooperative multiagent deep reinforcement learning (DRL)-based control method for signal-free intersections with free-direction lanes. Specifically, we first study the problem of cooperatively determining coordination of vehicles inside the intersection and lane changes of vehicles on the incoming arms. Then, a multiagent DRL-based control method for cooperatively determining coordination and lane change (LC) of vehicles for signal-free intersections with free-direction lanes, named CD-CLC, is proposed for maximizing nonconflicting vehicles crossing the intersection simultaneously while taking vehicle fairness into consideration, to minimize travel delays of vehicles and improve traffic efficiency. Extensive experiments have been conducted to compare CD-CLC with other state-of-the-art methods to demonstrate the effectiveness of the proposed approach.
- 1 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (18) , 37912-37927
A Multiagent DRL-Based Method for Cooperatively Determining Coordination and Lane Change of Vehicles at Signal-Free Intersections With Free-Direction Lanes
Owing to the growing population and rapid urbanization, intersections, where traffic converges from various directions, have become major bottlenecks for road capacity due to frequent congestion. Recent advances in connected and autonomous vehicle (CAV) technology enable signal-free intersections, where CAVs collaborate to cross intersections without collisions. Most existing signal-free intersection control methods focus on accommodating conflicts among vehicles inside the intersection and fixed-direction lanes are commonly adopted. However, the use of fixed-direction lanes is a legacy from conventional signalized intersections, where turning lanes are predetermined and fixed, so as to direct vehicles with different turning intentions to different lanes and avoid collisions. In this article, we aim to make full utilization of the capacity of signal-free intersections by making use of free-direction lanes, which allow vehicles to make right, straight or left turns from any lane. To this end, we propose a cooperative multiagent deep reinforcement learning (DRL)-based control method for signal-free intersections with free-direction lanes. Specifically, we first study the problem of cooperatively determining coordination of vehicles inside the intersection and lane changes of vehicles on the incoming arms. Then, a multiagent DRL-based control method for cooperatively determining coordination and lane change (LC) of vehicles for signal-free intersections with free-direction lanes, named CD-CLC, is proposed for maximizing nonconflicting vehicles crossing the intersection simultaneously while taking vehicle fairness into consideration, to minimize travel delays of vehicles and improve traffic efficiency. Extensive experiments have been conducted to compare CD-CLC with other state-of-the-art methods to demonstrate the effectiveness of the proposed approach.
- 13 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (11) , 15224-15237
Pervasive Indoor User Identification Leveraging Mobile Single-Station Localization
The utilization of Wi-Fi-based technology for pervasive indoor user identification has gained prominence due to its cost-effective nature and compatibility with user devices. Previous works proposed capturing the media access control (MAC) address emitted from a user’s device and using Information Elements (IE)-based MAC de-randomization methods to mitigate the impairment caused by random MAC. However, IE types of different Wi-Fi devices are not consistently differentiated, leading to identification errors in IE-based methods. Additionally, typical Wi-Fi fingerprinting approaches require densely pre-deployed Wi-Fi stations, contradicting the principle of pervasive localization. To address these challenges, we propose the mobile single station-based user identification (MS.Id) technique, which leverages Wi-Fi mobile single stations for pervasive indoor user identification. MS.Id includes mobile single station localization (MSL) and MAC de-randomization based on users’ spatio-temporal location and IE information (DR.LIE). MSL can be implemented on a standard mobile Wi-Fi station without extensive pre-deployment. DR.LIE performs MAC de-randomization using the LIC algorithm to identify users with random MAC addresses. Experimental results demonstrate that MS.Id outperforms previous IE-based user identification methods and multi-station localization techniques. MSL achieves a localization error of 1.15 meters which is better than multi-station with 12 APs of 1.40 meters. DR.LIE demonstrates an identification accuracy of 95.24% which is better than AIMAC of 85.48%.
