Direct torque control (DTC) suffers from the large ripples of the torque and flux, which leads to deteriorating the system performances. Replacing the conventional back-to-back converter with an indirect matrix conver...
详细信息
Direct torque control (DTC) suffers from the large ripples of the torque and flux, which leads to deteriorating the system performances. Replacing the conventional back-to-back converter with an indirect matrix converter (IMC) can reduce the torque and flux ripples in interior dual stator induction motor. However, the ripples can be still large, and the low harmonic distortion appears in the input current caused by the lack of synchronization between the two stages of the IMC. In order to overcome this problem with improved dynamic performances (torque, flux ripples, and reducing the low harmonic distortion in the input current), a new DTC based on space vector modulation (SVM) with the synchronization between the two stage of IMC is proposed. On the other hand, an adaptive PI controller based on a gradient descent algorithm using metaheuristic Bat algorithm is presented to generate: the desired stator voltages, and the switching states of the inverter stage by the SVM. Furthermore, the control scheme performance is enhanced by inserting a robust synergetic controller (SC) in the outer loop for speed regulation. The obtained simulation results under different operating conditions illustrate the benefit of the proposed control technique, also, it demonstrates the feasibility of the proposed control approach for real-world systems.
Field Programmable Gate Arrays (FPGA's) are being deployed in cloud data centers like Amazon Elastic compute cloud F1 instances for algorithmic acceleration to leverage the advantage of reconfigurability combined ...
详细信息
ISBN:
(纸本)9781728193694
Field Programmable Gate Arrays (FPGA's) are being deployed in cloud data centers like Amazon Elastic compute cloud F1 instances for algorithmic acceleration to leverage the advantage of reconfigurability combined with high throughput and low power consumption of FPGA's and hence balancing the dynamic workloads. gradientdescent is one such algorithm that is extensively used in core computation kernels to train the Machine Learning (ML) models. This paper proposes a custom-IP (Intellectual Property) for hardware acceleration of gradient descent algorithm (GDA) which is designed by exploring the inherent concurrency of GDA. The IP incorporates the Very High-Speed Integrated Circuit Hardware Description Language (VHDL) based description of GDA and AXI4 stream interface for connectivity. The Custom-IP is flexible and reusable by tuning generics defined in the VHDL code. The IP is interfaced to a 32-bit MicroBlaze soft-core processor which acts as host and manages run-time along with other peripherals to form a System on Chip (SoC) in which the design is partitioned into fixed hardware and flexible software. The Hardware/Software Co-design results show the 5x improvement in performance when implemented on Artix XC7A100T-CSG324 FPGA when compared to software implementation.
This paper is concerned with the secure state estimation problem for a class of cyber-physical systems(CPSs) with multiple sensors equipped with a decentralized event-triggering(ET) *** with the traditional secure sta...
详细信息
ISBN:
(数字)9789887581536
ISBN:
(纸本)9781665482561
This paper is concerned with the secure state estimation problem for a class of cyber-physical systems(CPSs) with multiple sensors equipped with a decentralized event-triggering(ET) *** with the traditional secure state estimators,the introduction of ET mechanism mitigates the unnecessary data transmission and brings more useful information for secure state *** make full use of the introduced event-triggering conditions which are not corrupted by the adversary,a novel data consolidation algorithm is proposed to consolidate the collected ***,a necessary and sufficient condition for the robust observability of the CPSs equipped with ET mechanism under sparse sensor attacks is provided,and a switched gradient descent algorithm is proposed to estimate the system state from the consolidated *** is shown that through consolidating the collected information,introducing ET mechanism helps improve the resilience against sparse sensor ***,a numerical simulation is provided to illustrate the effectiveness of the proposed algorithms.
We study the generalization ability of distributed learning equipped with a divide-and-conquer approach and gradient descent algorithm in a reproducing kernel Hilbert space (RKHS). Using special spectral features of t...
详细信息
We study the generalization ability of distributed learning equipped with a divide-and-conquer approach and gradient descent algorithm in a reproducing kernel Hilbert space (RKHS). Using special spectral features of the gradient descent algorithms and a novel integral operator approach, we provide optimal learning rates of distributed gradient descent algorithms in probability and partly conquer the saturation phenomenon in the literature in the sense that the maximum number of local machines to guarantee the optimal learning rates does not vary if the regularity of the regression function goes beyond a certain quantity. We also find that additional unlabeled data can help relax the restriction on the number of local machines in distributed learning.
作者：
Liu, YangUniv Maryland
Dept Human Dev & Quantitat Methodol 12308 Benjamin Bldg3942 Campus Dr College Pk MD 20742 USA
In exploratory factor analysis, latent factors and factor loadings are seldom interpretable until analytic rotation is performed. Typically, the rotation problem is solved by numerically searching for an element in th...
详细信息
In exploratory factor analysis, latent factors and factor loadings are seldom interpretable until analytic rotation is performed. Typically, the rotation problem is solved by numerically searching for an element in the manifold of orthogonal or oblique rotation matrices such that the rotated factor loadings minimize a pre-specified complexity function. The widely used gradient projection (GP) algorithm, although simple to program and able to deal with both orthogonal and oblique rotation, is found to suffer from slow convergence when the number of manifest variables and/or the number of latent factors is large. The present work examines the effectiveness of two Riemannian second-order algorithms, which respectively generalize the well-established truncated Newton and trust-region strategies for unconstrained optimization in Euclidean spaces, in solving the rotation problem. When approaching a local minimum, the second-order algorithms usually converge superlinearly or even quadratically, better than first-order algorithms that only converge linearly. It is further observed in Monte Carlo studies that, compared to the GP algorithm, the Riemannian truncated Newton and trust-region algorithms require not only much fewer iterations but also much less processing time to meet the same convergence criterion, especially in the case of oblique rotation.
