Conventional Vector Research Articles

A Modified Type-2 Neuro-fuzzy SVM based inverter fed IM drive

In indirect vector control, the conventional speed and current controllers operate satisfactorily when the operating point is constant. But the operating point is always dynamic; the reference voltages obtained in a closed-loop system feeding to the inverter contains more harmonics. Due to this, the pulses which are going to produce are uneven and in turn, understudy creates the inverter output voltages produced with more harmonics. In order to produce a better output voltage from the inverter, this paper presents the neuro-fuzzy type-2 (NFT2) space vector modulation technique. Performance comparison of the inverter is done with conventional space vector modulation (SVM) and neuro-fuzzy type-1 (NFT1) system using MATLAB simulation and experimental validation. The performance parameters of the induction motor based on current, torque, and speed with neuro-fuzzy type-2 space vector modulation is compared with conventional and type-1 neuro-fuzzy SVM. The % THD of inverter output voltages are also compared. The experimental validation of a DSpace-1104 is used to analyse the performance of the induction motor which is obtained by the simulation. The experimental validations are carried out considering 2HP induction motor in the lab.

240$^\circ$-Clamped PWM Applied to Transformerless Grid Connected PV Converters With Reduced Common Mode Voltage and Superior Performance Metrics

Transformerless voltage source inverters (VSIs) are one of the popular topologies for photovoltaic (PV) grid-connected applications due to the lowest component count and simple design. Because of the absence of galvanic isolation in such systems, problem of electromagnetic interference (EMI) due to high leakage current is highly pronounced. Recently various reduced common mode voltage (CMV) PWM (RCMV-PWM) methods which address these issues of CMV, and leakage current have been proposed. These schemes typically function at the expense of increased total harmonic distortion (THD), higher switching loss, high DC link current stress along with limited modulation index range. In this work, 240<inline-formula><tex-math notation="LaTeX">$^\circ$</tex-math></inline-formula> clamped PWM (240CPWM) is selected as an ideal candidate for grid connected VSIs in terms of all round performance of switching loss, THD, DC link current stress, CMV and leakage current. 240CPWM is a relatively new space vector based PWM method which can attain much superior performance by avoiding the zero states. After a brief overview of the 240CPWM concept, the paper provides a detailed comparison of 240CPWM with conventional space vector PWM (CSVPWM) and discontinuous PWM (DPWM1) methods in terms of each of the above metrics. A 3-phase 208 V 3 kW silicon IGBT based hardware prototype is built to validate the performance of 240CPWM and compared with other popular schemes under grid-connected mode. Experimental results show a 66.7&#x0025; reduction in the peak CMV and 50&#x0025; reduction in leakage current with the proposed scheme as compared to CSVPWM. Also, THD, DC link current stress and inverter switching loss are greatly reduced as an added advantage using the proposed method. A new, combined performance index is proposed to compare the performance of different PWM schemes, and it is shown that the 240CPWM achieves the best value for this index among the PWM methods studied.

Minicircles for CAR T Cell Production by Sleeping Beauty Transposition: A Technological Overview.

Development and application of chimeric antigen receptor (CAR) T cell therapy has led to a breakthrough in the treatment of hematologic malignancies. In 2017, the FDA approved the first commercialized CD19-specific CAR T cell products for treatment of patients with B-cell malignancies. This success increased the desire to broaden the availability of CAR T cells to a larger patient cohort with hematological but also solid tumors. A critical factor of CAR T cell production is the stable and efficient delivery of the CAR transgene into T cells. This gene transfer is conventionally achieved by viral vectors. However, viral gene transfer is not conducive to affordable, scalable, and timely manufacturing of CAR T cell products. Thus, there is a necessity for developing alternative nonviral engineering platforms, which are more cost-effective, less complex to handle and which provide the scalability requirement for a globally available therapy.One alternative method for engineering of T cells is the nonviral gene transfer by Sleeping Beauty (SB) transposition. Electroporation with two nucleic acids is sufficient to achieve stable CAR transfer into T cells. One of these vectors has to encode the gene of interest, which is the CAR , the second one a recombinase called SB transposase, the enzyme that catalyzes integration of the transgene into the host cell genome. As nucleic acids are easy to produce and handle SB gene transfer has the potential to provide scalability, cost-effectiveness, and feasibility for widespread use of CAR T cell therapies.Nevertheless, the electroporation of two large-size plasmid vectors into T cells leads to high T cell toxicity and low gene transfer rates and has hindered the prevalent clinical application of the SB system. To circumvent these limitations, conventional plasmid vectors can be replaced by minimal-size vectors called minicircles (MC ). MCs are DNA vectors that lack the plasmid backbone, which is relevant for propagation in bacteria, but has no function in a human cell. Thus, their size is drastically reduced compared to conventional plasmids. It has been demonstrated that MC-mediated SB CAR transposition into T cells enhances their viability and gene transfer rate enabling the production of therapeutic doses of CAR T cells. These improvements make CAR SB transposition from MC vectors a promising alternative for engineering of clinical grade CAR T cells.

