Lijun Zhu

Lijun Zhu

Auto-generated profile, based on publicly available information

Lijun Zhu

Huazhong University of Science and Technology

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

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

Representative Publications

View All Publications
10 August 2023 in IEEE Transactions on Robotics
Verified
Journal Article
Closed access
SXShaohang Xu
Shaohang Xu
LZLijun Zhu
Lijun Zhu
HZHai-Tao Zhang
Hai-Tao Zhang
CHChin Pang Ho
Chin Pang Ho
This article considers quadruped locomotion control in the presence of uncertainties. Two types of structured uncertainties are considered, namely, uncertain friction constraints and uncertain model dynamics. Then, a min-max optimization model is formulated based on robust optimization, and a robust min-max model predictive controller is proposed by recurrently solving the optimization model. We prove that the min-max optimization model is equivalent to a convex quadratic constrained quadratic program by exploiting the structure of uncertainties. Moreover, a two-stage optimization algorithm is proposed to solve the optimization problem efficiently, allowing for the deployment of the controller onto the real robot. The results show that the proposed optimization algorithm can improve solving frequency by $\sim$ 11× compared with Gurobi. The proposed controller is able to stabilize quadruped locomotion in challenging scenarios where the uncertainties are caused by significant disturbances and unknown environments.

19 January 2024 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
RLRenzhi Lu
Renzhi Lu
RBRuichang Bai
Ruichang Bai
Ruidong LiRuidong Li
Ruidong Li
LZLijun Zhu
Lijun Zhu
MSMingyang Sun
Mingyang Sun
YDYuemin Ding
Yuemin Ding
Load forecasting is critical to the task of energy management in power systems, for example, balancing supply and demand and minimizing energy transaction costs. There are many approaches used for load forecasting such as the support vector regression (SVR), the autoregressive integrated moving average (ARIMA), and neural networks, but most of these methods focus on single-step load forecasting, whereas multistep load forecasting can provide better insights for optimizing the energy resource allocation and assisting the decision-making process. In this work, a novel sequence-to-sequence (Seq2Seq)-based deep learning model based on a time series decomposition strategy for multistep load forecasting is proposed. The model consists of a series of basic blocks, each of which includes one encoder and two decoders; and all basic blocks are connected by residuals. In the inner of each basic block, the encoder is realized by temporal convolution network (TCN) for its benefit of parallel computing, and the decoder is implemented by long short-term memory (LSTM) neural network to predict and estimate time series. During the forecasting process, each basic block is forecasted individually. The final forecasted result is the aggregation of the predicted results in all basic blocks. Several cases within multiple real-world datasets are conducted to evaluate the performance of the proposed model. The results demonstrate that the proposed model achieves the best accuracy compared with several benchmark models.

18 December 2023 in IEEE Transactions on Robotics
Verified
Journal Article
Closed access
CJCheng Jiang
Cheng Jiang
Wen-Long LiWen-Long Li
Wen-Long Li
Wenpan LiWenpan Li
Wenpan Li
DWDong-Fang Wang
Dong-Fang Wang
LZLi-Jun Zhu
Li-Jun Zhu
Han DingHan Ding
Han Ding
Calibrating unknown transformation relationships is an essential task for multi-robot cooperative systems. Traditional linear methods are inadequate to decouple and simultaneously solve the unknown matrices due to their intercoupling. This paper proposes a novel dual-robot accurate calibration method that uses convex optimization and Lie derivative to solve the dual-robot calibration problem simultaneously. The key idea is that a convex optimization model based on dual-robot transformation chain is established using Lie representation of SE(3). The Jacobian matrix of the established optimization model is explicitly derived using the corresponding Lie derivative of SE(3). To balance the influence of the magnitudes of the rotational and translational optimization variables, a weight coefficient is defined. Due to the closure and smoothness of Lie group, the optimization model can be solved simultaneously using Newton-like iterative methods without additional orthogonalization processing. The performance of the proposed method is verified through simulation and actual calibration experiments. The results show that the proposed method outperforms the previous calibration methods in terms of accuracy and stability. The actual experiments are used to compare the proposed method with two existing calibration methods, and the mean measurement error of a certified ceramic sphere is reduced from 0.9205mm and 0.5363mm to 0.4381mm, respectively.

28 October 2019 in IEEE Transactions on Automatic Control
Verified
Journal Article
Closed access
GKGulam Dastgir Khan
Gulam Dastgir Khan
ZCZhiyong Chen
Zhiyong Chen
LZLijun Zhu
Lijun Zhu
A class of event-triggered stabilization technique works with a small gain theorem when the controlled system has a certain input-to-state stability (ISS) property from the measurement (sensor) disturbance to state. However, such an ISS property is not always achievable for nonlinear systems. In particular, early research suggests that global internal stabilizability does not imply global external stabilizability even for small sensor disturbances. This note proposes a new event-triggered stabilization technique that requires a milder condition that the controlled system has a certain ISS property from the input (actuator) disturbance to state. The technique is further applied to explicitly construct an event-triggered controller for the output regulation problem of a class of nonlinear systems. With the design, Zeno behavior is avoided.

18 October 2024 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
RLRenzhi Lu
Renzhi Lu
XWXiaotao Wang
Xiaotao Wang
YDYiyu Ding
Yiyu Ding
HZHai-Tao Zhang
Hai-Tao Zhang
FZFeng Zhao
Feng Zhao
YHYong He
Yong He
In this article, an optimal surrounding control algorithm is proposed for multiple unmanned surface vessels (USVs), in which actor-critic reinforcement learning (RL) is utilized to optimize the merging process. Specifically, the multiple-USV optimal surrounding control problem is first transformed into the Hamilton–Jacobi–Bellman (HJB) equation, which is difficult to solve due to its nonlinearity. An adaptive actor-critic RL control paradigm is then proposed to obtain the optimal surround strategy, wherein the Bellman residual error is utilized to construct the network update laws. Particularly, a virtual controller representing intermediate transitions and an actual controller operating on a dynamics model are employed as surrounding control solutions for second-order USVs; thus, optimal surrounding control of the USVs is guaranteed. In addition, the stability of the proposed controller is analyzed by means of Lyapunov theory functions. Finally, numerical simulation results demonstrate that the proposed actor-critic RL-based surrounding controller can achieve the surrounding objective while optimizing the evolution process and obtains 9.76% and 20.85% reduction in trajectory length and energy consumption compared with the existing controller.

23 January 2020 in IEEE Transactions on Automatic Control
Verified
Journal Article
Closed access
LZLijun Zhu
Lijun Zhu
DHDavid J. Hill
David J. Hill