- 13 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (11) , 15224-15237
Pervasive Indoor User Identification Leveraging Mobile Single-Station Localization
The utilization of Wi-Fi-based technology for pervasive indoor user identification has gained prominence due to its cost-effective nature and compatibility with user devices. Previous works proposed capturing the media access control (MAC) address emitted from a user’s device and using Information Elements (IE)-based MAC de-randomization methods to mitigate the impairment caused by random MAC. However, IE types of different Wi-Fi devices are not consistently differentiated, leading to identification errors in IE-based methods. Additionally, typical Wi-Fi fingerprinting approaches require densely pre-deployed Wi-Fi stations, contradicting the principle of pervasive localization. To address these challenges, we propose the mobile single station-based user identification (MS.Id) technique, which leverages Wi-Fi mobile single stations for pervasive indoor user identification. MS.Id includes mobile single station localization (MSL) and MAC de-randomization based on users’ spatio-temporal location and IE information (DR.LIE). MSL can be implemented on a standard mobile Wi-Fi station without extensive pre-deployment. DR.LIE performs MAC de-randomization using the LIC algorithm to identify users with random MAC addresses. Experimental results demonstrate that MS.Id outperforms previous IE-based user identification methods and multi-station localization techniques. MSL achieves a localization error of 1.15 meters which is better than multi-station with 12 APs of 1.40 meters. DR.LIE demonstrates an identification accuracy of 95.24% which is better than AIMAC of 85.48%.
- 8 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 1462-1474
Enhancing the Transferability of Adversarial Attacks via Multi-Feature Attention
Adversarial examples have posed a serious threat to deep neural networks due to their transferability. Existing transfer-based attacks tend to improve the transferability of adversarial examples by destroying intrinsic features. However, prior work typically employed single-dimensional or additive importance estimates, which provide inaccurate representations of features. In this work, we propose the Multi-Feature Attention Attack (MFAA), which fuses multiple layers of feature representations to disrupt category-related features and thus improve the transferability of the adversarial examples. First, MFAA introduces a layer-aggregation gradient (LAG) to obtain guidance maps, which reflect the importance of features in multiple scales. Second, it generates ensemble attention (EA), preserving object-specific features and offsetting model-specific features based on the guidance maps. Third, EA is iteratively disturbed to achieve high transferability of the adversarial examples. Empirical evaluation on the standard ImageNet dataset shows that adversarial examples crafted by MFAA can effectively attack different networks. Compared to the state-of-the-art transferable attacks, our attack improves the average attack success rate of the black-box model with defense from 88.5% to 94.1% on single-model attacks and from 86.6% to 95.1% on ensemble attacks. Our code is available at Github: https://github.com/KWPCCC/MFAA.
- 8 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 1462-1474
Enhancing the Transferability of Adversarial Attacks via Multi-Feature Attention
Adversarial examples have posed a serious threat to deep neural networks due to their transferability. Existing transfer-based attacks tend to improve the transferability of adversarial examples by destroying intrinsic features. However, prior work typically employed single-dimensional or additive importance estimates, which provide inaccurate representations of features. In this work, we propose the Multi-Feature Attention Attack (MFAA), which fuses multiple layers of feature representations to disrupt category-related features and thus improve the transferability of the adversarial examples. First, MFAA introduces a layer-aggregation gradient (LAG) to obtain guidance maps, which reflect the importance of features in multiple scales. Second, it generates ensemble attention (EA), preserving object-specific features and offsetting model-specific features based on the guidance maps. Third, EA is iteratively disturbed to achieve high transferability of the adversarial examples. Empirical evaluation on the standard ImageNet dataset shows that adversarial examples crafted by MFAA can effectively attack different networks. Compared to the state-of-the-art transferable attacks, our attack improves the average attack success rate of the black-box model with defense from 88.5% to 94.1% on single-model attacks and from 86.6% to 95.1% on ensemble attacks. Our code is available at Github: https://github.com/KWPCCC/MFAA.
- 6 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (10) , 14188-14201
MS-Loc: Toward Pervasive Indoor Localization Utilizing Mobile Single Site
Leveraging the widespread deployment of existing WiFi sites, WiFi-based techniques offer substantial potential for achieving pervasive indoor localization among various indoor localization techniques. Conventional WiFi-based indoor localization techniques primarily focus on providing fine-grained accuracy. However, previous techniques are not pervasive due to the following constraints: 1) they can hardly be implemented in environments with limited resources of WiFi sites; and 2) they are constrained by high hardware requirements, such as the need for multiple antennas. In this paper, we propose a novel technique called Mobile Single-site Localization (MS-Loc), which leverages a mobile single-site to perform indoor localization. Specifically, MS-Loc utilizes existing hardware at off-the-shelf mobile WiFi sites to achieve pervasive localization rather than relying on multiple sites or multiple antennas. Moreover, in MS-Loc, a tailor-designed path planning algorithm guides the movement of the mobile single-site to locate targets quickly and accurately. We conducted extensive experiments using a real-world testbed. The experimental results demonstrate that MS-Loc presents a competitive localization accuracy compared to previous techniques but is pervasive.