Application of artificial intelligence techniques in the operation of neutral-point clamped rectifier under perturbed conditions

This paper will demonstrate the viability of neural network algorithm for solving power quality problems in three-phase neutral-point clamped converter. A neural network algorithm is proposed here for the better performance of the converter under perturbed conditions. The neural network algorithm modifies conventional space vector pulse width modulation algorithm. The proposed algorithm generates ideal switching sequences for the converter by developing the best path for the reference vector in the hexagon of the space vector. The data set required to train the neural network was collected from well tuned PI controlled model. In this work, the implementation of neural network is described and simulation results are presented using MATLAB/Simulink software. Simulation results proves fast minimising the potential difference at the clamped bus of the converter with minimum switching losses and excellent performance of converter in terms of unity power factor, better total harmonic distortion at the source side.

Grid integration of multiple wind turbines using a multi-port converter—A novel simultaneous space vector modulation

In this paper, an innovative approach of integrating multiple doubly-fed induction generator (DFIG)-based wind turbines using a single multi-port converter unit, named as a unified expandable power converter (UEPC) is presented. A novel partial simultaneous space vector modulation (PSSVM) scheme, adopted to the unified structure, is proposed to enhance the operational efficiency and DC-link utilization. The unified configuration can integrate two or more DFIGs through a multi-port converter in a cost-effective and efficient way. The unified structure enjoys compactness, plug & play capability, fewer switches, and a lower installation cost. Comparing with the conventional sequential space vector modulation (SSVM), which only uses the individual active vectors, the proposed PSSVM can utilize common active vectors between different ports, which is the main salient feature of this method. The performance of PSSVM is also validated via an OPAL-RT real-time simulator. The real-time outcomes prove that the proposed PSSVM can improve the DC-link voltage utilization and switching loss compared with the SSVM. The effectiveness of the proposed PSSVM based unified configuration in both normal and faulty conditions are evaluated and compared with the conventional structure using MATLAB/Simulink. The simulation results endorse the finding that regardless of the shared switches in the unified expandable power converter, the independent control and maximum power extraction from the wind are guaranteed using the proposed PSSVM.

ОЖИДАЕМАЯ ПРОЦЕНТНАЯ СТАВКА И ЕЕ ВОЗДЕЙСТВИЕ НА ЦЕНЫ СЫРЬЕВЫХ ТОВАРОВ

The subject of the study is the impact of the expected interest rate on commodity prices. The purpose of the work is to determine the degree of influence of the expected interest rate on commodity prices. Structural and conventional vector autoregression models were used. The period 1998–2020 allows us to consider the impact of the expected interest rate on commodities in dynamics. The results demonstrated that the expected interest rate, measured through futures contracts, has the opposite effect on the prices of most commodities. The effect of shock persists for 3-5 months. The direct contribution of the expected interest rate to the variance of the forecast does not exceed 20%, but the indirect effect through the dollar exchange rate can be seen from the results of the analysis of the decomposition of the variance. The use of the expected interest rate for forecasting commodity prices is justified due to the better efficiency, according to the coefficient of average absolute error, compared with the “naïve” model. The results obtained can be used by investors and managers as an additional factor in the formation of deposit portfolios.

Separating P- and S-Waves Based on the Slope of Wavefield Events and Polarizability

In multicomponent surveys, the travel path of seismic waves from the subface interfaces to geophones may not be vertical to the ground; consequently, vertical and horizontal geophones receive mixed wavefields comprising P- and S-waves. Vector separation of wavefields is crucial for subsequent imaging and shear wave splitting analyses. However, conventional methods cannot address the nonlinear polarization of pure wave and can also produce incorrect results due to the presence of dipping interface. To address the above issues, we propose a vector wavefield separation method based on the slope of events and polarizability. The proposed method calculates the slope of the events for determining the propagation direction of the wavefield. The polarizability is used to evaluate if the resulting wavefield is purely composed of one type of body wave while the wavefield type is determined based on the angle between the polarization and propagation directions. When the wavefield is mixed, P/S separation is realized based on vector decomposition in an affine coordinate system optimized by propagation direction. We use both numerical and field data to demonstrate the effectiveness and applicability of the proposed method. The numerical example confirms that the proposed method yields more accurate results than the conventional vector decomposition method. The field data application also indicates that the new method can effectively resolve the P- and S-waves from the noisy seismic wavefield.

