Letteratura scientifica selezionata sul tema "SMOOTH LEAST MEAN SQUARE (SLMS)"

SUMMARYThere is increasing interest in using rehabilitation robots to assist post-stroke patients during rehabilitation therapy. The motion control of the robot plays an important role in the process of functional recovery training. Due to the change of the arm impedance of the post-stroke patient in the passive recovery training, the conventional motion control based on a proportional-integral (PI) controller is difficult to produce smooth movement of the robot to track the designed trajectory set by the rehabilitation therapist. In this paper, we model the dynamics of post-stroke patient arm as an impedance model, and propose an adaptive control scheme, which consists of an adaptive PI control algorithm and an adaptive damping control algorithm, to control the rehabilitation robot moving along predefined trajectories stably and smoothly. An equivalent two-port circuit of the rehabilitation robot and human arm is built, and the passivity theory of circuits is used to analyze the stability and smoothness performance of the robot. A slide Least Mean Square with adaptive window (SLMS-AW) method is presented for on-line estimation of the parameters of the arm impedance model, which is used for adjusting the gains of the PI-damping controller. In this paper, the Barrett WAM Arm manipulator is used as the main hardware platform for the functional recovery training of the post-stroke patient. Passive recovery training has been implemented on the WAM Arm, and the experimental results demonstrate the effectiveness and potential of the proposed adaptive control strategies.

Beamforming and side lobe level reduction of smart antenna are important tasks in mobile network. Adaptive signal processing algorithms are used for beam forming in smart antenna. In this paper, variable step-size sign least mean square (VS-SLMS) algorithm is used for beam forming of smart antenna with linear antenna array. The results are compared with the results obtained using sign least mean square (SLMS) algorithm. Variable step-size algorithm shows good results for beam forming compared to ordinary constant step-size algorithm.

Presently, the technology development trend of active noise control (ANC) systems is focused on implementing advanced adaptive filters in resource-constrained electronic appliances. Recently, several authors have proved that the use of two adaptive filter algorithms significantly improves the overall adaptive filter performance. However, the computational cost of these approaches is significantly increased since they use two filters simultaneously. Consequently, these filters cannot be implemented in these devices. To solve this problem, we propose a new ANC structure with switching selection based on filtered-x normalized least mean square (FxNLMS) and filtered-x sign least mean square (FxSLMS) algorithms to reduce the computational cost of the ANC system. The improvement of this factor has allowed us to introduce for the first time an advanced spike-based architecture, which can perform dual filter operations using dynamic routing, to be used in real ANC applications. The results have demonstrated that the computational cost of the proposed dual D-FxNLMS/SLMS algorithm is lower compared with previously reported solutions.

Features analysis is an important task which can significantly affect the performance of automatic bacteria colony picking. Unstructured environments also affect the automatic colony screening. This paper presents a novel approach for adaptive colony segmentation in unstructured environments by treating the detected peaks of intensity histograms as a morphological feature of images. In order to avoid disturbing peaks, an entropy based mean shift filter is introduced to smooth images as a preprocessing step. The relevance and importance of these features can be determined in an improved support vector machine classifier using unascertained least square estimation. Experimental results show that the proposed unascertained least square support vector machine (ULSSVM) has better recognition accuracy than the other state-of-the-art techniques, and its training process takes less time than most of the traditional approaches presented in this paper.

This article discusses the Elliptical Orbits Mode (EOM) as a method of approximating the function of changing the volume of fuel in the Underground Yank (UT). This research was conducted at the 45.507.21 Candirejo Tuntang Pertamina Gas Station. The calculation of the approximation method will be applied to the measuring book data from the Semarang Metrology Regency specifically for the Pertalite (Fuel Product of Pertamina) buried tank, because the calculation of the gas station is not smooth, it is necessary for a smoother data fitting by considering Residual Square Error (RSS) and Mean Square Error (MSE). The result of this research is the application of EOM(θ) measuring book with elliptical height control produces smaller RSS and MSE compared to using COM, EOM, Least Square degree two and three.

Turbulence measurements for rough-wall boundary layers are presented and compared to those for a smooth wall. The rough-wall experiments were made on a three-dimensional rough surface geometrically similar to the honed pipe roughness used by Shockling, Allen & Smits (J. Fluid Mech. vol. 564, 2006, p. 267). The present work covers a wide Reynolds-number range (Reθ = 2180–27 100), spanning the hydraulically smooth to the fully rough flow regimes for a single surface, while maintaining a roughness height that is a small fraction of the boundary-layer thickness. In this investigation, the root-mean-square roughness height was at least three orders of magnitude smaller than the boundary-layer thickness, and the Kármán number (δ+), typifying the ratio of the largest to the smallest turbulent scales in the flow, was as high as 10100. The mean velocity profiles for the rough and smooth walls show remarkable similarity in the outer layer using velocity-defect scaling. The Reynolds stresses and higher-order turbulence statistics also show excellent agreement in the outer layer. The results lend strong support to the concept of outer layer similarity for rough walls in which there is a large separation between the roughness length scale and the largest turbulence scales in the flow.

