Proposed Cost Function Research Articles

A Novel Structure of Actor-Critic Learning Based on an Interval Type-2 TSK Fuzzy Neural Network

In this article, a novel structure of actor-critic learning based on an interval type-2 Takagi–Sugeno–Kang fuzzy neural network (AC-IT2-TSK-FNN) is proposed. The proposed structure consists of two IT2-TSK-FNNs that represent the critic and the actor. Structure and parameter learnings are established for all the rules of the proposed structure. The antecedent and consequent parameters for the critic and actor are updated based on the minimization of the proposed cost function. Optimal values for the learning rates are developed and obtained to achieve stability using Lyapunov theory. The obtained results show the superiority of the proposed structure compared to other existing controllers when applied to nonlinear systems.

Improved Model Predictive Control for High-Power Current-Source Rectifiers Under Normal and Distorted Grid Conditions

This article proposes an improved model predictive control (MPC) method for high-power current-source rectifiers (CSR). Instead of introducing the virtual impedance based active damping notion into MPC framework, the predefined cost function directly takes grid-side current and capacitor voltage as the regulating objectives to suppress the $LC$ resonance and grid current distortions. The relationship and differences between the proposed cost function and its form with active damping function are analyzed in detail under normal grid conditions. Moreover, the number of discrete candidate vectors during each predictive horizon in the proposed scheme is reduced from nine to five by considering the constraint of switching states transition. Therefore the computational burden is relieved compared with conventional MPC under normal grid conditions. Finally, in order to keep sinusoidal grid current for CSR even working under severe distorted supply voltage, specific harmonic orders of grid current suppression items are added into the proposed cost function subsequently. The performance of proposed methods at low switching frequency ( $\leq$ 1 kHz) under normal and distorted grid conditions are tested through both simulation and experiments. For the implementation of all MPC methods, capacitor voltage is estimated by an observer to save hardware expense instead of using transducers.

Optimal Network Parameter Estimation: Single-Shot Exchange of Local Decisions

This letter considers a network of sensors that collectively sense a number of unknown parameters. Each sensor can possibly sense only a subset of the parameters, gather data only about these parameters, and has access to only the statistical model of the data that it collects locally. The objective is that each sensor forms optimal estimates for its designated parameters (i.e., the parameters that it can sense). This letter proposes an estimation cost function that strikes a balance between the sensors being autonomous in forming local estimates based on their locally available data and statistical models, and enforcing consistency among the local estimates formed for the parameters that are sensed by multiple sensors. Exact optimal estimators are characterized, and it is shown that the optimal estimators can be implemented in a distributed way, through a single-shot exchange of local decisions. Specifically, the distributed implementation consists of forming local estimates and exchanging certain sufficient statistics values in a single round of communication exchange among some of the sensors. Furthermore, the optimal performance under the proposed cost function is also compared analytically with the performance of the widely used mean squared error estimator.

Cascade-Free Fuzzy Finite-Control-Set Model Predictive Control for Nested Neutral Point-Clamped Converters With Low Switching Frequency

In this brief, a cascade-free fuzzy finite-control-set model predictive control (FCS-MPC) is proposed for nested neutral point-clamped converters with low switching frequency (SF). The main objective of the proposed method is to achieve a low SF operation. First, a new cost function is formulated to reduce the SF, and a fuzzy logic control (FLC) technique is employed to choose the weighting factors dynamically. As such, the FLC scheme has a good potential to avoid the constant of weighting factor for the conventional FCS-MPC strategy, especially in the presence of dynamic operation condition. Second, in order to enhance the dynamic response and improve the computationally efficient under low SF, a dynamic reference design with a simplified cascade-free strategy is embedded into the predictive controller. The novelty of this brief lies not only in combining the proposed cost function with the FLC technique for online determining the appropriate values of the weighting factors, but also in a compatible reference design that directly allows formulating the optimal control problem using a cascade-free structure. Finally, simulation results are presented for assessing the effectiveness of the proposed control strategy.

