A supervised machine learning (ML) algorithm is proposed for measuring refractive index (RI) values both below and above the RI of the fiber material using a multimode interference (MMI) fiber sensor. The sensor is constructed by splicing a coreless multimode fiber (CMF) segment between two single-mode fiber (SMF) leads. Measurement of low and high RI regimes is accomplished through the Decision Tree (DT) regression algorithm. The trained model algorithm demonstrates a wide dynamic range in RI measurement, covering the ranges of 1.30 to 1.39 (low RI regime) and 1.46 to 1.55 (high RI regime) without any RI ambiguity, achieving model accuracy of 99.77%. The guided and leaky modes mechanisms within the all-silica-based structure of the CMF are fundamentally insensitive to temperature, making it highly practical for deployment in conditions with varying temperatures without the need for any compensating scheme.

This paper reports a robust high refractive index (RI) sensor based on a multimode silica rod structure. The sensor was fabricated by splicing a segment of multimode silica rod between two single-mode fiber (SMF) leads. The measurement of high-RI is realized by monitoring the change of power spectrum caused by the reaction of leaky modes to different surrounding high-RI. The 0.5-cm long multimode silica rod sensing element exhibits a sensitivity of up to 115.12 dBm/RIU for high-RI ranging between 1.450 and 1.531. The sensor also demonstrated its potential in power transformer application by distinguishing various ageing levels of transformer oil. The leaky modes mechanism in all-silica based structure is essentially temperature insensitive, and therefore very practical to be deployed in extreme temperature conditions.

In this study, a novel and simple optical fiber sensor is proposed, simulated, and experimentally demonstrated for simultaneous measurement of high refractive index (RI) and temperature. The sensor consists of a section of single-mode-silica rod-single-mode (SSRS) fiber structure cascaded to fiber Bragg grating (FBG) (SSRS-FBG). The simultaneous measurement of high RI and temperature was realized in this study by monitoring the output power level and wavelength shift of the SSRS-FBG single dip transmission spectrum corresponding to different surrounding high RI and ambient temperature values. The experimental results showed that the cascaded sensor had a sensitivity of 108.07 dBm/RIU for high RI ranging at 1.45-1.531, while the temperature sensitivity was 9.31 pm/°C from 35 °C to 85 °C. This sensor is most suitable for high RI applications, such as for monitoring oil quality stored in terminal and power transformers.

A novel leaky mode coupler of multimode interference (MMI) fiber structure is proposed and experimentally demonstrated for surrounding refractive indices higher than that of silica fiber. The sensor is composed of a section of silica rod (SR) and multimode fiber (MMF) between two identical single-mode fibers (SMFs) to form a single-mode-silica rod-multimode-single-mode (SSRMS) fiber structure. In the proposed design, the SR acts as a mode coupler and a sensing element, while the MMF functions as the narrow spectral width generator. The output power at narrow spectral response of the SSRMS sensor was monitored corresponding to different surrounding high refractive indices (SHRI). The experimental results revealed that the SSRMS sensor exhibited high sensitivity up to −293.40 dB/RIU for SHRI ranging from 1.450 to 1.531. The sensor is highly suitable for SHRI applications, such as for monitoring insulation oil in a power transformer.

The accurate estimation of lithium-ion battery state of health (SOH) is crucial for battery monitoring, safety, and degradation assessment; however, it remains challenging because of the nonlinear nature of battery degradation, measurement noise, and variability in the battery aging trajectory. This study aims to solve these problems by proposing a hybrid attention-based BiLSTM–RF model, which combines wavelet-based signal denoising, incremental capacity analysis (ICA)-based feature extraction, stacked Bidirectional Long Short-Term Memory (BiLSTM) networks, multi-head self-attention, principal component analysis (PCA)-based feature compression, and ensemble regression using a Random Forest (RF) model with adaptive weighted fusion. The proposed framework was tested on the NASA battery datasets (B0005, B0006, B0007 and B0018) and was further validated on the Oxford Battery Degradation Dataset using leave-one-battery-out cross validation conditions. Experimental results indicated that, in general, the proposed framework outperformed the evaluated benchmark models (CNN-LSTM, BiLSTM, and RF models) in terms of the prediction error, with a minimum RMSE value of 0.0229 for NASA battery B0007 and 0.0024 for Oxford Cell3. Ablation analysis also showed that the combination of wavelet denoising, PCA compression, temporal sequence learning and ensemble regression played a role in the overall SOH estimation performance. These results show that the proposed hybrid approach is effective and stable for SOH estimation in different battery degradation trajectories under the tested experimental conditions.

Wide-bandgap (WBG) semiconductor devices such as silicon carbide (SiC) and gallium nitride (GaN) enable higher switching frequencies, greater efficiency, and increased power density, but their fast-switching transients and steep dv/dt and di/dt characteristics generate substantially elevated electromagnetic interference (EMI). This review synthesizes peer-reviewed studies to quantitatively compare EMI mitigation outcomes across source-side modulation, gate-driving, filtering, packaging, and AI-based techniques for WBG converters. Packaging-integrated common-mode screens cut CM current by up to 26 dB while raising partial-discharge inception voltage by 53%. Chaotic PWM with passive filtering reaches up to 50 dB attenuation with a 74% reduction in inductor volume. AI-based closed-loop adaptation attains up to 19.2 dB average attenuation with 98.5% CISPR 25 compliance, while RL-based filters reach 25-30 dB across wide frequency ranges. Active gate-driving reduces peak EMI by 19-39 dB. No single technique simultaneously delivers the highest suppression, efficiency, and lowest cost. Hybrid Si/WBG design currently offers the most balanced trade-off. These outcomes are consolidated into a taxonomy of propagation mechanisms, mitigation techniques, application-specific strategies across five domains, and open gaps in standardized testing and validation. These findings provide a practical, quantitative reference for engineers designing next-generation, EMC-compliant WBG power electronic systems.