Key Laboratory of Autonomous Systems and Networked Control

Guangzhou, China

Key Laboratory of Autonomous Systems and Networked Control

Guangzhou, China
SEARCH FILTERS
Time filter
Source Type

Ding Y.,SYSU CMU Shunde International Joint Research Institute | Li X.,Sun Yat Sen University | Zhang D.,Sun Yat Sen University | Tan H.,Key Laboratory of Autonomous Systems and Networked Control | Zhang Y.,Sun Yat Sen University
2016 3rd International Conference on Systems and Informatics, ICSAI 2016 | Year: 2016

In this paper, discontinuous signum-Activated ZD (Zhang dynamics) controllers are proposed as an attempt firstly to handle the agitator tank. For the agitator tank, there are two control variables, which are accurate concentration and moderate liquid level. In other words, the proposed discontinuous signum-Activated ZD controllers are designed to make the two variables of the agitator tank converge to desired trajectories. Moreover, signum function, as a kind of discontinuous activated function, has easier circuit implementation and thus is introduced as an activation function for handling agitator tank as an attempt. Besides, different tracking trajectories are used, and corresponding simulation results are shown in this paper. The simulation results, compared with the ZD controllers using the linear activation function, further verify the relative effectiveness of the discontinuous signum-Activated ZD controllers. In addition, the simulation results also simultaneously show that the discontinuous signum-Activated ZD controllers of the agitator tank may oscillate under certain frequencies. © 2016 IEEE.


Zhang Y.,Sun Yat Sen University | Zhang Y.,SYSU CMU Shunde International Joint Research Institute | Zhang Y.,Key Laboratory of Autonomous Systems and Networked Control | Qiu B.,Sun Yat Sen University | And 9 more authors.
Nonlinear Dynamics | Year: 2017

The pendulum control of the inverted-pendulum-on-a-cart (IPC) system is one of the most important issues in nonlinear control theory and has been widely investigated. Nevertheless, the control of pendulum tracking and swinging up has often been addressed separately. In this paper, by combining the zeroing dynamics and the conventional gradient dynamics, two concise zeroing-gradient (ZG) controllers (termed, z2g0 controller and z2g1 controller, respectively) are constructed for the IPC system. Importantly, the proposed z2g1 controller not only realizes the simultaneous control of pendulum swinging up and pendulum angle tracking, but also solves the singularity problem elegantly without using any switching strategy. Besides, the ZG method is compared with the optimal control method and the backstepping method. The theoretical analyses about the convergence performance of z2g0 and z2g1 controllers are further presented. Moreover, the boundedness of both control input u and its derivative (Formula presented.) of the z2g1 controller is proved. Three illustrative examples are carried out to demonstrate the tracking performance of z2g0 and z2g1 controllers for the pendulum tracking control. In particular, the efficacy and superiority of z2g1 controller for the control of pendulum tracking (including swinging up) of the IPC system in conquering the singularity problem are substantiated by comparative results. Furthermore, this paper investigates the robustness of the proposed ZG controllers (as well as the ZG design method) in the situations of time delay and disturbance. © 2017 Springer Science+Business Media Dordrecht


Zhang Y.,Sun Yat Sen University | Zhang Y.,SYSU CMU Shunde International Joint Research Institute | Zhang Y.,Key Laboratory of Autonomous Systems and Networked Control | Ding Y.,Sun Yat Sen University | And 7 more authors.
Information Processing Letters | Year: 2017

A new type of Zhang neural network (ZNN), which is activated by the signum-function array, is proposed for linear systems solving. Such a signum-function array activated ZNN is developed on the basis of a vector-valued error function instead of a scalar-valued norm-based energy function. Besides, a theorem is provided to illustrate the excellent finite-time convergence property of the new-type ZNN. In addition, the corresponding circuit schematic of the signum-function array activated ZNN is given. For better illustration, a representative simulative example is presented and the corresponding simulation result is shown to substantiate the efficacy of the proposed new-type ZNN for linear systems solving. Besides, the comparative simulation result further shows the desired finite-time convergent performance. © 2017 Elsevier B.V.


