Le Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Le Zhang has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Ye Du
Beijing Jiaotong University
8 shared publications - Junyan Guo
China Telecom Cybersecurity Technology Company Ltd., Beijing, China
4 shared publications Jikai Zhang
2 shared publications- Liyuan Chang
Beijing University of Posts and Telecommunications, Beijing, China
2 shared publications - Shuang Yao
Institute of Information and Communication, China Electric Power Research Institute Company Ltd., Beijing, China
2 shared publications - Yu Gu
Beijing Jiaotong University
2 shared publications
Latest Publications
- 25 April 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 275
A modeling and simulation framework for the study and analysis of cascading failures in satellite networks
- 25 April 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 275
A modeling and simulation framework for the study and analysis of cascading failures in satellite networks
- 20 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 4382-4402
SatFedGuard: Semi-Supervised Federated Contrastive Learning With RL-Assisted Bidirectional Distillation for Anomaly Traffic Detection in Satellite Networks
Federated learning-based intrusion detection methods for satellite networks enable model training without sharing local data, thereby ensuring network security while significantly reducing communication overhead. However, due to the difficulty of obtaining large-scale high-quality labeled data in satellite environments, a key challenge lies in how to train intrusion detection models using abundant unlabeled traffic data. We propose SatFedGuard, a semi-supervised federated contrastive learning approach for anomaly traffic detection in satellite networks. SatFedGuard effectively integrates unlabeled in-orbit data with labeled data from ground stations for model training. First, it models the unlabeled satellite traffic data using a contrastive learning framework. To address the challenge of non-IID data distribution, an attention-based dual-path aggregation strategy is designed to generate personalized models for each satellite by leveraging model similarities. Then, a bidirectional multi-granularity distillation method between larger and smaller models is implemented, where reinforcement learning is employed to optimize the weights of different loss terms dynamically. Experiments on two satellite network traffic datasets under non-IID settings demonstrate that the proposed method significantly improves anomaly detection performance while reducing dependence on in-orbit labeled data, achieving F1-Scores of 93.38% (↑11.63%) and 99.80% (↑8.72%), respectively.
- 20 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 4382-4402
SatFedGuard: Semi-Supervised Federated Contrastive Learning With RL-Assisted Bidirectional Distillation for Anomaly Traffic Detection in Satellite Networks
Federated learning-based intrusion detection methods for satellite networks enable model training without sharing local data, thereby ensuring network security while significantly reducing communication overhead. However, due to the difficulty of obtaining large-scale high-quality labeled data in satellite environments, a key challenge lies in how to train intrusion detection models using abundant unlabeled traffic data. We propose SatFedGuard, a semi-supervised federated contrastive learning approach for anomaly traffic detection in satellite networks. SatFedGuard effectively integrates unlabeled in-orbit data with labeled data from ground stations for model training. First, it models the unlabeled satellite traffic data using a contrastive learning framework. To address the challenge of non-IID data distribution, an attention-based dual-path aggregation strategy is designed to generate personalized models for each satellite by leveraging model similarities. Then, a bidirectional multi-granularity distillation method between larger and smaller models is implemented, where reinforcement learning is employed to optimize the weights of different loss terms dynamically. Experiments on two satellite network traffic datasets under non-IID settings demonstrate that the proposed method significantly improves anomaly detection performance while reducing dependence on in-orbit labeled data, achieving F1-Scores of 93.38% (↑11.63%) and 99.80% (↑8.72%), respectively.