This paper combines the piecewise Cubic Hermite (CH) interpolation algorithm and the weighted least square support vector machine (WLS-SVM) to improve identification accuracy for marine crafts built based on the characteristic model. The characteristic model is first used to describe the heading dynamics of marine crafts and is a superior model to the traditional response model in both accuracy and complexity. Especially in order to improve identification accuracy, a CH-based data preprocessing strategy is utilized to densify and smooth data for further accurate identification. Subsequently, the combination of the linear kernel function and the Gaussian kernel function is introduced in the conventional WLS-SVM method, which renders global and local performance improvements compared with the conventional WLS-SVM method. Finally, informative maneuvers composed of Zigzag and Sine are carried out to test the performance of the improved identification method. Compared to the conventional LS-SVM method based on the response model, the root mean square error of the proposed CH-MK-WLS-SVM method based on the characteristic model is reduced by an order of magnitude in the presence of sensor noise.

Methods suitable for the determination and classification of green timber mix (western hemlock and amabilis fir), with respect to species and moisture content, were developed and tested using near infrared spectroscopy and chemometrics. One thousand two hundred samples were distributed into a calibration set (720 samples) and a prediction set (480 samples). Partial least squares (PLS) and least squares-support vector machines (LS-SVM) for both regression (PLSR and LS-SVR) and classification (PLS-DA and LS-SVC) with different spectral preprocessing methods were implemented. LS-SVM outperformed PLS models for both regression and classification. The coefficient of determination (R2p) and root mean square error (RMSEP) of prediction for the best LS-SVR model with spectra pretreated by smooth and first derivative were 0.9824 and 8.7%, respectively, for wood moisture content prediction in the range of 30% to 253%. The best classification model was LS-SVC with spectra pretreated by smooth and second derivative, with overall accuracies of 99.8% in the prediction set, when the samples were divided into four classes. NIRS combined with LS-SVM can be used as a rapid alternative method for qualitative and quantitative analysis of green hem-fir mix before kiln drying. The results could be helpful for sorting green hem-fir mixes with an on-line application.

This paper shows the operation of the microgrid using a new adaptive PI controller based operational (control) scheme. The core of the proposed control scheme is the suggested adaptive PI controller. The parameters of the PI controller are adaptively tuned using a variable step-size least mean fourth algorithm with no need for any system model to operate this adaptive controller. The main merit of the proposed scheme is that it stabilizes the magnitude and frequency of the voltage at any loading condition such as variable balanced loads, variable unbalanced loads, and nonlinear loads. The proposed scheme has a simple structure and accurate performance. In addition, the structure of proposed scheme provides a seamless transition toward any loss or reconnection of any inverter in the microgrid. Furthermore, the suggested operational scheme is flexible enough to enable the microgrid to be operative in a grid-connected mode and to transfer from the voltage control mode to power control mode with a smooth transitional procedure. To validate the meritorious performance of the suggested scheme, its performance is compared to similar schemes based on a linear controller (regular PI controller), single-neuron PI controller (adaptive PI controller), recursive least square-support vector machine based PI controller (another adaptive PI controller), and nonlinear controller (sliding mode controller) for different operations of the microgrid.