Chen D.,Sun Yat Sen University | Chen D.,SYSU CMU Shunde International Joint Research Institute | Chen D.,Key Laboratory of Autonomous Systems and Networked Control | Zhang Y.,Sun Yat Sen University | And 2 more authors.
IEEE Transactions on Automation Science and Engineering | Year: 2017

In this paper, a hybrid multi-objective scheme is proposed to complete simultaneously four objectives, i.e., the specified primary task for the end-effector, obstacle avoidance, joint-physical limits avoidance, and repetitive motion of redundant robot manipulators. In addition, corresponding theoretical analysis is given, which guarantees the validity of the proposed scheme. Then, the proposed hybrid multi-objective scheme is reformulated as a dynamical quadratic program (DQP) problem. The optimal solution of the DQP problem is found by the PLPE (piecewise-linear projection equation) neural network, i.e., PLPENN, and also by the corresponding numerical algorithm implemented on the computer. Furthermore, simulation and comparison based on a six-link planar redundant robot manipulator substantiate the effectiveness and accuracy of the proposed scheme. At last, a hardware experiment is conducted on a six-link physical robot manipulator system, which substantiates the physical realizability, operational stability, and safety of the proposed hybrid multi-objective scheme. © 2015 IEEE.


Chen D.,Sun Yat Sen University | Chen D.,Key Laboratory of Autonomous Systems and Networked Control | Chen D.,SYSU CMU Shunde International Joint Research Institute | Zhang Y.,Sun Yat Sen University | And 2 more authors.
International Journal of Systems Science | Year: 2017

Dual-arm redundant robot systems are usually required to handle primary tasks, repetitively and synchronously in practical applications. In this paper, a jerk-level synchronous repetitive motion scheme is proposed to remedy the joint-angle drift phenomenon and achieve the synchronous control of a dual-arm redundant robot system. The proposed scheme is novelly resolved at jerk level, which makes the joint variables, i.e. joint angles, joint velocities and joint accelerations, smooth and bounded. In addition, two types of dynamics algorithms, i.e. gradient-type (G-type) and zeroing-type (Z-type) dynamics algorithms, for the design of repetitive motion variable vectors, are presented in detail with the corresponding circuit schematics. Subsequently, the proposed scheme is reformulated as two dynamical quadratic programs (DQPs) and further integrated into a unified DQP (UDQP) for the synchronous control of a dual-arm robot system. The optimal solution of the UDQP is found by the piecewise-linear projection equation neural network. Moreover, simulations and comparisons based on a six-degrees-of-freedom planar dual-arm redundant robot system substantiate the operation effectiveness and tracking accuracy of the robot system with the proposed scheme for repetitive motion and synchronous control. © 2017 Informa UK Limited, trading as Taylor & Francis Group


Yin M.,Guangdong University of Technology | Yin M.,Key Laboratory of Autonomous Systems and Networked Control | Gao J.,Charles Sturt University | Lin Z.,Peking University | Lin Z.,Shanghai JiaoTong University
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2016

Low-rank representation (LRR) has recently attracted a great deal of attention due to its pleasing efficacy in exploring low-dimensional subspace structures embedded in data. For a given set of observed data corrupted with sparse errors, LRR aims at learning a lowest-rank representation of all data jointly. LRR has broad applications in pattern recognition, computer vision and signal processing. In the real world, data often reside on low-dimensional manifolds embedded in a high-dimensional ambient space. However, the LRR method does not take into account the non-linear geometric structures within data, thus the locality and similarity information among data may be missing in the learning process. To improve LRR in this regard, we propose a general Laplacian regularized low-rank representation framework for data representation where a hypergraph Laplacian regularizer can be readily introduced into, i.e., a Non-negative Sparse Hyper-Laplacian regularized LRR model (NSHLRR). By taking advantage of the graph regularizer, our proposed method not only can represent the global low-dimensional structures, but also capture the intrinsic non-linear geometric information in data. The extensive experimental results on image clustering, semi-supervised image classification and dimensionality reduction tasks demonstrate the effectiveness of the proposed method. © 1979-2012 IEEE.


Zhang Y.,Sun Yat Sen University | Zhai K.,Sun Yat Sen University | Chen D.,SYSU CMU Shunde International Joint Research Institute | Jin L.,SYSU CMU Shunde International Joint Research Institute | Hu C.,Key Laboratory of Autonomous Systems and Networked Control
Mathematics and Computers in Simulation | Year: 2015

Zhang-gradient (ZG) method is a combination of Zhang dynamics (ZD) and gradient dynamics (GD) methods which are two powerful methods for online time-varying problems solving. ZG controllers are designed using the ZG method to solve the tracking control problem of a certain system. In this paper, the design process of the ZG controllers with explicit as well as implicit tracking control of the double-integrator system is presented in detail. In addition, the corresponding computer simulations are conducted with different values of the design parameter λ to illustrate the effectiveness of ZG controllers. However, even though the ZG controllers are powerful, there is still a challenge in the simulation practice. Specifically, different settings of simulation options in MATLAB ordinary differential equation (ODE) solvers may lead to different simulation results (e.g.,failure and success). For better comparison, the successful and failed simulation results are both presented. The differences in simulation results remind us to pay more attention to MATLAB defaults and options when we conduct such simulations. © 2015 International Association for Mathematics and Computers in Simulation (IMACS).