State-of-charge-constraint-based energy management strategy of plug-in hybrid electric vehicle with bus route

For the plug-in hybrid electric vehicle (PHEV) with repeated routes, the route information is very useful for energy optimization. SOC (state of charge) of power battery is the most important parameter for PHEV, so the information is usually utilized to calculate reference SOC. However, it is not the optimal trajectory, which might restrict the improvement of economic performance. In this paper, the route information is processed by dynamic programming and a group of optimal state of charge trajectories are obtained. Then, a novel state-of-charge-constraint-based cost function is firstly proposed, which can be obtained by presented method based on history cycle data. Next, a nonlinear model predictive controller is designed as energy management strategy to evaluate different cost functions. Experiment results show that proposed cost function can save fuel consumption by 5.9% and 10.8% compared with other two common cost functions.

Automatic targetless camera–LIDAR calibration by aligning edge with Gaussian mixture model

AbstractThis paper presents a calibration algorithm that does not require an artificial target object to precisely estimate a rigid‐body transformation between a camera and a light detection and ranging (LIDAR) sensor. The proposed algorithm estimates calibration parameters by minimizing a cost function that evaluates the edge alignment between two sensor measurements. In particular, the proposed cost function is constructed using a projection model‐based many‐to‐many correspondence of the edges to fully exploit measurements with different densities (dense photometry and sparse geometry). The alignment of the many‐to‐many correspondence is represented using the Gaussian mixture model (GMM) framework. Here, each component of the GMM, including weight, displacement, and standard deviation, is derived to suitably capture the intensity, location, and influential range of the edge measurements, respectively. The derived cost function is optimized by the gradient descent method with an analytical derivative. A coarse‐to‐fine scheme is also applied by gradually decreasing the standard deviation of the GMM to enhance the robustness of the algorithm. Extensive indoor and outdoor experiments validate the claim that the proposed GMM strategy improves the performance of the proposed algorithm. The experimental results also show that the proposed algorithm outperforms previous methods in terms of precision and accuracy by providing calibration parameters of standard deviations less than 0.6° and 2.1 cm with a reprojection error of 1.78 for a 2.1‐megapixel image (2,048 × 1,024) in the best case.

Cost-Function-Based Microgrid Decentralized Control of Unbalance and Harmonics for Simultaneous Bus Voltage Compensation and Current Sharing

Power quality of inverter-based microgrids is a challenging issue due to nonlinearities of inverters, multiple resonance modes of network impedance, and unbalanced and nonlinear loading condition. The ideal solution is to assure the best power quality at the common bus to which loads are connected, and share the negative and harmonic sequence currents between the inverters without any communication. However, it is difficult to achieve this objective with existing virtual-impedance-based methods when nonlinearities of inverters, e.g., dead time, are not negligible. In this paper, a novel cost-function-based method is proposed to solve this problem. The proposed cost functions are optimized using continuous-control-set model predictive control. As unknown nonlinearities of inverters can be observed by introducing disturbance models, even when they are not negligible, the presented method can compensate bus voltage unbalance and harmonics and share the compensated current between inverters autonomously, with neither communication nor upper level controller. Parameter tuning of the proposed method is the key to achieve the balance between bus voltage quality and current sharing, and between stability and fast response. Experimental results demonstrate that the proposed control scheme is effective even under conditions of large inverter nonlinearities.

MC-SSM: Nonparametric Semantic Image Segmentation With the ICM Algorithm

In the last few years, there has been considerable interest in scene parsing. This task consists of assigning a predefined class label to each pixel (or pre-segmented region) in an image. To best address the complexity challenge of this task, first, we propose a new geometric retrieval strategy to select nearest neighbors from a database containing fully segmented and annotated images. Then, we introduce a novel and simple energy-minimization model. The proposed cost function of this model combines efficiently different global nonparametric semantic likelihood energy terms. These terms are computed from the (pre-)segmented regions of the (query) image and their structural properties (location, texture, color, context, and shape). Different from the traditional approaches, we use a simple and local optimization procedure derived from the iterative conditional modes algorithm to optimize our energy-based model. Experimental results on two challenging datasets: 1) microsoft research Cambridge dataset and 2) Stanford background dataset demonstrate the feasibility and the success of the proposed approach. Compared to existing annotation methods that require training classifiers for each object and learning many parameters, our method is easy to implement, has a few parameters, and combines different criteria.

