HLHanxiao Liu

Hanxiao Liu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Hanxiao Liu has a total of 34 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Rollout-Based Decision-Making for Autonomous Driving Left Turn at Unsignalized Intersections

    by Hanxiao Liu, Ronghui Zhang, Tianlong Wan

    Unsignalized intersections pose significant safety and efficiency challenges for autonomous driving due to ambiguous right-of-way and complex, multi-directional vehicle interactions within overlapping conflict zones. In this study, we propose a Rollout-Enhanced Attention (REA) framework, which synergizes a rollout algorithm with a multi-head cross-attention mechanism to optimize driving decisions. Specifically, the multi-head cross-attention mechanism effectively extracts interaction features between the ego vehicle and surrounding vehicles, prioritizing those with higher collision risks. Incorporating these features, the rollout algorithm executes online multi-step lookahead planning by leveraging the action-value estimates from a Deep Q-Network (DQN), thereby optimizing action selection. Extensive experiments in unsignalized intersection scenarios with varying traffic densities demonstrate that the REA model outperforms standard DQN and other baselines. Furthermore, comparisons between different rollout variations highlight the framework’s advantages in safety, efficiency, and decision-making accuracy.

  • Event-Triggered Quantized Fault-Tolerant Control for Multiagent Systems Under Aperiodic DoS Attacks: A Mode-Switching Mechanism

    by Chun Liu, Haiwen Wang, Hanxiao Liu, Xiaoqiang Ren, Xiaofan Wang

    This paper presents a mode-switching event triggered quantized fault-tolerant control strategy for multi-agent systems under cyber-physical threats. The strategy uniquely prioritizes the communication layer, streamlining controller design while simultaneously mitigating actuator faults, denial-of-service attacks, bandwidth constraints, and communication frequency restrictions. Firstly, the distributed architecture is adopted for state estimation storage and the design of a fault-tolerant controller to guarantee ultimately bounded consensus. Secondly, a dual-mode event-triggered quantization strategy is developed to facilitate inter-agent communication with limited resources under both attack-free and attack-active scenarios. In the normal mode, a triggering threshold with time-varying contraction characteristics is utilized to enhance quantization precision, thereby refining control performance during consensus convergence. Upon switching to the abnormal mode, the triggering threshold is dynamically reconfigured to simultaneously counteract the combined effects of sequential denial-of-service attacks and actuator faults. Furthermore, the upper bound of the control input, the lower bound of the inter-event intervals, and the average number of bits required to achieve consensus are analytically derived. Simulation results demonstrate the effectiveness and superiority of the proposed strategy.


  • Swarm Trajectory Planning Using Dynamic Corridors for Dead Zone Avoidance

    by Xuan Wei, Juncheng Li, Hanxiao Liu, Panpan Zhou, Xu Fang

    Efficient obstacle avoidance in Unmanned Aerial Vehicle (UAV) formations remains a significant challenge, particularly within dynamic and cluttered environments. While substantial advancements have been made in existing methods for formation control and obstacle avoidance, they frequently encounter constraints when segments of the UAV swarm become trapped in dead zones, which significantly degrade overall operational performance. To mitigate this issue, we propose a novel approach aimed at enhancing the swarm’s resilience by enabling UAVs to detect and proactively avoid potential dead zones. The proposed method constructs dynamic flight corridors by inflating the global trajectories of individual UAVs and systematically assessing their deviation from the primary formation path. This framework facilitates the early detection of potential dead zones, ensuring the swarm maintains both a safe and cohesive flight configuration. Extensive simulations across diverse environmental settings demonstrate the robustness and generalizability of the proposed method, underscoring its effectiveness in improving both formation stability and obstacle avoidance capabilities.

  • Optimal Scheduling of Importance-Based DoS Attack in Multisystems

    by Jiacheng Wang, Hanxiao Liu, Chun Liu, Xiaoqiang Ren, Xiaofan Wang

    This paper is concerned with the security of remote state estimation in multi-systems. Different from most works, we study a novel denial of service (DoS) attack, which we call an importance-based DoS attack. Here, subsystem importance is defined as the quantitative impact of each subsystem on the overall state estimation accuracy, which is critical for ensuring correct decision-making. This work assumes that there are N subsystems, and the DoS attacker can only target M out of N communication channels due to energy constraints. The attacker’s goal is to determine the optimal strategy to degrade the performance of remote state estimation. Two attack scenarios are considered: a model-based scenario where the attacker has knowledge of the overall dynamics and parameters of the system and a model-free scenario where the attacker lacks access to the wireless channel model. The problem is formulated as a Markov decision process (MDP) with an infinite time horizon. Additionally, the optimal attack strategies corresponding to different importance function forms under various scenarios are considered, and solutions for some specific cases are provided. Numerical examples are presented to demonstrate the effectiveness of the proposed strategies.


