Zhiqiang Zhang

Dr. Zhiqiang Zhang

orcid

Research Focus & Metrics

Academic Career Length

18Years (2009-Present)

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

Representative Publications

View All Publications
14 November 2019 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
YNYongqiang Nai
Yongqiang Nai
QYQingyu Yang
Qingyu Yang
ZZZhiqiang Zhang
Zhiqiang Zhang
Effectively compensating unknown intermittent actuator faults in uncertain decentralized nonlinear systems is a very difficult problem, and very few results have been obtained. In this article, to address this issue, an adaptive neural output feedback compensation control scheme based on command-filtered backstepping is developed. First, we design a bank of observers to estimate the system states and utilize neural networks with random hidden nodes to approximate the unknown functions of these observers. Second, a smooth projection algorithm is used to online update estimated parameters in the controllers such that the possible ceaseless increase in the estimated parameters caused by intermittent actuator faults can be eliminated. Due to the presence of intermittent jumps of unknown parameters, a modified Lyapunov function is developed to analyze the system stability. It is proved that the boundedness of all closed-loop system signals is ensured and the ultimate bound of the tracking error depends on design parameters, adjustable jumping amplitude of Lyapunov function, and minimum fault time interval. Third, by analyzing the system transient performance, the peaking phenomenon at the starting instant of the system operation can be removed, and a root mean square type of bound is established to illustrate that the transient tracking error performance is tunable by design parameters. Finally, simulations studies are done to illustrate the effectiveness of the theoretical results.

1 January 2022 in Applied Soft Computing
Verified
Journal Article
Closed access
ZZZhiqiang Zhang
Zhiqiang Zhang
QYQingyu Yang
Qingyu Yang

2 September 2022 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
ZZZhiqiang Zhang
Zhiqiang Zhang
QYQingyu Yang
Qingyu Yang
YZYanyang Zi
Yanyang Zi
QTQi Tan
Qi Tan
Single-layer representation learning (SLRL) is very promising in automatically learning features for gearbox fault diagnosis. Most of the existing autoencoder (AE)-based SLRL methods perform feature learning by mining intrinsic correlations of the data, whereas the learned features may be insufficiently discriminative when the vibration signals are very complex. To cope with this problem, we propose a new SLRL method named discriminative sparse AE (DSAE), which incorporates label information of the data through a discriminator. By exploring intrinsic correlations and exploiting label information, DSAE is able to learn sufficient discriminative features from complex vibration signals. The proposed DSAE is evaluated by experiments on a gearbox dataset. The results show that DSAE exhibits superior and more reliable performance with acceptable training efficiency in terms of diagnosis accuracy, stability, and anti-noise capability, by comparing with the current state-of-the-art SLRL methods. More significantly, a meaningful discover is that the discriminator in DSAE can also play the role of feature selection with better ability over traditional feature selectors.

5 August 2020 in IEEE Transactions on Systems, Man, and Cybernetics: Systems
Verified
Journal Article
Closed access
ZZZhiqiang Zhang
Zhiqiang Zhang
QYQingyu Yang
Qingyu Yang
QTQi Tan
Qi Tan
Although representation learning (RL) has achieved great success in intelligent fault diagnosis, existing RL methods still have two deficiencies: 1) all learned bases are employed for fault diagnosis, which may degrade the computational efficiency and diagnosis accuracy and 2) it is unable to know which bases are more useful or less useful for fault diagnosis. In this work, we present a novel RL method, namely, sparse filtering with adaptive basis weighting (SFABW) whose architecture is a three-layer neural network. The first and the last two layers are responsible for basis learning and basis weighting, respectively. We formulate a loss function to model such architecture and develop an iterative algorithm to minimize it, also we prove the convergence of this algorithm in theory. Through optimizing the whole network, we are able to obtain a group of bases together with their weights simultaneously. A subset of top-ranked bases with great weights is retained while the rest bases are discarded. The experimental results on a motor bearing dataset and a gear dataset have demonstrated the effectiveness of our method.

30 August 2024 in Electronics
Verified
Journal Article
Open access
PDF file
DLDanhua Liu
Danhua Liu
DHDan Huang
Dan Huang
XCXiming Chen
Ximing Chen
JDJian Dou
Jian Dou
LTLi Tang
Li Tang
ZZZhiqiang Zhang
Zhiqiang Zhang
Power theft and other abnormal power consumption behaviors seriously affect the safety, reliability, and stability of the power grid system. The traditional abnormal power consumption detection methods have complex models and low accuracy. In this paper, an unsupervised abnormal power consumption detection method based on multi-cluster feature selection and the Gaussian mixture model is proposed. First of all, twelve features are extracted from the load sequence to reflect the overall form, fluctuation, and change trend of the user’s electricity consumption. Then, multi-cluster feature selection algorithm is employed to select a subset of important features. Finally, based on the selected features, the Gaussian mixture model is formulated to cluster the normal power users and abnormal power users into different groups, so as to realize abnormal power consumption detection. The proposed method is evaluated through experiments based on a power load dataset from Anhui Province, China. The results show that the proposed method works well for abnormal power consumption detection, with significantly superior performance comapred to the traditional approaches in terms of the popular binary evaluation indicators like recall rate, precision rate, and F-score.

1 October 2009
Verified
Conference Paper
Closed access
LFLuo Feng
Luo Feng
RLRuan Liting
Ruan Liting
WSWu Shunjun
Wu Shunjun
ZQZhao Qiansheng
Zhao Qiansheng
ZZZhiqiang Zhang
Zhiqiang Zhang
In the radar field, Non-Linear Frequency Modulation (NLFM) signals are widely used. The conventional design method of side-lobe suppression filter for NLFM signal is based on the principle of stationary phase with the window functions and the modified spectrum algorithm. A new filtering technique of Modified Spectrum Filter based on Mismatched Window (MSFMW) algorithm is proposed in this paper. This method respectively designs the NLFM signal and the pulse compression filter based on different window functions. So the spectrum magnitude of the filtered waveform is proportional to the window function which is different from the one used for waveform design. The performance of the pulse compression can be modified by the different window functions. The simulation results demonstrate that the Ratio of Main-lobe peak to Side-lobe peak (RMS) can be increased by 0.8 ~ 7.7dB, without changing the main-lobe width.