Chenxi Zhu

Chenxi Zhu

Auto-generated profile, based on publicly available information

Chenxi Zhu

School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, China

Research Focus & Metrics

Academic Career Length

17Years (2010-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2010 in Journal of Communications and Networks
Verified
Journal Article
Closed access
TGTolga Girici
Tolga Girici
CZChenxi Zhu
Chenxi Zhu
JAJonathan R. Agre
Jonathan R. Agre
AEAnthony Ephremides
Anthony Ephremides
In this work we consider the problem of downlink resource allocation for proportional fairness of long term received rates of data users and quality of service for real time sessions in an OFDMA-based wireless system. The base station allocates available power and subchannels to individual users based on long term average received rates, quality of service (QoS) based rate constraints and channel conditions. We formulate and solve a joint bandwidth and power optimization problem, solving which provides a performance improvement with respect to existing resource allocation algorithms. We propose schemes for flat as well as frequency selective fading cases. Numerical evaluation results show that the proposed method provides better QoS to voice and video sessions while providing more and fair rates to data users in comparison with existing schemes.

10 May 2021 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
Yangjie CaoYangjie Cao
Yangjie Cao
ZZZhiyi Zhou
Zhiyi Zhou
CZChenxi Zhu
Chenxi Zhu
Pengsong DuanPengsong Duan
Pengsong Duan
Xianfu ChenXianfu Chen
Xianfu Chen
Jie LiJie Li
Jie Li
WiFi-based identity recognition is predominant because of its noninvasive and ubiquitous advantages. However, existing approaches show slow training speed and limited applicability. In this article, we propose a lightweight deep learning model, named as lightweight WiFi-based identification (LW-WiID), to address these technical challenges. LW-WiID reconstructs original data of channel state information into frequency energy graph, which contains not only the temporal feature of the gait but also the spatial feature among subcarriers, ensuring the accuracy of identity recognition. Furthermore, a novel Balloon mechanism is designed to achieve the lightweight. Through information integration crossing both layers and channels, the Balloon mechanism effectively reduces the number of model parameters. Experimental results demonstrate that LW-WiID achieves an accuracy of 99.7% on a 50-person gait data set while the model size is compressed to 5.53% of the existing identity recognition approaches with the same accuracy.

6 February 2023 in Electronics
Verified
Journal Article
Open access
PDF file
HWHaobo Wang
Haobo Wang
CZChenxi Zhu
Chenxi Zhu
Yangjie CaoYangjie Cao
Yangjie Cao
YZYan Zhuang
Yan Zhuang
Jie LiJie Li
Jie Li
Xianfu ChenXianfu Chen
Xianfu Chen
Deep neural networks are susceptible to interference from deliberately crafted noise, which can lead to incorrect classification results. Existing approaches make less use of latent space information and conduct pixel-domain modification in the input space instead, which increases the computational cost and decreases the transferability. In this work, we propose an effective adversarial distribution searching-driven attack (ADSAttack) algorithm to generate adversarial examples against deep neural networks. ADSAttack introduces an affiliated network to search for potential distributions in image latent space for synthesizing adversarial examples. ADSAttack uses an edge-detection algorithm to locate low-level feature mapping in input space to sketch the minimum effective disturbed area. Experimental results demonstrate that ADSAttack achieves higher transferability, better imperceptible visualization, and faster generation speed compared to traditional algorithms. To generate 1000 adversarial examples, ADSAttack takes 11.08s and, on average, achieves a success rate of 98.01%.

4 September 2025 in IEEE/ASME Transactions on Mechatronics
Verified
Journal Article
Closed access
CZChenxi Zhu
Chenxi Zhu
HSHongwei Sun
Hongwei Sun
CCCheng Cao
Cheng Cao
DZDailin Zhang
Dailin Zhang
Tracking error prediction and compensation is essential to ensure the accuracy of robotic motion trajectories in various tasks. Due to the structural characteristics of serial robots, the tracking error is influenced by pose-dependent dynamics. To enhance the tracking control accuracy of robots, this study proposes a physics-informed machine learning (PIML)-assisted method for tracking error prediction and compensation. This method explicitly considers the influence of pose-dependent robotic dynamics and the characteristics of the closed-loop control mechanism through a dynamics-embedded neural network. Specifically, the theoretical current is calculated as a critical input for the neural network model, providing a physical constraint during the parameter update process. Moreover, a current-error dual-loop architecture is designed to help the neural network capture the mutual influence between the control current and tracking error within the robotic closed-loop control mechanism. Finally, the errors predicted by the proposed method are utilized for trajectory precompensation. The PIML-assisted tracking error prediction and compensation approach effectively combines the strengths of mechanistic and data-driven models, offering a robust solution for ensuring high-accuracy robotic tracking control. Experimental results demonstrate that the proposed approach achieves a 19.04% reduction in prediction error and significantly reduces tracking error by 74.74%, which is 4.25% greater than the improvement obtained by normal ML-based compensation methods.

22 February 2024 in IEEE Sensors Journal
Verified
Journal Article
Closed access
WZWanpeng Zhang
Wanpeng Zhang
KKKang Kong
Kang Kong
CZChenxi Zhu
Chenxi Zhu
PZPeng Zhang
Peng Zhang
DZDailin Zhang
Dailin Zhang
The shield machine is a critical piece of equipment for tunnel construction and necessitates high-precision attitude measurement for ensuring construction quality. During tunneling through hard rocks, the machine produces intense vibrations, which detrimentally affect the precision of attitude measurement instruments, subsequently impacting the accuracy of the machine’s attitude assessment. In this paper, we present a novel approach that introduces a Double-Weighted Information Fusion method to coordinate the capabilities of fiber optic gyroscopes and inclinometers. This method dynamically adjusts two parameters, allocating distinct weights to the posterior estimates of the Kalman Filter (KF) and the original angular velocity values. Additionally, it integrates a Complementary Filter with a KF, aiming to counteract the phase lag introduced by the KF during periods of attitude change. Extensive comparison experiments validate the algorithm’s effectiveness. Experimental results demonstrate that the proposed Double-weighted Fusion algorithm decreases the average angular measurement error by 50% relative to the KF, and by 46.7% compared to the Adaptive Kalman Filter. The proposed method enhances the system’s capability to measure attitude angles. It is well-suited for high-precision attitude measurement of engineering equipment operating under conditions of strong vibrations.

21 March 2022 in Journal of Sensors
Verified
Journal Article
Open access
PDF file
CZChenxi Zhu
Chenxi Zhu
HWHaobo Wang
Haobo Wang
YZYan Zhuang
Yan Zhuang
Jie LiJie Li
Jie Li
Yangjie CaoYangjie Cao
Yangjie Cao
Deep neural networks-based image classification systems could suffer from adversarial attack algorithms, which generate input examples by adding deliberately crafted yet imperceptible noise to original inputs. To reduce the impact on human visual sense and to ensure adversarial attack ability, the input image needs to be modified by pixels in considerable iterations which is time consuming. By using sparse mapping network to map the input into a higher dimensional space, searching space of adversarial perturbation distribution is enlarged to better acquire perturbation information. Taking both searching speed and searching effectiveness into consideration, sparsity limitation is introduced to suppress unnecessary neurons during parameter updating process. Based on different eye sensitivity of different colors, maps of each color channel are disturbed by perturbations with different strengths to reduce visual perception. Numerical experiments confirm that compared with the state-of-the-art adversarial attack algorithms, the proposed SparseAdv performs a relatively high attack ability, better imperceptible visualization, and faster generation speed.