Search

Weather forecasts are essential to aviation safety. Unreliable forecasts not only cause problems to pilots and air traffic controllers, but also lead to aviation accidents and incidents. This study develops a long short-term memory (LSTM) integrating both multiple linear regression and the Pearson’s correlation coefficients to improve forecasting. A numerical dataset of 10 weather features (sea pressure, temperature, dew point temperature, relative humidity, wind speed, wind direction, sunshine rate, global solar radiation, visible mean, and cloud amount) is applied on every calendar day in a year to train and validate the LSTM for temperature forecasting. It is shown that data standardization is necessary to rescale the data to improve training convergence and reduce training time. In addition, feature selection by multiple linear regression and by Pearson’s correlation coefficients are shown effective to the forecast accuracy of the LSTM. By selecting only the sensitive features (sea pressure, dew point temperature, relative humidity and relative humidity), the temperature forecasting errors can be reduced from RMSE 4.0274 to 2.2215 and MAPE 23.0538% to 5.0069%. LSTM deep learning with data standardization and feature selection is effective in forecasting for aviation safety.

The performance of a thermoelectric energy generator (TEG) can be improved by having thermocouple materials with higher Seebeck coefficient, thermocouple configuration with better thermal isolation, and thin-film layer deposition/etching with effective electrical insulation. A TEG design by the 1P6M (1 polysilicon layer and 6 metal layers) standard CMOS process in semiconductor foundry service (TSMC) is proposed in this work to achieve better energy harvesting performance. The in-plane thermocouple has higher Seebeck effect by the single polysilicon layer at different doping rates. The TEG configuration is with double cavity design to seal the thermocouples above-and-below for better thermal isolation, and the deposition layers are in mask-less design for better electrical insulation. Measurement results of a $5\times5$ mm2 TEG chip show that higher thermoelectric conversion, better thermal isolation and electrical insulation can be realized. The voltage factor 15.604 V/cm2K and power factor $0.105~\mu \text{W}$ /cm2K2 are about $5.40\times $ and $2.34\times $ , respectively, of those of the previous work by the CMOS process with two polysilicon layers. This design, implementation, and experiment have achieved the best performance in all TEGs by CMOS process in semiconductor foundry.

A thermoelectric energy generator (TEG) is to convert the temperature difference between the hot and cold junctions into electrical energy by Seebeck effect. This work proposes a TEG design implemented by standard CMOS process to improve energy harvesting performance and to expedite the fabrication process. The TEG has double cavity with the upper cavity of $5.2~\mu \text{m}$ and the lower cavity of $10~\mu \text{m}$ in depth, which are created by dry etching post process for thermal isolation to improve energy conversion. The size and geometry of the thermocouples are determined to match their thermal/electrical resistance for optimal performance. A $5\times5$ mm2 TEG chip with four $2\times2$ mm2 cells, two in electrical series and another two in parallel, is implemented by CMOS process in semiconductor foundry service (TSMC D35 2P4M). Experiments validate that the voltage/power factor of the cells in the TEG chip agree very well. In simulation, the double cavity design achieves the performance 30% higher than that in previous work with only lower cavity. In experimental measurement, the TEG has voltage factor 2.889 V/cm2K and power factor $0.0450~\mu \text{W}$ /cm2K2. This is by far best TEG performance by matured, standard CMOS process in semiconductor foundry.

Thermoelectric energy generator (TEG) is a device which is used to convert temperature difference between the hot and cold junctions into electrical energy. This paper aims at developing a TEG chip of small layout 5 mm2 having three 2.01 mm $\times0.72$ mm cells in a total of 10 080 thermocouples by semiconductor foundry service (TSMC) for the advantages in batch production, low cost, and scalability. The design has the unique features to improve TEG performance by having (1) thin-film polycrystalline silicon-germanium Si0.9Ge0.1 layer for higher thermoelectric conversion efficiency, (2) isolation cavity for better thermal insulation, and (3) structural support for improved thermocouple dimensional stability. It is shown that the TEG chip can deliver power factor $0.125~\mu $ W/cm2K2 and voltage factor 25.91 V/cm2 K. The experimental verification demonstrates that the 5 mm2 chip has output power $2.083~\mu $ W and voltage 3.4 V at 20 K temperature gradient. This is by far the TEG design with best thermoelectric performance implemented by the semiconductor foundry service.