- 15 April 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 198
SGEOM: Sample gradient entropy based oversampling method for the class-imbalanced problem in software defect prediction
- 15 April 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 198
SGEOM: Sample gradient entropy based oversampling method for the class-imbalanced problem in software defect prediction
- 17 March 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 272
Gateway placement optimization for cascading survivability enhancement in low earth orbit satellite networks
- 17 March 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 272
Gateway placement optimization for cascading survivability enhancement in low earth orbit satellite networks
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 3635-3654
EF-CPPA: Escrow-Free Conditional Privacy-Preserving Authentication Scheme for Real-Time Emergency Messages in Smart Grids
Timely and secure emergency message delivery is critical to resilient smart-grid operation and rapid disturbance response. However, existing schemes remain inadequate, leaving smart grids vulnerable to security and privacy threats and causing verification bottlenecks, particularly when nonlinear emergency measurements cannot be homomorphically aggregated, which prevents bandwidth-efficient in-network aggregation and scalable batch verification. We propose EF-CPPA, an escrow-free, conditional privacy-preserving authentication scheme for real-time emergency messaging in smart grids. EF-CPPA enables smart meters to deliver authenticated emergency messages to the CC via power gateways verifiable as legitimate relays, while ensuring the confidentiality, integrity, and unlinkability of embedded nonlinear measurements. EF-CPPA further provides conditional anonymity with accountable tracing, as well as origin authentication, intra-domain verification, and scalable batch verification under bursty multi-meter messaging. An ECDLP-based escrow-free key-generation mechanism reduces reliance on the CC and enables efficient node joining and revocation. Security analysis shows that EF-CPPA achieves existential unforgeability under chosen-message attacks (EUF-CMA) and satisfies the stated security and privacy requirements. Performance evaluation demonstrates low computational, communication, energy, and node-management overhead, making EF-CPPA suitable for security-critical, time-sensitive smart-grid emergency messaging.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 3635-3654
EF-CPPA: Escrow-Free Conditional Privacy-Preserving Authentication Scheme for Real-Time Emergency Messages in Smart Grids
Timely and secure emergency message delivery is critical to resilient smart-grid operation and rapid disturbance response. However, existing schemes remain inadequate, leaving smart grids vulnerable to security and privacy threats and causing verification bottlenecks, particularly when nonlinear emergency measurements cannot be homomorphically aggregated, which prevents bandwidth-efficient in-network aggregation and scalable batch verification. We propose EF-CPPA, an escrow-free, conditional privacy-preserving authentication scheme for real-time emergency messaging in smart grids. EF-CPPA enables smart meters to deliver authenticated emergency messages to the CC via power gateways verifiable as legitimate relays, while ensuring the confidentiality, integrity, and unlinkability of embedded nonlinear measurements. EF-CPPA further provides conditional anonymity with accountable tracing, as well as origin authentication, intra-domain verification, and scalable batch verification under bursty multi-meter messaging. An ECDLP-based escrow-free key-generation mechanism reduces reliance on the CC and enables efficient node joining and revocation. Security analysis shows that EF-CPPA achieves existential unforgeability under chosen-message attacks (EUF-CMA) and satisfies the stated security and privacy requirements. Performance evaluation demonstrates low computational, communication, energy, and node-management overhead, making EF-CPPA suitable for security-critical, time-sensitive smart-grid emergency messaging.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (7) , 15035-15055
LCE-PPDA: Lightweight Certificateless and Escrow-Free Privacy-Preserving Data Aggregation for UAV-Assisted IoT-Enabled Smart Grids
The convergence of uncrewed aerial vehicles (UAVs) and the Internet of Things (IoT) is expected to enhance sensing coverage, connectivity, and resilience in distributed smart grids, especially in remote or infrastructure-sparse regions. In this UAV-assisted, IoT-enabled paradigm, UAVs act as aerial relays that collect, aggregate, and forward sensing data between ground devices and control centers. However, privacy-preserving data aggregation (PPDA) in such settings still faces key-escrow vulnerabilities, certificate management overhead, incomplete privacy protection, and high computational and energy costs, particularly for signature verification at UAV relays and decryption at control centers. To address these challenges, we propose LCE-PPDA, a lightweight, certificateless, and escrow-free PPDA scheme tailored for UAV-assisted, IoT-enabled smart grids. LCE-PPDA eliminates key escrow through joint key generation, adopts a hierarchical timing structure with macro-interval rekeying and micro-interval reporting, and supports ciphertext-level in-network aggregation with both individual and batch authentication at UAV relays. To ensure privacy with accountability, it integrates certificateless signatures, dynamic pseudonyms, and session-bound key masking, achieving end-to-end confidentiality, conditional anonymity, unlinkability, and accountable traceability. Formal analysis shows that LCE-PPDA achieves correctness and EUF-CMA security in the random-oracle model (ROM) under the ECDLP assumption against both Type-I and Type-II adversaries. Performance evaluation further demonstrates that LCE-PPDA reduces computational, communication, and energy overheads compared with representative schemes, providing a scalable and lightweight foundation for secure PPDA in UAV-assisted, IoT-enabled smart grids.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (7) , 15035-15055
LCE-PPDA: Lightweight Certificateless and Escrow-Free Privacy-Preserving Data Aggregation for UAV-Assisted IoT-Enabled Smart Grids
The convergence of uncrewed aerial vehicles (UAVs) and the Internet of Things (IoT) is expected to enhance sensing coverage, connectivity, and resilience in distributed smart grids, especially in remote or infrastructure-sparse regions. In this UAV-assisted, IoT-enabled paradigm, UAVs act as aerial relays that collect, aggregate, and forward sensing data between ground devices and control centers. However, privacy-preserving data aggregation (PPDA) in such settings still faces key-escrow vulnerabilities, certificate management overhead, incomplete privacy protection, and high computational and energy costs, particularly for signature verification at UAV relays and decryption at control centers. To address these challenges, we propose LCE-PPDA, a lightweight, certificateless, and escrow-free PPDA scheme tailored for UAV-assisted, IoT-enabled smart grids. LCE-PPDA eliminates key escrow through joint key generation, adopts a hierarchical timing structure with macro-interval rekeying and micro-interval reporting, and supports ciphertext-level in-network aggregation with both individual and batch authentication at UAV relays. To ensure privacy with accountability, it integrates certificateless signatures, dynamic pseudonyms, and session-bound key masking, achieving end-to-end confidentiality, conditional anonymity, unlinkability, and accountable traceability. Formal analysis shows that LCE-PPDA achieves correctness and EUF-CMA security in the random-oracle model (ROM) under the ECDLP assumption against both Type-I and Type-II adversaries. Performance evaluation further demonstrates that LCE-PPDA reduces computational, communication, and energy overheads compared with representative schemes, providing a scalable and lightweight foundation for secure PPDA in UAV-assisted, IoT-enabled smart grids.