  • Promotion of cooperation in a structured population with environmental feedbacks

    by Ding Lyu, Hanxiao Liu, Chuang Deng, Xiaofan Wang

    Cooperation is a representative altruistic behavior in which individuals contribute public goods to benefit their neighborhoods and even larger communities in social networks. The defective behavior is more likely to bring higher payoffs than the cooperative behavior, which makes the cooperative behavior hard to maintain and sustain. Many mechanisms were proposed to promote cooperation within a social dilemma, in which some recent studies introduced the impact of dynamically changing environments on players’ payoffs and strategies in social-ecological systems, and evolutionary-ecological systems. However, degree heterogeneity, an important structural property of many real-world complex networks such as social networks, academic collaboration networks, and communication networks, is rarely explored and studied in such eco-evolutionary games. In this research, we propose a Public Goods Game model on social networks with environmental feedback and analyze how the environmental factor and network structure affect the evolution of cooperation. It is found that as the initial environmental factors and the cooperation-enhancement defection-degradation ratio increase, the steady cooperation level of the social network significantly increases, and the dynamic environment will eventually evolve into a high-return environment; On the other hand, even if the initial environmental benefit coefficient is high, when the cooperation-enhancement defection-degradation ratio is less than a threshold, the dynamic environment will gradually degrade into a low-return environment. The steady cooperation level of the social network first gradually increases as the network structure becomes more heterogeneous, but it will decrease once the heterogeneity of the social network exceeds a certain threshold.


  • Optimal Privacy-Preserving Transmission Schedule Against Eavesdropping Attacks on Remote State Estimation

    by Jiaying Zou, Hanxiao Liu, Chun Liu, Xiaoqiang Ren, Xiaofan Wang

    This letter is concerned with the preservation of privacy in remote state estimation of cyber-physical systems. A privacy-preserving transmission scheduling strategy against eavesdropping is proposed, incorporating three operational modes for sensors: silence, direct transmission, and noise-injected transmission. This strategy is designed to minimize the transmission cost and estimation error covariance for the remote estimator while maximizing the estimation error covariance for eavesdroppers. Threshold structures are demonstrated for optimal transmission schedules in different scenarios. Additionally, a novel correlation between the optimal transmission choice and the magnitude of injection noise is presented, particularly pertinent to scenarios involving direct transmission and transmission with injection noise. This correlation is important in balancing transmission information integrity against privacy concerns. Finally, several numerical examples are presented to demonstrate the effectiveness of the theoretical results.


  • Linear Attacks Against Remote State Estimation: Performance Analysis Under an Encryption Scheme

    by Hanxiao Liu, Yuqing Ni, Xiaofan Wang, Lihua Xie

    This paper is concerned with the performance analysis of innovation-based remote state estimation with an encryption scheme under linear attacks. We investigate performances from the perspectives of both the system defender and the malicious attacker. For the system defender, we begin by detailing the design of an encryption mechanism that achieves an optimal trade-off between the estimation performance and the detection probability. Subsequently, we explore two scenarios where the attacker is aware of the existence of the proposed scheme. In these scenarios, we formulate the problems of seeking optimal linear attack strategies as convex optimization problems, and semi-definite programming (SDP) is employed to find their numerical solutions. In particular, a closed-form expression of the worst-case attacks is obtained when the attacker can observe the system state through its own sensor deployment. Moreover, we conduct a comparative analysis between these two scenarios and prove that the worst-case attack strategies have their respective advantages and disadvantages under different cases. Finally, some numerical examples are given to illustrate the theoretical results.


  • Constrained Residual Race: An Efficient Hybrid Controller for Autonomous Racing

    by Mengmeng Zou, Yiwen Lu, Hanxiao Liu, Yilin Mo, Xiaoqiang Ren

    Deep reinforcement learning (DRL) has shown promise in solving the autonomous racing problem without requiring an accurate dynamics model, but random sampling of actions can lead to inefficient exploration. This paper proposes an efficient racing method, Constrained Residual Race (CRR), which combines a geometry-based reactive obstacle avoidance algorithm, Follow-The-Gap Method (FTG), with DRL. CRR generates safe guided actions with FTG and complementary actions with DRL, which are then fused into a final control action through a constraint equation. Using FTG as a prior controller enhances exploration efficiency and produces safe driving actions. Constraining the DRL policy output ensures that it does not overwrite the prior policy during early stages of exploration. Additionally, the DRL policy improves over the performance of the prior controller during training, reducing the lap time of the car. Validation on the F1TENTH simulator shows that CRR achieves lower lap times in most cases and exhibits excellent generalization capability.

  • A Privacy-Preserving Approach Against Eavesdropping Attacks on Remote State Estimation for Cyber–Physical Systems

    by Hanxiao Liu, Jiaying Zou, Xiaoqiang Ren, Xiaofan Wang

    This paper is concerned with the issue of privacy preservation against eavesdropping attacks on remote state estimation for cyber-physical systems. We aim to propose a privacy-preserving approach for remote state estimation to achieve an optimal trade-off between the estimation performance and the privacy-preserving level. We first investigate the design of a privacy-preserving scheme and present a closed-form solution for a signal system. Then, we present how to design the privacy-preserving signal for multi-systems. Note that different subsystems or local state information possess estimation performance and privacy-preserving requirements, we are inspired to propose a novel importance-aware privacy-preserving scheme. Correspondingly, we present how to design a privacy-preserving scheme for multi-systems to achieve an overall balance between estimation performance and privacy preservation. Finally, some numerical examples are given to illustrate the theoretical results.