Helicobacter pylori (Hp) infection and overexpression of hepatoma-derived growth factor (HDGF) are involved in gastric carcinogenesis. However, the relationship between Hp-induced gastric diseases and HDGF upregulation is not yet completely clear. This study aimed to elucidate the role of HDGF in Hp-induced gastric inflammation and carcinogenesis. HDGF expression in gastric biopsy and serum from patients was analyzed by immunohistochemical and ELISA analysis, respectively. Hp and gastric cells coculture system was employed to delineate the mechanism underlying HDGF overexpression during Hp infection. The gastric pathologies of wild type and HDGF knockout mice after Hp infection were investigated by immunohistochemical, immunoblot, and immunofluorescence analyses. HDGF level was significantly elevated in patients with Hp infection or intestinal metaplasia (IM, a precancerous lesion), and HDGF overexpression was positively correlated with Hp load, IM, and neutrophil infiltration in gastric biopsy. Consistently, patients with Hp infection or IM had significantly higher serum HDGF level. By using coculture assay, Hp infection led to HDGF upregulation and secretion in gastric cells. In mice model, HDGF ablation significantly suppressed the Hp-induced neutrophil infiltration and inflammatory TNF-α/COX-2 signaling, thereby relieving the tissue damage in stomach. This was further supported by that recombinant HDGF (rHDGF) stimulated the differentiation/chemotaxis of cultured neutrophils and oncogenic behaviors of gastric cells. Time series studies showed that Hp infection elicited an inflammatory TNF-α/HDGF/COX-2 cascade in stomach. HDGF secretion by Hp infection promotes the neutrophils infiltration and relays Hp-induced inflammatory signaling. Thus, HDGF may constitute a novel diagnostic marker and therapeutic target for Hp-induced gastritis and carcinogenesis.

With the advances of thin-film process technology in foundry services, recent development is to design thermoelectric generator (TEG) with in-plane thermocouple of polysilicon or polycrystalline silicon germanium by complementary metal oxide semiconductor (CMOS) or BiCMOS process. However, the small thermal resistance and adversely large electrical resistance in a thermocouple of micron dimension lead to small temperature gradient and excessive heat loss, hence resulting in poor harvester performance. The optimal design of thermocouple dimension is therefore critical to the TEG performance. A model of the effect of thermocouple dimension on the TEG performance is developed. For the TEG design by TSMC CMOS process with P- and N-thermoleg of thickness $0.275~\mu \text{m}$ and $0.180~\mu \text{m}$ , respectively, the model shows that the optimal thermocouple is $60\,\,\mu \text {m} \times 2\,\,\mu \text{m}$ (length $\times $ width). For design by TSMC BiCMOS process with P- and N-thermoleg of thickness $0.380~\mu \text{m}$ and $0.200~\mu \text{m}$ , respectively, it is $45\,\,\mu \text {m} \times 2\,\,\mu \text{m}$ . The numerical simulation and experimental verification on the latter show that better TEG performance can be achieved by optimal thermocouple dimension to match the electrical and thermal resistances.

The design of thermoelectric energy harvester in hybrid configuration by CMOS process has been known desirable. Thermal isolation cavity on the top and/or bottom of a thermocouple to increase its thermal resistance is necessary to achieve better harvester performance. However, the thermocouple becomes suspended and may suffer from poor dimension stability on wearable device applications. This paper aims at improving the dimension stability of thermoelectric energy harvester by a structural support design. A harvester with a thermal isolation cavity of 10 $\mu \text{m}$ in depth and a structural support of 4 $\mu \text{m}$ in length is developed to minimize heat leakage and to provide the dimension stability of thermocouples. Compared with many previous works, the harvester design implemented by BiCMOS standard process (TSMC 0.18- $\mu \text{m}$ SiGe 3P6M) is shown to achieve a superior power factor of 0.125 $\mu \text{W}$ /cm2K2 and a voltage factor of 23.53 V/cm2K.

Burn-related neuropathy is common and often involves pain, paresthesia, or muscle weakness. Irisin, an exercise-induced myokine after cleavage from its membrane precursor fibronectin type III domain-containing 5 (FNDC5), exhibits neuroprotective and anti-inflammatory activities. A rat model of third-degree burn on the right hind paw was used to investigate the therapeutic role of irisin/FNDC5. Rats received burn injury and were treated with intrathecal recombinant adenovirus containing the irisin sequence (Ad-irisin) at 3 weeks postburn. One week later, mechanical allodynia was examined. The expression of irisin in cerebrospinal fluid (CSF) was detected. Ipsilateral gastrocnemius muscle and lumbar spinal cord were also obtained for further investigation. Furthermore, the anti-apoptotic effect of recombinant irisin in SH-SY5Y cells was evaluated through tumor necrosis factor alpha (TNF&alpha, ) stimulus to mimic burn injury. We noted intrathecal Ad-irisin attenuated pain sensitization and gastrocnemius muscle atrophy by modulating the level of irisin in CSF, and the expression of neuronal FNDC5/irisin and TNF&alpha, in the spinal cord. Ad-irisin also ameliorated neuronal apoptosis in both dorsal and ventral horns. Furthermore, recombinant irisin attenuated TNF&alpha, induced SH-SY5Y cell apoptosis. In summary, irisin attenuated allodynia and muscle wasting by ameliorating neuroinflammation-induced neuronal apoptosis.