- 27 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (2) , 3196-3211
Distributed Anti-Cascading Routing Scheme Based on Fuzzy Logic in LEO Satellite Networks
LEO satellite networks provide global users with satellite internet services, yet the risk of cascading failures severely threatens the reliable operation of the network. Failures rapidly propagate from congested and overloaded satellites to normal satellites, causing the network to be trapped in a vicious cycle of periodic failures, unable to return to a stable state capable of providing regular services. Existing routing schemes focus on optimizing network performance under normal conditions, paying little attention to fault tolerance mechanisms after satellite failures, or the mechanisms are too simplistic to handle large-scale cascading scenarios of satellite failures, ultimately leading to rapid failure propagation or even complete network paralysis. To address these issues, this paper proposes a distributed anti-cascading routing scheme based on fuzzy logic, aimed at promptly initiating takeover routing after the network triggers cascading failures, preventing further failure propagation, and restoring the network to stable operation. A strategy for collecting and sharing information of two-hop neighboring satellites is introduced, and a routing decision method based on fuzzy logic is constructed, five main update steps for the membership functions based on a detailed analysis of cascading scenarios. Satellites can make routing decisions to reduce the risk of cascading based on two-hop neighbor information. Compared with six routing schemes under a case constellation, simulation results show that the proposed scheme can ensure stable recovery within 35% satellite failure. The results offering valuable suggestions for network managers.
- 27 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (2) , 3196-3211
Distributed Anti-Cascading Routing Scheme Based on Fuzzy Logic in LEO Satellite Networks
LEO satellite networks provide global users with satellite internet services, yet the risk of cascading failures severely threatens the reliable operation of the network. Failures rapidly propagate from congested and overloaded satellites to normal satellites, causing the network to be trapped in a vicious cycle of periodic failures, unable to return to a stable state capable of providing regular services. Existing routing schemes focus on optimizing network performance under normal conditions, paying little attention to fault tolerance mechanisms after satellite failures, or the mechanisms are too simplistic to handle large-scale cascading scenarios of satellite failures, ultimately leading to rapid failure propagation or even complete network paralysis. To address these issues, this paper proposes a distributed anti-cascading routing scheme based on fuzzy logic, aimed at promptly initiating takeover routing after the network triggers cascading failures, preventing further failure propagation, and restoring the network to stable operation. A strategy for collecting and sharing information of two-hop neighboring satellites is introduced, and a routing decision method based on fuzzy logic is constructed, five main update steps for the membership functions based on a detailed analysis of cascading scenarios. Satellites can make routing decisions to reduce the risk of cascading based on two-hop neighbor information. Compared with six routing schemes under a case constellation, simulation results show that the proposed scheme can ensure stable recovery within 35% satellite failure. The results offering valuable suggestions for network managers.
- 27 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 11 (12) , 22562-22577
UAV-Enabled IoT: Cascading Failure Model and Topology-Control-Based Recovery Scheme
The emergence of unmanned aerial vehicle (UAV)-enabled Internet of Things (IoT) represents a new paradigm of intelligent monitoring that transcends geographical limitations. However, UAVs are highly sensitive to disruptive factors, and their failure can easily trigger cascading failures due to a surge in traffic. To address this, this article proposes a cascading failure model to help understand cascading risks, incorporating the unique online and offline modes of UAVs. For the issue of IoT devices not being fully covered after a cascading failure, a topology-control-based recovery scheme is proposed. Surviving UAVs are divided into two categories, where backbone UAVs (BUAVs) hover at specific positions to form a reliable backbone network spanning the service area, and cruising UAVs patrol nearby to achieve extensive coverage. To provide a solution, the optimization model is transformed into two stages, and approximate algorithms are developed separately. The improved combination of cuckoo search and a density-based clustering algorithm is employed to determine the positions for BUAVs. Meanwhile, a multistep approach utilizing the nondominated sorting genetic algorithm II is introduced to optimize the cruising paths of UAVs. Case study shows that the proposed cascading failure model is reasonable; the proposed recovery scheme can achieve up to 91% coverage, with most IoT devices covered by cruising UAVs maintaining more than half of their service time; the recovered network can still maintain 50% of IoT devices within coverage under up to 20% link failure. These findings offer network managers practical insights for enhancing network resilience.