Zhang Y.,Sun Yat Sen University | Zhang Y.,SYSU CMU Shunde International Joint Research Institute | Zhang Y.,Key Laboratory of Autonomous Systems and Networked Control | Qu L.,Sun Yat Sen University | And 5 more authors.
Soft Computing | Year: 2016

To solve complex problems such as multi-input function approximation by using neural networks and to overcome the inherent defects of traditional back-propagation neural networks, a single hidden-layer feed-forward sine-activated neural network, sine neural network (SNN), is proposed and investigated in this paper. Then, a double-stage weights and structure determination (DS-WASD) method, which is based on the weights direct determination method and the approximation theory of using linearly independent functions, is developed to train the proposed SNN. Such a DS-WASD method can efficiently and automatically obtain the relatively optimal SNN structure. Numerical results illustrate the validity and efficacy of the SNN model and the DS-WASD method. That is, the proposed SNN model equipped with the DS-WASD method has great performance of approximation on multi-input function data. © 2014, Springer-Verlag Berlin Heidelberg.


Mao M.,Sun Yat Sen University | Li J.,Sun Yat Sen University | Li J.,Key Laboratory of Autonomous Systems and Networked Control | Jin L.,Sun Yat Sen University | And 4 more authors.
Neurocomputing | Year: 2016

Inevitable noises and limited computational time are major issues for time-variant matrix inversion in practice. When designing a time-variant matrix inversion algorithm, it is highly demanded to suppress noises without violating the performance of real-time computation. However, most existing algorithms only consider a nominal system in the absence of noises, and may suffer from a great computational error when noises are taken into account. Some other algorithms assume that denoising has been conducted before computation, which may consume extra time and may not be suitable in practice. By considering the above situation, in this paper, an enhanced discrete-time Zhang neural network (EDTZNN) model is proposed, analyzed and investigated for time-variant matrix inversion. For comparison, an original discrete-time Zhang neural network (ODTZNN) model is presented. Note that the EDTZNN model is superior to ODTZNN model in suppressing various kinds of bias noises. Moreover, theoretical analyses show the convergence of the proposed EDTZNN model in the presence of various kinds of bias noises. In addition, numerical experiments including an application to robot motion planning are provided to substantiate the efficacy and superiority of the proposed EDTZNN model for time-variant matrix inversion. © 2016 Elsevier B.V.


Chen D.,Sun Yat Sen University | Chen D.,SYSU CMU Shunde International Joint Research Institute | Chen D.,Key Laboratory of Autonomous Systems and Networked Control | Zhang Y.,Sun Yat Sen University | And 2 more authors.
IET Control Theory and Applications | Year: 2016

In this study, a minimum jerk norm (MJN) scheme with an obstacle avoidance constraint is proposed and applied to a redundant robot arm, of which the joint jerks keep bounded for a human-friendly robot control. To achieve superior tracking performances of the redundant robot arm, the proposed jerk bounded MJN scheme is improved by the feedback control. More importantly, the effectiveness on obstacle avoidance of the proposed scheme is guaranteed by the variable-magnitude escape jerk theorem. Besides, for the purpose of implementation on the practical robot system, the corresponding discrete formulas with their theoretical analyses are presented. Then the proposed scheme is reformulated as a dynamical quadratic program which is solved by a piecewise-linear projection equation neural network. Furthermore, the path-tracking simulation and comparison substantiate the effectiveness and accuracy of such a scheme with the smooth and human-friendly joint variables applied to the obstacle avoidance of a six degrees of freedom jerk bounded robot arm. At last, the experimental application conducted on a practical redundant robot arm system further shows the physical realisability and the safety of the proposed scheme. © The Institution of Engineering and Technology 2016.

Loading Key Laboratory of Autonomous Systems and Networked Control collaborators
Loading Key Laboratory of Autonomous Systems and Networked Control collaborators