- 6 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (10) , 14188-14201
MS-Loc: Toward Pervasive Indoor Localization Utilizing Mobile Single Site
Leveraging the widespread deployment of existing WiFi sites, WiFi-based techniques offer substantial potential for achieving pervasive indoor localization among various indoor localization techniques. Conventional WiFi-based indoor localization techniques primarily focus on providing fine-grained accuracy. However, previous techniques are not pervasive due to the following constraints: 1) they can hardly be implemented in environments with limited resources of WiFi sites; and 2) they are constrained by high hardware requirements, such as the need for multiple antennas. In this paper, we propose a novel technique called Mobile Single-site Localization (MS-Loc), which leverages a mobile single-site to perform indoor localization. Specifically, MS-Loc utilizes existing hardware at off-the-shelf mobile WiFi sites to achieve pervasive localization rather than relying on multiple sites or multiple antennas. Moreover, in MS-Loc, a tailor-designed path planning algorithm guides the movement of the mobile single-site to locate targets quickly and accurately. We conducted extensive experiments using a real-world testbed. The experimental results demonstrate that MS-Loc presents a competitive localization accuracy compared to previous techniques but is pervasive.
- 5 December 2024
- Published by Elsevier in Vehicular Communications
- Vol. 51
Redundant task offloading with dual-reliability in MEC-assisted vehicular networks
- 5 December 2024
- Published by Elsevier in Vehicular Communications
- Vol. 51
Redundant task offloading with dual-reliability in MEC-assisted vehicular networks
- 1 January 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 7 (4) , 2877-2889
Packet Delivery Ratio Fingerprinting: Toward Device-Invariant Passive Indoor Localization
Passive indoor localization for mobile Wi-Fi devices, e.g., smartphones, has attracted increasing attention from research communities recently. Existing passive localization techniques leverage received signal strength (RSS) of packets transmitted by target Wi-Fi devices and do not require a dedicated software installed on the devices. However, RSS-based passive localization techniques 1) are device-dependent, which results in poor localization accuracy for a wide variety of mobile devices; and 2) cannot perform real-time passive localization. In this paper, we present a novel passive localization technique, namely packet delivery ratio (PDR) fingerprinting, to address these problems. In PDR fingerprinting, a lowest-power and highest-modulation scheme (LPHMS) is proposed to generate device-invariant packet delivery ratio, which replaces RSS to construct fingerprints, to achieve device-invariant localization accuracy. Moreover, instead of passively monitoring packets rarely sent by mobile devices, in PDR fingerprinting, access points (APs) actively transmit request-to-send (RTS) frames to trigger target devices to reply clear-to-send (CTS) frames to calculate packet delivery ratio. The RTS/CTS mechanism enables PDR fingerprinting to perform real-time localization. We have conducted extensive experiments in a real-world testbed. The experimental results demonstrate that PDR fingerprinting presents a competitive localization accuracy compared to RSS-based passive fingerprinting methods but is device-invariant.
- 1 January 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 7 (4) , 2877-2889
Packet Delivery Ratio Fingerprinting: Toward Device-Invariant Passive Indoor Localization
Passive indoor localization for mobile Wi-Fi devices, e.g., smartphones, has attracted increasing attention from research communities recently. Existing passive localization techniques leverage received signal strength (RSS) of packets transmitted by target Wi-Fi devices and do not require a dedicated software installed on the devices. However, RSS-based passive localization techniques 1) are device-dependent, which results in poor localization accuracy for a wide variety of mobile devices; and 2) cannot perform real-time passive localization. In this paper, we present a novel passive localization technique, namely packet delivery ratio (PDR) fingerprinting, to address these problems. In PDR fingerprinting, a lowest-power and highest-modulation scheme (LPHMS) is proposed to generate device-invariant packet delivery ratio, which replaces RSS to construct fingerprints, to achieve device-invariant localization accuracy. Moreover, instead of passively monitoring packets rarely sent by mobile devices, in PDR fingerprinting, access points (APs) actively transmit request-to-send (RTS) frames to trigger target devices to reply clear-to-send (CTS) frames to calculate packet delivery ratio. The RTS/CTS mechanism enables PDR fingerprinting to perform real-time localization. We have conducted extensive experiments in a real-world testbed. The experimental results demonstrate that PDR fingerprinting presents a competitive localization accuracy compared to RSS-based passive fingerprinting methods but is device-invariant.