Crude Oil Spot Price Forecasting Using Ivanov‐Based LASSO Vector Autoregression

This paper proposes a forecasting methodology that investigates a set of different sparse structures for the vector autoregression (VAR) model using the Ivanov‐based least absolute shrinkage and selection operator (LASSO) framework. The variant auxiliary problem principle method is used to solve the various Ivanov‐based LASSO‐VAR variants, which is supported by parallel computing with simple closed‐form iteration and linear convergence rate. A test case with ten crude oil spot prices is used to demonstrate the improvement in forecasting skills gained from exploring sparse structures. The proposed method outperformed the conventional vector autoregressive model.

Acceleration strategies for Tridimensional Coupled hydromechanical problems based on CPU and GPU programming in MATLAB.

Large-scale fluid flow in porous media demands intense computations and occurs in the most diverse applications, including groundwater flow and oil recovery. This article presents novel computational strategies applied to reservoir geomechanics. Advances are proposed for the efficient assembly of finite element matrices and the solution of linear systems using highly vectorized code in MATLAB. In the CPU version, element matrix assembly is performed using conventional vectorization procedures, based on two strategies: the explicit matrices, and the multidimensional products. Further assembly of the global sparse matrix is achieved using the native sparse function. For the GPU version, computation of the complete set of element matrices is performed with the same strategies as the CPU approach, using gpuArray structures and the native CUDA support provided by MATLAB Parallel Computing Toolbox. Solution of the resulting linear system in CPU and GPU versions is obtained with two strategies using a one-way approach: the native conjugate gradient solver (pcg), and the one provided by the Eigen library. A broad discussion is presented in a dedicated benchmark, where the different strategies using CPU and GPU are compared in processing time and memory requirements. These analyses present significant speedups over serial codes.

Sliding Mode Predictive Current Control of Permanent Magnet Synchronous Motor With Cascaded Variable Rate Sliding Mode Speed Controller

To solve the problem that the conventional vector control system of permanent magnet synchronous motor is susceptible to motor parameters and load disturbances, a sliding mode predictive current control strategy of permanent magnet synchronous motor with cascaded variable rate sliding mode speed controller based on extended state observer (ESO) is proposed, which improves the dynamic and static performance of the system. First, a sliding mode speed controller is designed based on the novel variable rate exponential reaching law, and the ESO is used to estimate and compensate for parameter mismatch and disturbance, which improves the system&#x2019;s rapidity and robustness while reducing the chattering. Then, to solve the problem that the traditional deadbeat predictive control of current loop ignores the influence of cross-coupling electromotive force, based on the above reaching law, a sliding mode controller is designed to equate the back electromotive force to effectively compensate the deadbeat predictive current controller, which not only retains the rapidity of current but also enhances the robustness. Finally, simulation and experimental results show the effectiveness of the proposed control strategy.

Joint Deep-Learning-Enabled Impact of Holistic Care on Line Coagulation in Hemodialysis.

In order to investigate the impact of holistic care on line coagulation and safety in hemodialysis and to address limitations of the conventional ultrasound flow vector imaging (VFM) technique, which requires proprietary software to acquire raw Doppler and scatter tracking data, a combined deep-learning-enabled holistic care approach to line coagulation in hemodialysis is proposed. First, velocity along the direction of the sound beam, which is provided by the color Doppler echocardiogram, is obtained as the radial velocity component using a velocity scale. Moreover, the left ventricular wall contour is automatically identified using a U-Net network and the left ventricular wall velocity is calculated as the boundary condition of continuity equation by a retrained PWC-Net model. Likewise, the velocity component of each blood mass in the vertical direction of the sound beam is obtained by solving the continuity equation (i.e. tangential velocity component). Finally, the velocity vector map of cardiac flow field was synthesized and visualized in the flow diagram. For this purpose, sixty patients admitted to receive hemodialysis from February 2019 to June 2020 were randomly divided into two groups of 30 patients where the control group implemented conventional care and the study group implemented all-round care on the basis of conventional care. The nursing effects of both groups were compared. Incidence of pipeline coagulation and complications in the study group were lower than those in the control group and the difference was statistically significant (P < 0.05). The nursing detail score, nursing attitude score, nursing professionalism score, and total satisfaction score in the study group were higher than those in the control group and the difference was statistically significant (P < 0.05). Applying all-round nursing in hemodialysis can effectively reduce the incidence of line coagulation complications and improve the safety of hemodialysis, as well as improve patients' satisfaction with nursing care.

