the proceedings contain 221 papers. the topics discussed include: a novel feature weighted twin-hypersphere support vector machine for pattern recognition;an EMD-RF based short-term wind power forecasting method;feedb...
ISBN:
(纸本)9781538626184
the proceedings contain 221 papers. the topics discussed include: a novel feature weighted twin-hypersphere support vector machine for pattern recognition;an EMD-RF based short-term wind power forecasting method;feedback-aided PID-type iterative learningcontrol against initial state error;quantum noise protection via weak measurement for quantum mixed states;a K-shell improved method for the importance of complex network nodes;moving object real-time detection and tracking method based on improved Gaussian mixture model;multi-objective optimization for thermal power plant operation based on improved working condition;and motor imagery ECoG signal classification using sparse representation with elastic net constraint.
Deep learning has become a crucial technology for making breakthroughs in many fields. Nevertheless, it still faces two important challenges in theoretical and applied aspects. the first lies in the shortcomings of gr...
详细信息
ISBN:
(纸本)9789819785049;9789819785056
Deep learning has become a crucial technology for making breakthroughs in many fields. Nevertheless, it still faces two important challenges in theoretical and applied aspects. the first lies in the shortcomings of gradient descent based learning schemes which are time-consuming and difficult to determine the learningcontrol hyperparameters. Next, the architectural design of the model is usually tricky. In this paper, we propose a semi-adaptive synergetic two-way pseudoinverse learning system, wherein each subsystem encompasses forward learning, backward learning, and feature concatenation modules. the whole system is trained using a non-gradient descent learning algorithm. It simplifies the hyperparameter tuning while improving the training efficiency. the architecture of the sub-systems is designed using a data-driven approach that enables automated determination of the depth of the subsystems. We compare our method withthe baselines of mainstream non-gradient descent based methods and the results demonstrate the effectiveness of our proposed method. the source code for this paper is available at http://***/B-berrypie/Semi-adaptive-Synergetic-Two-way-Pseudoinverse-learning-System.
this paper deals with iterative learningcontrol for a linear conformable fractional differential equation. A conformable D-type learning updating law is proposed to derive the convergence results for such type equati...
详细信息
ISBN:
(纸本)9781538626184
this paper deals with iterative learningcontrol for a linear conformable fractional differential equation. A conformable D-type learning updating law is proposed to derive the convergence results for such type equations varying withthe initial state is (not) coincident withthe desired initial state. Finally, two numerical examples are given to illustrate the results.
In this work, we conduct stability analysis for a class of multi-module impulsive controlsystems via an event-driven scheme. By designing some event-driven conditions and a proper event-driven impulsive control law, ...
详细信息
ISBN:
(纸本)9781538626184
In this work, we conduct stability analysis for a class of multi-module impulsive controlsystems via an event-driven scheme. By designing some event-driven conditions and a proper event-driven impulsive control law, we establish some sufficient stability criteria for the considered systems. the proposed event-drivencontrol scheme is advantageous to reduce the utilization of communication and computation resources. Further, we study the impulsive synchronization problem for two continuous-time dynamical systems with different initial values. Finally, an example of Chua's circuit with simulations results are provided to illustrate the validity of the method.
In this work, a control scheme with compensation along the iteration axis is discussed for discrete time nonlinear systems with random data loss. the loss of output data from sensor to controller is considered, and th...
详细信息
ISBN:
(纸本)9781538626184
In this work, a control scheme with compensation along the iteration axis is discussed for discrete time nonlinear systems with random data loss. the loss of output data from sensor to controller is considered, and the data missing is described through a variable satisfying the Bernoulli distribution. the lost output value is estimated by using the time-varying parameter and the output value of the last iteration to compensate the influence of data loss on the plant. A numerical simulation example verifies the validity of the algorithm.
In this paper, we propose an output feedback tracking control scheme for a class of continuous-time nonlinear systems without specific model. A radial basis function neural network (RBFNN) observer is designed to onli...
详细信息
ISBN:
(纸本)9781538626184
In this paper, we propose an output feedback tracking control scheme for a class of continuous-time nonlinear systems without specific model. A radial basis function neural network (RBFNN) observer is designed to online estimate the unmeasured inner state variables only using the input and output data. Based on the designed RBFNN observer, a sliding mode controller is derived to guarantee that the system states follow the desired trajectories. Simulation results on an example show the effectiveness and tracking performance of the proposed scheme.
In this paper, the right coprime factorization method based on operator theory is applied to deal withthe stability issue of nonlinear feedback system, wherein the inverse of the right factor obtained from the isomor...
详细信息
ISBN:
(纸本)9781538626184
In this paper, the right coprime factorization method based on operator theory is applied to deal withthe stability issue of nonlinear feedback system, wherein the inverse of the right factor obtained from the isomorphism-based factorization method is discussed and is proved to be stable, thus the Bezout identity is satisfied withthe designed controllers. Meanwhile, the nonlinear feedback system is stable.
Linearization technique is inevitable for a nonlinear control system design. However, the traditional linearization methods require model information, which is difficult to obtain for the complex nonlinear system. In ...
详细信息
ISBN:
(纸本)9781538626184
Linearization technique is inevitable for a nonlinear control system design. However, the traditional linearization methods require model information, which is difficult to obtain for the complex nonlinear system. In this article, a new local dynamic linearization method is proposed via a mean-value theorem and can be estimated by using the I/O data only. then a new adaptive iterative learningcontrol is proposed by using the optimal technology. the simulation verifies the monotonic convergence and practicability of this method.
this thesis discusses the decentralized iterative learningcontrol for large-scale discrete-time single-input single output (SISO) systems, which is interconnected by non-affine nonlinear systems. In view of the struc...
详细信息
ISBN:
(纸本)9781538626184
this thesis discusses the decentralized iterative learningcontrol for large-scale discrete-time single-input single output (SISO) systems, which is interconnected by non-affine nonlinear systems. In view of the structure of the system, the P-type learning algorithm is constructed. Under certain assumptions, the algorithm can make sure that the error precision required in each subsystem is attained through repeated iteration. the given example indicates that the proposed scheme is effective.
the note considers an iterative learningcontrol scheme for a kind of switched repetitive systems. the manipulated systems are specified by arbitrary switching signals with respective to both time variable and iterati...
详细信息
ISBN:
(纸本)9781538626184
the note considers an iterative learningcontrol scheme for a kind of switched repetitive systems. the manipulated systems are specified by arbitrary switching signals with respective to both time variable and iteration index. By employing Lebesgue-p norm, the learning performance is analyzed and a sufficient condition of convergence is derived. Results show that the concerned control law works well for tracking problem of the switched systems when the switching rules are expanded to time-iteration domain. Simulation is included to verify the validity of the approach.
暂无评论