With the depletion of non-renewable resources and growing public awareness about the advantages of green energy, alternative renewable sources are evolving as a significant source of energy since past few years. Furthermore, the electrical grid is on the verge of a paradigm shift, from centralized power generation, transmission, and huge power grids towards distributed generation (DG). DG fundamentally uses small-scale generators like photovoltaic (PV) panels, wind turbine, fuel cells, small and micro hydropower, diesel generator set, etc., and is limited to small distribution networks to produce power close to the end users. Renewable energy sources (RES) are essential components of DG because they are more environment friendly than conventional power generators and once established maintenance cost is also low. One of the most popular renewable energy source is solar energy because it is abundant, accessible and can be easily converted into electricity. The electricity produced from SPV system can be utilized by the local loads within the microgrid or it can be integrated with conventional grid. Microgrid (MG), which is a cluster of distributed generation, renewable sources, and local loads connected to the utility grid provides solution to manage local generations and loads as a single grid level entity. It has the potential to maximize overall system efficiency, power quality, and energy surety for critical loads. A microgrid can operate either in stand-alone mode or grid connected mode. Due to abundant availability of solar energy, an SPV based microgrid is widely used around the world. Due to intermittent nature of solar energy, stand-alone SPV based microgrid needs an energy storage system also, whereas in grid connected system, the microgrid is connected to conventional grid which takes care of the solar intermittency by having bi-directional flow of power. Depending on the technical specifications, grid-connected solar PV- based microgrid can be single-stage or double-stage. In single stage configuration, PV array is directly connected to a DC/AC converter whereas in double-stage configuration, DC/DC converter is coupled in between the solar PV array and PV inverter and provides the desired fixed DC voltage to the inverter. The present work aims at modelling, design, development and control of a solar PV vii based microgrid for enhanced performance. Also, the characterization studies of the developed system have been carried out. Modeling of the system is required in order to predict its behaviour under both steady and dynamic states. Characterization studies such as sensitivity and reliability analysis are used to evaluate the performance of the system. Sensitivity analysis is the performance evaluation technique for evaluating the change in the system’s performance with respect to the change in its parameters. The sensitivity functions for solar cell and boost converter with respect to influential parameters have been developed using first derivative of Taylor’s series. Reliability analysis for electrical and electronic components of the system have been performed using pareto analysis and reliability model of the PV based microgrid has been developed using reliability block diagram for different PV array configurations. The Fault tree analysis (FTA) model of the system has been developed to find the cause of failure and to step the events leading to failure serially. Further, Markov’s model has been used to develop the reliability functions of individual components and hence, the reliability of complete grid connected PV system has been calculated. Solar PV system gives maximum power under uniform shading. But many a times PV panels are non-uniformly irradiated and this condition is known as called partial shading condition (PSC). PSC occur due to shadow of big trees, nearby buildings and dense clouds etc. PSC in PV system is an inevasible situation and exhibits multiple peaks, consisting of a single global maximum power point and many local maximum power points, in its power-voltage curve. PSC makes tracking of global maximum power point more difficult and also reduces the efficiency of the system. The conventional MPPT control algorithms work well under uniform shading condition but under partial shading scenario, they may not be able to track global peak out of multiple peaks. Therefore, an efficient controller is required to overcome the raised issue. Further, various PV array configurations such as series, series-parallel, total cross tied, bridge linked etc. may be used to improve the system efficiency. In the present work, novel maximum power point control algorithms viz. an asymmetrical fuzzy logic control (AFLC) and asymmetrical interval type-2 FLC (AIT-2 FLC) are developed for stand-alone PV system under partial shading condition. The developed algorithms are tested for different PV array configurations. viii In stand-alone PV system, the power supplied to the load depends upon the available solar energy. The output of SPV is intermittent in nature as it depends on the environmental conditions. This intermittency problem can be addressed by adding an energy storage system along with PV system. Battery is the most commonly used energy storage device and is very pivotal in maintaining continuity of power to the load. But when two or more energy sources are connected, then control of dc link voltage at common coupling point (CCP) is an area of concern. Therefore, in a SPV system with BESS a controller is required which can maintain constant DC link voltage irrespective of system transients. The PI controller is commonly used controller for controlling dc- link voltage, but it cannot regulate DC-link voltage under dynamic operating conditions and have overshoots and long settling time in its response. Suitable intelligent controllers are designed to replace the conventional PI controller, as they provide a better transient response. In order to overcome the drawbacks of the conventional PI control algorithm, nonlinear autoregressive moving average-L2 (NARMA-L2) control algorithm is proposed and developed for the stand-alone PV system with BESS. The proposed control scheme maintains the voltage across DC-link under change in irradiation and load condition. In a grid connected SPV based microgrid, the output of boost converter i.e., DC link is connected to voltage source inverter which is connected to grid at the point of common coupling (PCC). Voltage source inverter converts the generated DC power from PV system to AC of required voltage and frequency, as well as maintains the balance of power between the SPV system, load, and grid. The inverter is regulated by the interfacing controllers for effective operation and grid synchronization. The interfacing controllers are used to control the output of PV inverter for its efficient utilization and for improving power quality at PCC by providing reactive power compensation, harmonics compensation and load balancing. Conventional control algorithm like synchronous reference frame theory (SRFT) uses proportional integral (PI) controller for DC-link voltage regulation. These controllers are not best suited for SPV based microgrid as the overshoots and long settling time in their response are inevitable. In order to overcome this, novel smooth Least Mean Square (SLMS), improved zero attracting LMS (IZALMS) and reweighted L0 norm variable step size continuous mixed p-norm (RL0-VSSCMPN) based adaptive interfacing control algorithms are proposed ix and developed for the PV based microgrid. The efficacy of the proposed control algorithms has been tested on hardware prototype developed in the laboratory using MicroLab box (dSPACE 1202). The developed prototype system acts as distribution static compensator (DSTATCOM) and consists of inverter that is tied in parallel to the grid at the point of common coupling. FLUKE power analyzer has been used to measure the response of the system. The research work presented in the thesis is expected to provide good exposure to design, development and control of the solar PV based microgrid.