An Evaluation of Speech-Based Recognition of Emotional and Physiological Markers of Stress

Life in modern societies is fast-paced and full of stress-inducing demands. The development of stress monitoring methods is a growing area of research due to the personal and economic advantages that timely detection provides. Studies have shown that speech-based features can be utilised to robustly predict several physiological markers of stress, including emotional state, continuous heart rate, and the stress hormone, cortisol. In this contribution, we extend previous works by the authors, utilising three German language corpora including more than 100 subjects undergoing a Trier Social Stress Test protocol. We present cross-corpus and transfer learning results which explore the efficacy of the speech signal to predict three physiological markers of stress—sequentially measured saliva-based cortisol, continuous heart rate as beats per minute (BPM), and continuous respiration. For this, we extract several features from audio as well as video and apply various machine learning architectures, including a temporal context-based Long Short-Term Memory Recurrent Neural Network (LSTM-RNN). For the task of predicting cortisol levels from speech, deep learning improves on results obtained by conventional support vector regression—yielding a Spearman correlation coefficient (ρ) of 0.770 and 0.698 for cortisol measurements taken 10 and 20 min after the stress period for the two corpora applicable—showing that audio features alone are sufficient for predicting cortisol, with audiovisual fusion to an extent improving such results. We also obtain a Root Mean Square Error (RMSE) of 38 and 22 BPM for continuous heart rate prediction on the two corpora where this information is available, and a normalised RMSE (NRMSE) of 0.120 for respiration prediction (−10: 10). Both of these continuous physiological signals show to be highly effective markers of stress (based on cortisol grouping analysis), both when available as ground truth and when predicted using speech. This contribution opens up new avenues for future exploration of these signals as proxies for stress in naturalistic settings.

An Effective Decoupling Control with Simple Structure for Induction Motor Drive System Considering Digital Delay

Digital processing poses a considerable time delay on controllers of induction motor (IM) driving system, which degrades the effects of torque/flux decoupling, slows the motor torque response down, or even makes the entire system unstable, especially when operating at a low switching frequency. The existing methods, such as feed-forward and feed-back decoupling methods based on the proportional integral controller (PI), have an intrinsic disadvantage in the compromise between high performance and low switching frequency. Besides, the digital delay cannot be well compensated, which may affect the system loop and bring instability. Conventional complex vector decoupling control based on an accurate IM model employs complicated decoupling loops that may be degraded by digital delay leading to discrete error. This article aims to give an alternative complex vector decoupling solution with a simple structure, intended for optimized decoupling and improving the system dynamic performance throughout the entire operating range. The digital delay-caused impacts, including secondary coupling effect and voltage vector amplitude/phase inaccuracy, are specified. Given this, the digital delay impact is canceled accurately in advance, simplifying the entire decoupling process greatly while achieving uncompromised decoupling performance. The simulation and experimental results prove the effectiveness and feasibility of the proposed decoupling technique.

Air‐gap flux‐oriented vector control of dual stator induction generator used in wind energy conversion system with novel 13‐zone time optimized space vector‐based hybrid bus clamping PWM