- 27 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 11 (12) , 22562-22577
UAV-Enabled IoT: Cascading Failure Model and Topology-Control-Based Recovery Scheme
The emergence of unmanned aerial vehicle (UAV)-enabled Internet of Things (IoT) represents a new paradigm of intelligent monitoring that transcends geographical limitations. However, UAVs are highly sensitive to disruptive factors, and their failure can easily trigger cascading failures due to a surge in traffic. To address this, this article proposes a cascading failure model to help understand cascading risks, incorporating the unique online and offline modes of UAVs. For the issue of IoT devices not being fully covered after a cascading failure, a topology-control-based recovery scheme is proposed. Surviving UAVs are divided into two categories, where backbone UAVs (BUAVs) hover at specific positions to form a reliable backbone network spanning the service area, and cruising UAVs patrol nearby to achieve extensive coverage. To provide a solution, the optimization model is transformed into two stages, and approximate algorithms are developed separately. The improved combination of cuckoo search and a density-based clustering algorithm is employed to determine the positions for BUAVs. Meanwhile, a multistep approach utilizing the nondominated sorting genetic algorithm II is introduced to optimize the cruising paths of UAVs. Case study shows that the proposed cascading failure model is reasonable; the proposed recovery scheme can achieve up to 91% coverage, with most IoT devices covered by cruising UAVs maintaining more than half of their service time; the recovered network can still maintain 50% of IoT devices within coverage under up to 20% link failure. These findings offer network managers practical insights for enhancing network resilience.
- 15 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 21 (2) , 1672-1689
A Dynamic Cascading Failure Model for LEO Satellite Networks
The advancement of space technology has positioned LEO satellite networks as a pivotal solution for global digitalization. However, satellites can fail due to various internal and external threats. Their limited buffer capacity and processing ability can exacerbate failures, posing a risk of cascading failures to the entire network. To delve deeper into the cascading process of LEO satellite networks, we have developed a dynamic cascading failure model. We propose a method to calculate the minimum time required for satellite movement to cover discrete zone units on Earth. By considering the dynamic changes in satellite coverage, we model the probability of user data reaching satellites using a Poisson process. Additionally, we have established models for satellites’ processing ability and buffer capacity to depict a more realistic satellite state transition. Satellites can become congested or even overloaded when they receive an excessive amount of data and require a corrective phase before returning to standard operations. We present a dynamic cascading algorithm, ensuring that the sequence of observations doesn’t impact the final cascading process outcome. Through detailed case studies involving three constellations, we found that cascading failures tend to propagate mainly to satellites within two hops. Smaller constellations are more susceptible to periodic cascading failures, while mega-constellations might experience rapid, severe avalanche effects. These findings offer practical insights for satellite network managers aiming to mitigate cascading risks.
- 15 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 21 (2) , 1672-1689
A Dynamic Cascading Failure Model for LEO Satellite Networks
The advancement of space technology has positioned LEO satellite networks as a pivotal solution for global digitalization. However, satellites can fail due to various internal and external threats. Their limited buffer capacity and processing ability can exacerbate failures, posing a risk of cascading failures to the entire network. To delve deeper into the cascading process of LEO satellite networks, we have developed a dynamic cascading failure model. We propose a method to calculate the minimum time required for satellite movement to cover discrete zone units on Earth. By considering the dynamic changes in satellite coverage, we model the probability of user data reaching satellites using a Poisson process. Additionally, we have established models for satellites’ processing ability and buffer capacity to depict a more realistic satellite state transition. Satellites can become congested or even overloaded when they receive an excessive amount of data and require a corrective phase before returning to standard operations. We present a dynamic cascading algorithm, ensuring that the sequence of observations doesn’t impact the final cascading process outcome. Through detailed case studies involving three constellations, we found that cascading failures tend to propagate mainly to satellites within two hops. Smaller constellations are more susceptible to periodic cascading failures, while mega-constellations might experience rapid, severe avalanche effects. These findings offer practical insights for satellite network managers aiming to mitigate cascading risks.