Compared with silicon‐based Insulated Gate Bipolar Transistors (IGBTs), silicon carbide (SiC) Metal‐Oxide‐Semiconductor Field‐Effect Transistors (MOSFETs) are characterized by higher operating temperatures, switching speeds and switching frequencies, and are considered the next evolutionary step for future electric drives. The application of SiC MOSFETs in the field of electrified vehicles has brought many benefits, such as higher efficiency, higher power density, and simplified cooling system, and can be seen as an enabler for high‐power fast battery charging. This article reviews the benefits of SiC MOSFETs in different electrified vehicle (EV) application scenarios, including traction inverters, on‐board converters, and off‐board charging applications. However, replacing Si‐IGBTs with SiC MOSFETs introduces several new technical challenges, such as stronger electromagnetic interference (EMI), reliability issues, potential electric machine insulation failure due to high transient voltages, and cooling difficulties. Compared to mature silicon‐based semiconductor technologies, these challenges have so far hindered the widespread adoption of SiC MOSFETs in automotive applications. To fully exploit the advantages of SiC MOSFETs in automotive applications and enhance their reliability, this paper explores future technology developments in SiC MOSFET module packaging and driver design, as well as novel electric machine drive strategies with higher switching frequencies, and optimized high‐frequency machine design.

The perceptual quality of stereoscopic images plays an essential role in the human perception of visual information. However, most available stereoscopic image quality assessment (SIQA) methods evaluate 3D visual experience using hand-crafted features or shallow architectures, which cannot model the visual properties of stereo images well. In this paper, we use convolutional neural networks (CNNs) to learn deeper local quality-aware structures for stereo images. With different inputs, two CNN models are designed for no-reference SIQA tasks. The one-column CNN model directly accepts a cyclopean view as the input, and the three-column CNN model jointly considers the cyclopean, left and right views as CNN inputs. The two SIQA frameworks share the same implementation approach: First, to overcome the obstacle of limited SIQA datasets, we accept image patches that have been cropped from corresponding stereopairs as inputs for local quality-sensitive feature extraction. Next, a local feature selection algorithm is used to remove related features on non-salient patches, which could cause large prediction errors. Finally, the reserved local visual structures of salient regions are aggregated into a final quality score in an end-to-end manner. Experimental results on three public SIQA databases demonstrate that our method outperforms most state-of-the-art no-reference (NR) SIQA methods. The results of a cross-database experiment also show the robustness and generality of the proposed method.

Motivated by the success of Convolutional Neural Networks (CNNs) in image-related applications, in this paper, we design an effective method for no-reference 3D image quality assessment (3D IQA) through CNN-based feature extraction and consolidation strategy. In the first and most vital stage, quality-aware features, which reflect the inherent quality of images, are extracted by a fine-tuned CNN model exploiting the concept of transfer learning. This fine-tuning strategy solves the large-scale training data dependence existing in current deep-learning-based IQA algorithms. In the second stage, features from left and right view are consolidated by linear weighted fusion where the weight for each image is obtained from its saliency map. In addition, statistical characteristics of the disparity map are also considered in a multi-scale manner as additional features. In the final stage of quality mapping, the objective score for each stereoscopic pair is gained by Support Vector Regression. Experimental results on the public databases show that our approach outperforms many existing no-reference and even full-reference methods.

Accurate mathematical model is the basis and precondition of electric motor control. The existing motor model under concentrated parameters is difficult to describe the harmonic and nonlinear characteristics of the motor, while the motor model under distributed parameters can accurately describe the torque and flux characteristics of the motor precisely at all working points, which provides the foundation for model-based motor control. The stator flux and torque observer have an essential influence on the performance of direct torque control (DTC) for PMSM. The concentrated parameter model of PMSM considering only the fundamental wave vector of the magnetic field for DTC will have obvious deficiencies in the performance such as torque ripple. This paper therefore uses the reconstructed magnetic co-energy (MCE) model to establish a torque and flux observer based on distributed parameters, which is introduced into the DTC algorithm of PMSM as an observer. The optimized DTC algorithm makes full use of the flux and torque harmonic description capabilities of the distributed parameter model, which improves control performance. The superiority of the distributed parameter observer is verified by simulation and bench test.

Dual three-phase (DTP) electric machines are increasingly favored in industrial applications and automotive electric drive systems. Different machine types are used in different scenarios due to their unique advantages. However, the corresponding parametric models developed for different machine types increase development and calibration costs, and they are poorly compatible with each other. In this context, this article introduces a unified mathematical model that combines the electromagnetic characteristics of different DTP machines by introducing the concept of a “Faraday reference frame” for the use of permanent magnet synchronous machines (PMSMs), reluctance synchronous machines (RSMs), externally excited synchronous machines (EESMs), and induction machines (IMs). A unified state-space model of DTP machines is developed. The model adapted to different machine types is discussed. Furthermore, the unified flux and torque observers are built on the basis of the proposed model. The proposed observers are validated in a circuit-level simulation and an electromagnetic-level co-simulation with finite element analysis (FEA) modeling of different machines. By comparing the observer output with the machine output, the model can observe the flux at $s\alpha /s\beta $ subplane with an estimation error less than 4.9% NRMSE. The algorithm is validated at the physical level in an experimental environment, achieving torque estimation accuracy greater than 99.38% under steady-state operating conditions.