This work presents a novel speed sensor less air-gap flux-oriented vector control of dual stator induction generator (DSIG) used for grid connected, variable speed wind energy generation system. The DSIG consists of two stator windings electrically separated by an angle of 30° with dissimilar pole numbers, which are in the ratio of 1:3. This arrangement offers reduced magnetic core saturation and less harmonics in stator currents, which results into reduced stator winding losses. Moreover, presence of two stator windings makes DSIG suitable to be used in high power application. In conventional rotor-flux-oriented control, the rotor flux is oriented along the d-axis of the rotor winding. But the rotor flux is not easily accessible. Air-gap flux-oriented vector control (AGFOVC) is advantageous as compared to conventional vector control because the air-gap flux is easily accessible and does not require any special arrangement for the measurement of variables. Moreover incorporation of this novel speed sensor less air-gap flux-oriented vector control also eliminates the use of speed encoder as the rotor speed is estimated from the machine terminal variables. Elimination of speed encoder not only reduces the cost of the overall system but also nullifies the error due to mismeasurement of speed and enhances control flexibility and robustness. As a result of which robustness of the system increases. Here, two separate converters are used for the two stator windings of the machine. And the converter switching's are being performed by space vector-based novel dwell time optimized 13-zone hybrid PWM. This special PWM results in better harmonic performance, reduced torque pulsation, and minimal switching loss. The simulation of the overall system is performed in MATLAB environment and the simulated results are validated experimentally using dSPACE controller board DS1104. Close proximity in simulated and experimental waveforms justifies the effectiveness of the proposed work.

Two-Dimensional Broad Learning System for Data Analytic

Recently a novel learning algorithm called the broad learning system (BLS) has been established by the theory of pseudo-inverse and compressed sensing technology. As an alternative to deep learning, BLS has gained extensive attention in data analytics. Unlike the existing deep structures based learning tools (such as deep neural networks), it is able to rapidly achieve incremental learning, and does not require tedious retraining to remodel the system. However, given image data modeling tasks, one-dimensional BLS requires a conventional vectorization operation, the dimension of input will be quite large for high resolution image. This operation may result in huge computational complexity, thereby increasing the time of model construction. To resolve this issue, a two-dimensional version of BLS, namely 2D-BLS, is first proposed to fast construct randomized learner models by using matrix inputs. Specifically, the proposed method employs left and right projecting vectors to replace the usual high dimensional input weight. Moreover, the theoretical analyses on the superiority of 2D-BLS against BLS are presented in terms of the algorithm complexity and the difference of random parameter space. Empirical results based on MNIST dataset, NORB 3D object image recognition dataset, ORL dataset, YaleB dataset and Handwritten Digit dataset demonstrate that the proposed 2D-BLS outperforms the original BLS, some extensions of BLS and some 2-D machine learning methods in terms of the efficiency of model construction. This result also shows good potential of 2D-BLS for image data analytics.

Space vector PWM ‐based current error boundary investigations for three‐level VSI fed induction motor drive

Field oriented control (FOC) of induction motor (IM) powered by a three-level voltage source inverter (3L -VSI) with space vector (SV) based continuous and dis -continuous pulse width modulation (PWM) is presented in this paper. The current error variations in each sector with above-mentioned PWM techniques are monitored and analyzed mathematically. These boundary investigations play a great role in the implementation of Current Error Space Phasor (CESP) based hysteresis current controller (HCC) which incorporates the advantages of a constant switching frequency PWM fed three-level inverter (3LI) into highly dynamic current controlled – voltage source inverter. As a preliminary step, FOC of IM with the conventional space vector PWM and bus clamping PWM (BC30 and BC60) fed 3LI have been simulated using SIMULINK toolbox of MATLAB. Transient analysis of the drive with acceleration, deceleration, and load change studies with different switching schemes are analyzed with d SPACE RTI1103 experimentation on a 2.2 kW three-phase IM drive in detail.

Robust Agrobacterium-Mediated Transient Expression in Two Duckweed Species (Lemnaceae) Directed by Non-replicating, Replicating, and Cell-to-Cell Spreading Vectors.

Plant-based transient expression systems have recognized potential for use as rapid and cost-effective alternatives to expression systems based on bacteria, yeast, insect, or mammalian cells. The free-floating aquatic plants of the Lemnaceae family (duckweed) have compact architecture and can be vegetatively propagated on low-cost nutrient solutions in aseptic conditions. These features provide an economically feasible opportunity for duckweed-based production of high-value products via transient expression of recombinant products in fully contained, controlled, aseptic and bio-safe conditions in accordance with the requirements for pharmaceutical manufacturing and environmental biosafety. Here, we demonstrated Agrobacterium-mediated high-yield transient expression of a reporter green fluorescent protein using deconstructed vectors based on potato virus X and sweet potato leaf curl virus, as well as conventional binary vectors, in two representatives of the Lemnaceae (Spirodela polyrhiza and Landoltia punctata). Aseptically cultivated duckweed populations yielded reporter protein accumulation of >1 mg/g fresh biomass, when the protein was expressed from a deconstructed potato virus X-based vector, which is capable of replication and cell-to-cell movement of the replicons in duckweed. The expression efficiency demonstrated here places duckweed among the most efficient host organisms for plant-based transient expression systems, with the additional benefits of easy scale-up and full containment.