We developed a methodology for the neural network boosting of logistic regression aimed at learning an additional model structure from the data. In particular, we constructed two classes of neural network-based models...
详细信息
We developed a methodology for the neural network boosting of logistic regression aimed at learning an additional model structure from the data. In particular, we constructed two classes of neural network-based models: shallow-dense neural networks with one hidden layer and deep neural networks with multiple hidden layers. Furthermore, several advanced approaches were explored, including the combined actuarial neural network approach, embeddings and transfer learning. The model training was achieved by minimizing either the deviance or the cross-entropy loss functions, leading to fourteen neural network-based models in total. For illustrative purposes, logistic regression and the alternative neural network-based models we propose are employed for a binary classification exercise concerning the occurrence of at least one claim in a French motor third-party insurance portfolio. Finally, the model interpretability issue was addressed via the local interpretable model-agnostic explanations approach.
The regression analysis of large data sets by parallelization is investigated in the work. Different approaches to parallel computing are proposed. The parallelization of the gradientdescent method and the stochastic...
详细信息
The regression analysis of large data sets by parallelization is investigated in the work. Different approaches to parallel computing are proposed. The parallelization of the gradientdescent method and the stochastic gradientdescent using OpenMP and MPI technologies, as well as their hybrid, is performed. The efficiency of the proposed parallel algorithms is analyzed. A number of numerical experiments were performed. Acceleration of 5 times was achieved using the six-core architecture of a personal computer. By varying the number of threads and processor cores, the possibility of further optimization of the computational process is established. The problem of erroneous exchange is considered, which often has a negative effect on acceleration during parallelization using OpenMP. The advantages and disadvantages of using each of the technologies are analyzed. Without reducing the generality, the software product is developed in the work, which can be used in the regression analysis of big data processing in a wide range of economic problems.
In this paper, an "Adaptive Receding Horizon Controller (ARHC)" is exemplified in the suboptimal control of a Furuta pendulum. A dynamic model of strongly overestimated inertia and friction parameters is use...
详细信息
ISBN:
(纸本)9781665444996
In this paper, an "Adaptive Receding Horizon Controller (ARHC)" is exemplified in the suboptimal control of a Furuta pendulum. A dynamic model of strongly overestimated inertia and friction parameters is used in an RHC controller to track the nominal trajectory under cost terms penalizing the control forces. The so obtained "optimized" trajectory is tracked by an adaptive controller that uses a realistic approximate dynamic model of the controlled system. Since the approximate and the actual model contain considerably smaller inertia and friction parameters than that used for optimization the cautiously optimized trajectory can be precisely tracked by the actual system without suffering from heavy force burdens. The adaptivity is guaranteed by a "Fixed Point Iteration"-based approach that in this manner easily can be combined with the mathematical framework of optimal controllers. Instead of using Lagrange multipliers, the optimization happens through explicitly applying the dynamic model in forward Euler integration. The operation of the method is exemplified via numerical simulations.
作者：
Rizk, HananUniv Grenoble Alpes
GIPSA Lab 11 Rue Math F-38400 St Martin Dheres France ERI
Photovolta Dept PV Joseph Tito St Cairo 12622 Egypt
Our work is devoted to optimal control of a double-pipe heat exchanger system modeled as coupled hyperbolic partial differential equations (PDEs) of first order in time and space based on adjoint method-calculus of va...
详细信息
ISBN:
(纸本)9781665427494
Our work is devoted to optimal control of a double-pipe heat exchanger system modeled as coupled hyperbolic partial differential equations (PDEs) of first order in time and space based on adjoint method-calculus of variations, which should be associated with numerical techniques to get solutions. A gradient descent algorithm, which is proved to be convergent after few iterations, is applied to solve the optimal control problem of this heat exchanger (HE) system. Based on the real data in a system, the simulation results demonstrate the effectiveness of the proposed control approach to control the temperatures along the pipes of the HE system, and that the safety of both system and performance is ensured. From these results, we concluded that the gradient descent algorithm is convergent and the method is effective in the optimal control problem of coupled hyperbolic PDEs.
In this paper, an optimal state feedback controller is designed for a nonaffine nonlinear system. The nonaffine system is converted into affine system using mean value theorem, since the analysis of affine nonlinear s...
详细信息
ISBN:
(纸本)9781665425360
In this paper, an optimal state feedback controller is designed for a nonaffine nonlinear system. The nonaffine system is converted into affine system using mean value theorem, since the analysis of affine nonlinear system is easy compared to nonaffine systems. Feedback linearization is used for linearizing the affine system. Then a gradient descent algorithm-based optimization approach is designed for feedback linearized system for optimal performance. The performance of the proposed method is presented by conducting simulation study on magnetic levitation system. A significant reduction in the minimum value of cost function is obtained using proposed method compared with the conventional LQR technique and PSO method.
暂无评论