Modeling and predictive control of InGap/GaAs/Ge triple-junction solar cells to increase the energy conversion efficiency

Physical studies of the last decade indicate that multi-junction solar cell has higher energy conversion efficiency than single-junction cells. However, choosing the type of material, modeling and finding the parameters of this type of cell has always been one of the leading challenges in this topic. Most of the proposed models are assumed to have predetermined electrical parameters of the cell or regardless of the effect of the tunnel junction, which reduces of the system. This paper discusses the modeling of solar cell triplejunction InGap/GaAs /Ge, taking into account the effect of a tunnel junction and finds the parameters of each subcellular deal. Also, predictive control will be used to control the active and reactive power of the single-phase inverter. The use of this method eliminates the need for the modulation module and the phase-lock loop (PLL). And simplifies the control algorithm for its digital implementation. The proposed cost function of this paper will be such that in addition to controlling the power of the inverter, the closed loop stability will be guaranteed based on the Lyapunov theory. Finally, the performance of the system using the software MATLAB/SIMULINK simulation will be evaluated.

Modeling and predictive control of InGap/GaAs/Ge triple-junction solar cells to increase the energy conversion efficiency

Physical studies of the last decade indicate that multi-junction solar cell has higher energy conversion efficiency than single-junction cells. However, choosing the type of material, modeling and finding the parameters of this type of cell has always been one of the leading challenges in this topic. Most of the proposed models are assumed to have predetermined electrical parameters of the cell or regardless of the effect of the tunnel junction, which reduces of the system. This paper discusses the modeling of solar cell triplejunction InGap/GaAs /Ge, taking into account the effect of a tunnel junction and finds the parameters of each subcellular deal. Also, predictive control will be used to control the active and reactive power of the single-phase inverter. The use of this method eliminates the need for the modulation module and the phase-lock loop (PLL). And simplifies the control algorithm for its digital implementation. The proposed cost function of this paper will be such that in addition to controlling the power of the inverter, the closed loop stability will be guaranteed based on the Lyapunov theory. Finally, the performance of the system using the software MATLAB/SIMULINK simulation will be evaluated.

Model Predictive Control for Dual-Active-Bridge Converters Supplying Pulsed Power Loads in Naval DC Micro-Grids

Pulsed power loads (PPLs) are becoming prevalent in medium-voltage naval dc micro-grids. To alleviate their effects on the system, energy storages are commonly installed. For optimal performance, their interface converters need to have fast dynamics and excellent disturbance rejection capability. Moreover, these converters often need to have voltage transformation and galvanic isolation capability since common energy storage technologies such as batteries and supercaps are typically assembled with low-voltage strings. In order to address these issues, a moving discretized control set model predictive control (MDCS-MPC) is proposed in this paper and applied on a dual-active-bridge converter. Fixed switching frequency is maintained, enabling easy passive components design. The proposed MDCS-MPC has a reduced prediction horizon, which allows low computational burden. The operating principle of the MDCS-MPC is introduced in the development of a cost function, which provides stiff voltage regulation. Resonance damping and sampling noise resistance can also be achieved with the proposed cost function. An adaptive step is introduced to enable a fast transition. Assessments on the performance of the proposed MDCS-MPC are conducted. Comparisons with other control methods are also provided. Experimental validations on a 300 V/300 V 20-kHz 1-kW dual-active-bridge converter are carried out to verify the theoretical claims.

Cooperative path planning for leader – follower formation of Multi UAV based on the minimum energy consumption for load transportation

Today, unmanned aerial vehicles are highly considered for both military and commercial fields due to low cost, high maneuverability, and good survival. One of the most important design challenges of multi-UAV systems is mission planning. In the broad class of unmanned aerial vehicles (UAV), Quadrotor is an important member. The capabilities of this vehicle in load transportation has attracted the attention of many research groups around the world. In this paper, path planning based on the minimum energy consumption is studied for load transportation. The purpose of the present study is to investigate the effect of proposed cost function in order to obtain optimal path for transporting loads based on reducing the energy consumption of quadrotors. The results indicate the 35.29% energy consumption of multi UAV compared to the energy consumption of an individual quadrotor. This also leads to an increase in load carrying capacity. On the other hand, the leader- follower formation is preserved until the end of the path based on the defined relationships. The simulation results illustrate the power and efficiency of the method to overcome the high nonlinearity nature of the problem such as path optimization of multi-rotor helicopters.