Temporally Generalizable Land Cover Classification: A Recurrent Convolutional Neural Network Unveils Major Coastal Change through Time

The ability to accurately classify land cover in periods before appropriate training and validation data exist is a critical step towards understanding subtle long-term impacts of climate change. These trends cannot be properly understood and distinguished from individual disturbance events or decadal cycles using only a decade or less of data. Understanding these long-term changes in low lying coastal areas, home to a huge proportion of the global population, is of particular importance. Relatively simple deep learning models that extract representative spatiotemporal patterns can lead to major improvements in temporal generalizability. To provide insight into major changes in low lying coastal areas, our study (1) developed a recurrent convolutional neural network that incorporates spectral, spatial, and temporal contexts for predicting land cover class, (2) evaluated this model across time and space and compared this model to conventional Random Forest and Support Vector Machine methods as well as other deep learning approaches, and (3) applied this model to classify land cover across 20 years of Landsat 5 data in the low-lying coastal plain of North Carolina, USA. We observed striking changes related to sea level rise that support evidence on a smaller scale of agricultural land and forests transitioning into wetlands and “ghost forests”. This work demonstrates that recurrent convolutional neural networks should be considered when a model is needed that can generalize across time and that they can help uncover important trends necessary for understanding and responding to climate change in vulnerable coastal regions.

Transduction Properties of an Adenovirus Vector Containing Sequences Complementary to a Liver-Specific microRNA, miR-122a, in the 3'-Untranslated Region of the E4 Gene in Human Hepatocytes from Chimeric Mice with Humanized Liver.

Replication-incompetent adenovirus (Ad) vectors are promising gene delivery vehicles, especially for hepatocytes, due to their superior hepatic tropism; however, in vivo application of an Ad vector often results in hepatotoxicity, mainly due to the leaky expression of Ad genes from the Ad vector genome. In order to reduce the Ad vector-induced hepatotoxicity, we previously developed an Ad vector containing the sequences perfectly complementary to a liver-specific microRNA (miRNA), miR-122a, in the 3'-untranslated region (UTR) of the E4 gene. This improved Ad vector showed a significant reduction in the leaky expression of Ad genes and hepatotoxicity in the mouse liver and primary mouse hepatocytes; however, the safety profiles and transduction properties of this improved Ad vector in human hepatocytes remained to be elucidated. In this study, we examined the transgene expression and safety profiles of Ad vectors with miR-122a-targeted sequences in the 3'-UTR of the E4 gene in human hepatocytes from chimeric mice with humanized liver. The transgene expression levels of Ad vectors with miR-122a-targeted sequences in the 3'-UTR of the E4 gene were significantly higher than those of the conventional Ad vectors. The leaky expression levels of Ad genes of Ad vectors with miR-122a-targeted sequences in the 3'-UTR of the E4 gene in the primary human hepatocytes were largely reduced, compared with the conventional Ad vectors, resulting in an improvement in Ad vector-induced cytotoxicity. These data indicated that this improved Ad vector was a superior gene delivery vehicle without severe cytotoxicity for not only mouse hepatocytes but also human hepatocytes.

Kernelized support tensor train machines

Tensor, a multi-dimensional data structure, has been exploited recently in the machine learning community. Traditional machine learning approaches are vector- or matrix-based, and cannot handle tensorial data directly. In this paper, we propose a tensor train (TT)-based kernel technique for the first time, and apply it to the conventional support vector machine (SVM) for high-dimensional image classification with very small number of training samples. Specifically, we propose a kernelized support tensor train machine that accepts tensorial input and preserves the intrinsic kernel property. The main contributions are threefold. First, we propose a TT-based feature mapping procedure that maintains the TT structure in the feature space. Second, we demonstrate two ways to construct the TT-based kernel function while considering consistency with the TT inner product and preservation of information. Third, we show that it is possible to apply different kernel functions on different data modes. In principle, our method tensorizes the standard SVM on its input structure and kernel mapping scheme. This reduces the storage and computation complexity of kernel matrix construction from exponential to polynomial. The validity proof and computation complexity of the proposed TT-based kernel functions are provided elaborately. Extensive experiments are performed on high-dimensional fMRI and color images datasets, which demonstrates the superiority of the proposed scheme compared with the state-of-the-art techniques.