Performance Comparison of Lattice-Matched AlInN/GaN/AlGaN/GaN Double-Channel Metal–Oxide–Semiconductor High-Electron Mobility Transistors with Planar Channel and Multiple-Mesa-Fin-Channel Array

In this work, Al0.83In0.17N/GaN/Al0.18Ga0.82N/GaN epitaxial layers used for the fabrication of double-channel metal–oxide–semiconductor high-electron mobility transistors (MOSHEMTs) were grown on silicon substrates using a metalorganic chemical vapor deposition system (MOCVD). A sheet electron density of 1.11 × 1013 cm−2 and an electron mobility of 1770 cm2/V-s were obtained. Using a vapor cooling condensation system to deposit high insulating 30-nm-thick Ga2O3 film as a gate oxide layer, double-hump transconductance behaviors with associated double-hump maximum extrinsic transconductances (gmmax) of 89.8 and 100.1 mS/mm were obtained in the double-channel planar MOSHEMTs. However, the double-channel devices with multiple-mesa-fin-channel array with a gmmax of 148.9 mS/mm exhibited single-hump transconductance behaviors owing to the better gate control capability. Moreover, the extrinsic unit gain cutoff frequency and maximum oscillation frequency of the devices with planar channel and multiple-mesa-fin-channel array were 5.7 GHz and 10.5 GHz, and 6.5 GHz and 12.6 GHz, respectively. Hooge’s coefficients of 7.50 × 10−5 and 6.25 × 10−6 were obtained for the devices with planar channel and multiple-mesa-fin-channel array operating at a frequency of 10 Hz, drain–source voltage of 1 V, and gate–source voltage of 5 V, respectively.

Outcomes and indications of the single approach to double-channel core decompression and bone grafting with structural bone support for osteonecrosis of the femoral head

BackgroundTo report the outcomes of the single approach to double-channel core decompression and bone grafting with structural bone support (SDBS) for osteonecrosis of the femoral head (ONFH) and define the indications.MethodsOne-hundred-and-thirty-nine hips in 96 patients (79 males, 17 females; mean age 37.53±10.31 years, range 14–58 years; mean body mass index 25.15±3.63 kg/m2) were retrospectively analysed. The Harris hip score (HHS) was used to assess hip function, and radiographs were used to assess the depth of femoral head collapse. Treatment failure was defined as the performance of total hip arthroplasty (THA). The variables assessed as potential risk factors for surgical failure were: aetiology, Japanese Osteonecrosis Investigation Committee (JIC) type, age, and Association Research Circulation Osseous (ARCO) stage. Complications were recorded.ResultsThe HHS increased from 79.00±13.61 preoperatively to 82.01±17.29 at final follow-up (P=0.041). The average HHS improvement was 3.00±21.86. The combined excellent and good rate at final follow-up (65.6%) was significantly higher than that before surgery (34.5%) (P=0.000). On radiographic evaluation, 103 (74.1%) hips remained stable, while 36 (25.9%) had femoral head collapse or aggravation of ONFH. THA was performed in 18 hips. Thus, the overall femoral head survival rate was 87.05% (121/139). The success rate was adversely affected by JIC type, but not by aetiology, age, or ARCO stage. The only complication was a subtrochanteric fracture in one patient.ConclusionThe SDBS effectively delays or even terminates the progression of ONFH, especially type B and C1. The SDBS is a good option for early-stage ONFH.

Comparative Analysis of Efficiency of the Machine Learning Methods for Gesture Recognition Using Double-Channel Electromyography

The paper is devoted to the efficiency analysis of the machine learning methods for gesture recognition, which are applied to the surface double-channel electromyography data. The comparative analysis was conducted for recognition of eight types of palm movements. The results of the analysis lead to conclusion that it is necessary to consider the muscle groups’ location for better recognition accuracy and the increase of the number of considered gestures.