Improving the performance of a class of adaptive fuzzy controller based on stable and fast on-line learning algorithm

In this paper, a novel stable and fast nonlinear learning scheme called adaptive Takagi–Sugeno–Kang fuzzy controller using a reinforcement actor-critic learning (ATSKFC-RACL) is proposed. In the adopted ATSKFC-RACL approach, the actor is represented by the TSK fuzzy controller, which the parameters are updated on-line according to the proposed cost function and the critic output. The proposed cost function relies on the change of the reward function, the change of the actor parameters, the actor error and the change of the actor error, which gives the feedback from the external environment to the critic network to facilities the learning through its own faults and accelerates it. The critic network is implemented by a neural network, which the parameters are updated on-line using the novel laws that depend on Lyapunov criteria to avoid the shortcoming of the gradient descent such as sucking in local minima or instability. Furthermore, Lyapunov stability analysis is studied for choosing the conditions of the learning rates to guarantee the stability of the proposed ATSKFC-RACL. The learning of the proposed ATSKFC-RACL scheme is faster than the TSK fuzzy controller due to the critic network and the proposed cost function. Furthermore, the proposed scheme can reduce the effect of the system uncertainties and disturbances. The proposed scheme is applied for controlling nonlinear systems. The results show that the performance of the proposed ATSKFC-RACL approach is a stable and fast compared with other existing techniques.

Direct Predictive Speed Control With a Sliding Manifold Term for PMSM Drives

This paper proposes a new direct predictive speed control (DPSC) method for permanent-magnet synchronous machines (PMSMs), which eliminates the cascaded loops structure that is usually used in the conventional controller. A novel sliding manifold term is added to the cost function to realize the speed/current tracking. The proposed cost function is simple and easy to implement. Therefore, in comparison with the other DPSC methods, the proposed method does not need long prediction horizons, dynamically adjustable weighting factors or large weighting factors for the control variables with large time constants. The performance of the proposed method has been experimentally investigated and compared with: 1) conventional DPSC scheme and 2) a predictive current control with an outer PI controller (PI-PCC). The results demonstrated enhancement in the dynamic performance in comparison with the PI-PCC method and improvement in the steady-state response in comparison with the conventional DPSC technique.

Kernel Principal Component Analysis Allowing Sparse Representation and Sample Selection

Kernel principal component analysis (KPCA) is a kernelized version of principal component analysis (PCA). A kernel principal component is a superposition of kernel functions. Due to the number of kernel functions equals the number of samples, each component is not a sparse representation. Our purpose is to sparsify coefficients expressing in linear combination of kernel functions, two types of sparse kernel principal component are proposed in this paper. The method for solving sparse problem comprises two steps: (a) we start with the Pythagorean theorem and derive an explicit regression expression of KPCA and (b) two types of regularization $l_1$-norm or $l_{2,1}$-norm are added into the regression expression in order to obtain two different sparsity form, respectively. As the proposed objective function is different from elastic net-based sparse PCA (SPCA), the SPCA method cannot be directly applied to the proposed cost function. We show that the sparse representations are obtained in its iterative optimization by conducting an alternating direction method of multipliers. Experiments on toy examples and real data confirm the performance and effectiveness of the proposed method.

A multi-objective optimization problem in mixed and natural convection for a vertical channel asymmetrically heated

This paper deals with a multi-objective topology optimization problem in an asymmetrically heated channel, based on both pressure drop minimization and heat transfer maximization. The problem is modeled by assuming steady-state laminar natural convection f low. The incompressible Navier-Stokes equations coupled with the convection-diffusion equation, under the Boussinesq approximation, are employed and are solved with the finite volume method. In this paper, we discuss some limits of classical pressure drop cost function for buoyancy driven flow and, we then propose two new expressions of objective functions: the first one takes into account work of pressure forces and contributes to the loss of mechanical power while the second one is related to thermal power and is linked to the maximization of heat exchanges. We use the adjoint method to compute the gradient of the cost functions. The topology optimization problem is first solved for a Richardson (Ri) number and Reynolds number (Re) set respectively to Ri ∈ {100,200,400} and Re = 400. All these configuration are investigated next in order to demonstrate the efficiency of the new expressions of cost functions. We compare two types of interpolation functions for both the design variable field and the effective diffusivity. Both interpolation techniques have pros and cons and give slightly the same results. We notice that we obtain less isolated solid elements with the sigmoid type interpolation functions.Then, we choose to work with the sigmoid and solve the topology optimization problem in case of pure natural convection, by setting Rayleigh number to {3×103,4×104,5×105}. In all considered cases, our algorithm succeeds to enhance one of the phenomenon modeled by the proposed cost functions without deteriorating the other one. The optimized design obtained suppresses any reversal flow at the exit of the channel. We also show that the thermal exchanges are improved by computing the Nusselt numbers and bulk temperature. We conclude that the new expressions of objective functions are well suited to deal with natural convection optimization problem in a vertical channel.