AbstractThe mean sea surface (MSS) is an important reference surface for oceanographic or geodetic applications such as sea level studies or the geodetic determination of the steady-state ocean circulation. Models of the MSS are derived from averaged along-track radar altimetry which provides instantaneous measurements of the sea surface heights (SSH). SSH observations corrected for tides and other physical signals and can be modeled as the sum of the MSS and sea level anomalies (SLA) which describe the temporal variability of the ocean. The typical MSS products are defined as grids of heights at a specific reference epoch and result from spatial and temporal prediction and filtering of the along-track SSH observations, whereas SLA products are computed with respect to an MSS model and are also defined as e.g. daily or averaged monthly grids.In this contribution a one-step least-squares approach is used to estimate a continuous spatio-temporal model of the MSS and filtered SLAs from along-track altimetric SSH measurements using C1-smooth finite element spaces for the spatial representation. The finite elements are defined on triangulations with different edge lengths and, thus, different spatial resolutions for MSS and SLA modeling. To model the temporal ocean variability finite B-Splines base functions are combined with the spatial finite elements to construct a spatio-temporal model. This contribution presents a concept to adapt the triangulations to the spatial characteristics of the signal of the MSS and SLA in a study region south of Africa. Least-squares residuals are studied to detect areas which show unmodeled spatial signal. These serve as input for the refinement of the triangulation. The results show that the residuals are indeed a good indicator for unmodeled signal, but as they are significantly influenced by unmodeled temporal signals as well, the refinement has only a small local impact on the obtained MSS and SLA models.

“Peak-locking” causes deterministic mean measurement bias in most of the existing cross-correlation based algorithms for PIV data analysis. This phenomenon is inherent to the typical smooth curve-fitting through discrete correlation values which are used to obtain the sub-pixel accuracy in velocity. In this paper we introduce a new algorithm/method for obtaining the sub-pixel accuracy, which eliminates the peak-locking effect. We refer to this procedure as “correlation mapping method”. In an ideal case, the second exposure (image 2) in a PIV measurement can be regarded as a mapping of the first exposure (image 1) where the mapping rules are affected by displacement, deformation, out of plane motion, etc. The correlation mapping method is based on shifting of image 1 by certain sub-pixel value, thus generating a virtual image (2′), whose gray level can be expressed in terms of the original image and the sub-pixel displacement. Thus, the correlation map of images 1 and 2′ is also a function of the intensity distribution in image 1 and the displacement. This correlation map is matched with the measured correlation map of images 1 and 2, providing a system of equations, one for discrete point in the correlation map with the sub-pixel values as unknowns. Solving these equations for each point in the vicinity of the correlation peak generates a series of sub-pixel displacements. Least square fitting is then used to determine the sub-pixel displacement with minimal difference between the real and virtual correlation values. This method is applied to several experimental and synthetic flow image pairs. In most cases the results show substantial improvements in sub-pixel accuracy in comparison to other algorithms and it eliminates the peak locking bias.

A ribbed square duct (P/e = 10, e/Dh = 0.10, and rotating with an angular velocity ωz) subjected to sand ingestion is studied using an Eulerian-Lagrangian framework. Particle sizes of 10μm and 50μm with response times (normalized by friction velocity and hydraulic diameter) of 0.06875 and 1.71875 respectively are considered. The calculations are performed for a nominal bulk Reynolds number of 20,000 and a Rotation number of 0.35 (based on hydraulic diameter and mean velocity) under fully-developed conditions. It is found that at any given instant in time about 67% of the total number of 10 micron particles are concentrated in the vicinity (within 0.05 Dh) of the duct surfaces, compared to 99% of the 50 micron particles. Both the particle sizes show preferential concentration near the smooth walls. The 10 micron particles, which are more sensitive to changes in flow, exhibit selective concentration on the trailing wall as compared to negligible concentration of the 50 micron particles. Both the particle sizes show negligible accumulation on the leading wall due to the action of Coriolis forces which push the particles towards the trailing wall. At the side walls of the duct, the 10 micron particles exhibit a high potential to erode the region in the vicinity of the rib due to secondary flow impingement. At the ribbed walls, while the 10 micron particles exhibit a fairly uniform but low propensity for erosion, the 50 micron particles show a much higher tendency to erode the surface. The trailing wall rib face facing the flow is by far the most susceptible to erosion and deposition for all particle sizes. In comparison to the front surface, the top and back surfaces of the rib do not exhibit a large propensity to be eroded. The potential for erosion in the rotating duct at the trailing wall and rib is at least an order of magnitude larger than in a stationary duct, whereas the leading wall has little or no erosion.