Model Predictive Direct Power Control With Fixed Switching Frequency and Computational Amount Reduction

The model predictive direct power control (MPDPC) has attracted significant attention due to its outstanding dynamic response and high power factor. However, the variable switching frequency makes the design of alternating current (ac) filter more challenging, and the heavy computational burden limits the application of MPDPC. This paper proposed a new cost function and four steps’ MPDPC (FSMPDPC) scheme for T-type inverters. The proposed cost function can reduce the number of division operations and does not require calculating the duty cycles of all vectors. Meanwhile, the four steps’ calculation process is divided into four steps to reduce the number of cycle calculations. The first three steps are assigned to adjust the active and reactive powers, and the neutral-point (NP) voltage is balanced in the fourth step. An experimental platform of a T-type inverter is established to demonstrate the superiorities of the proposed FSMPDPC. The results show that FSMPDPC improves the steady-state performance of the T-type inverter with lower current total harmonic distortion (THD) and lower ripples in the active and reactive powers. In addition, the proposed algorithm eases the computational burden of the digital signal processor (DSP).

Clustered multitask non-negative matrix factorization for spectral unmixing of hyperspectral data

The algorithm based on a clustered multitask network is proposed to solve spectral unmixing problem in hyperspectral imagery. In the proposed algorithm, the clustered network is employed. Each pixel in the hyperspectral image is considered as a node in this network. The nodes in the network are clustered using the fuzzy c-means clustering method. Diffusion least mean square strategy has been used to optimize the proposed cost function. To evaluate the proposed method, experiments are conducted on synthetic and real datasets. Simulation results based on spectral angle distance, abundance angle distance, and reconstruction error metrics illustrate the advantage of the proposed algorithm, compared with other methods.

An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson–Gaussian noise

Much research has been devoted to the problem of low-dose computerized tomography (CT) image reconstruction with a scan protocol by both lowering the X-ray tube current (low-mAs) and reducing the total number of projection views(sparse view). However, the CT transmission data may be severely corrupted by X-ray quanta noise and system electronic noise. Recently, a nonlocal means (NLM) regularization method for low-dose CT reconstruction was proposed. Although this method is effective in suppressing both Poisson–Gaussian noise and artifacts, it has two disadvantages: the heavy computational burden and blurred edge information in CT reconstruction. This paper proposes an adaptive patch-wise regularization method for low-dose CT reconstruction from available CT transmission data in Poisson–Gaussian noise. The proposed cost function includes a penalized weighted least-square term and two patch-wise regularization terms which are combined with a novel adaptive regularization parameter. The two regularization terms take advantage of image redundant information across different scales. By exploiting both global and local structure information, the reconstruction accuracy is enhanced. In addition, we select the adaptive regularization parameter based on the texture of the image patch so that the edge information is further maintained. Furthermore, our algorithm updates a patch rather than a pixel, the computation burden is greatly reduced. Finally, experiment results show that the proposed adaptive patch-wise regularization method for low-dose CT reconstruction is superior to several conventional regularization-based approaches in terms of computation efficiency and resolution quality.

New insights on the optimality of parameterized Wiener filters for speech enhancement applications

This work presents a unified framework for defining a family of noise reduction techniques for speech enhancement applications. The proposed approach provides a unique theoretical foundation for some widely-applied soft and hard time-frequency masks, which encompasses the well-known Wiener filter and the heuristically-designed Binary mask. These techniques can now be considered as optimal solutions of the same minimization problem. The proposed cost function is defined by two design parameters that not only establish a desired trade-off between noise reduction and speech distortion, but also provide an insightful relationship with the mask morphology. Such characteristic may be useful for applications that require online adaptation of the suppression function according to variations of the acoustic scenario. Simulation examples indicate that the derived conformable suppression mask has approximately the same quality and intelligibility performance capability of the classical heuristically-defined parametric Wiener filter. The proposed approach may be of special interest for real-time embedded speech enhancement applications such as hearing aids and cochlear implants.