Hepatitis is an important health problem worldwide. Novel molecular targets are in demand for detection and management of hepatitis. Hepatoma-derived growth factor (HDGF) has been delineated to participate in hepatic fibrosis and liver carcinogenesis. However, the relationship between hepatitis and HDGF remains unclear. This study aimed to elucidate the role of HDGF during hepatitis using concanavalin A (ConA)-induced hepatitis model. In cultured hepatocytes, ConA treatment-elicited HDGF upregulation at transcriptional level and promoted HDGF secretion while reducing intracellular HDGF protein level and cellular viability. Similarly, mice receiving ConA administration exhibited reduced hepatic HDGF expression and elevated circulating HDGF level, which was positively correlated with serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels. By using HDGF knockout (KO) mice, it was found the ConA-evoked cell death was prominently alleviated in KO compared with control. Besides, it was delineated HDGF ablation conferred protection by suppressing the ConA-induced neutrophils recruitment in livers. Above all, the ConA-mediated activation of tumor necrosis factor-α (TNF-α)/interleukin-1β (IL-1β)/interleukin-6 (IL-6)/cyclooxygenase-2 (COX-2) inflammatory signaling was significantly abrogated in KO mice. Treatment with recombinant HDGF (rHDGF) dose-dependently stimulated the expression of TNF-α/IL-1β/IL-6/COX-2 in hepatocytes, further supporting the pro-inflammatory function of HDGF. Finally, application of HDGF antibody not only attenuated the ConA-mediated inflammatory cascade in hepatocytes, but also ameliorated the ConA-induced hepatic necrosis and AST elevation in mice. In summary, HDGF participates in ConA-induced hepatitis via neutrophils recruitment and may constitute a therapeutic target for acute hepatitis.

Methods based on discrete cosine transform (DCT) have been proposed for digital watermarking of audio signals; however, the watermark is often vulnerable to data compression and signal processing. This paper presents an effective audio watermarking method by energy averaging of DCT coefficients such that an audio signal with watermark is robust to data processing. The method is to divide an audio signal into segments by three parameters defining the segment length, the segment sequence of watermark location, and the frequency range of DCT coefficients for watermark location. An error correcting code is also integrated to improve audio signal quality after watermarking. Experimental results show that the method is robust to data compression and many other kinds of signal processing. No original signal is required for decoding the watermark. Comparison of watermarking performance with a recent work validates that the watermarking method has better audio quality and higher robustness.

With the advent of semiconductor foundry services, thermoelectric energy generator (TEG) by CMOS/BiCMOS process has been preferable for production yield, cost effectiveness, and device scalability. The thermocouples in semiconductor TEG often have thermal isolation cavity to reduce heat loss. Both die- and wafer-packaging are proposed in this work to seal the cavity, prevent moisture/dirt from contamination, and avoid short circuit condition in the thermocouples for TEG applications in wearable devices. For a TEG design by 3P6M (3 polysilicon layers and 6 metal layers) BiCMOS process, the die-level packaging by aluminum film of thickness 50 μm over the TEG is shown to be effective, though there is about 23 % temperature gradient loss by the packaging. The voltage factor is 18.22 V/cm2K, about 77 % of that before packaging. For a TEG design by 1P6M CMOS process, the wafer-level packaging by bonding two identical TEG device wafers is shown to have about 22 % loss in temperature gradient. Yet the TEG performance is increased significantly on the same footprint because of doubling area density, and hence doubling both the power factor and voltage factor. The power factor and voltage factor is 0.164 μW/cm2 K2 and 24.32 V/cm2K, respectively, on the TEG in series circuit after packaging.

This paper proposes an image watermarking method based on the fast discrete cosine transform (DCT) algorithm for implementation in digital signal processor. A digital watermark can be effectively embedded and efficiently extracted without the host image. The keys in watermarking process include four frequency coefficients in DCT, two random permutation vectors, and a quantization matrix for normalizing the watermark and the host image. The fast DCT algorithm has been shown to reduce the complexity of two-dimensional image transformation so that embedding/decoding an image watermark can be completed in real time within 0.33 seconds. The quality of both watermarked image and extracted (retrieved) watermark remains excellent. It is shown that the watermarking method is efficient in and robust to data cropping, transmission loss, and compression/decompression.

Discrete cosine transform (DCT) has been an international standard in Joint Photographic Experts Group (JPEG) format to reduce the blocking effect in digital image compression. This paper proposes a fast discrete cosine transform (FDCT) algorithm that utilizes the energy compactness and matrix sparseness properties in frequency domain to achieve higher computation performance. For a JPEG image of8×8block size in spatial domain, the algorithm decomposes the two-dimensional (2D) DCT into one pair of one-dimensional (1D) DCTs with transform computation in only 24 multiplications. The 2D spatial data is a linear combination of the base image obtained by the outer product of the column and row vectors of cosine functions so that inverse DCT is as efficient. Implementation of the FDCT algorithm shows that embedding a watermark image of 32 × 32 block pixel size in a 256 × 256 digital image can be completed in only 0.24 seconds and the extraction of watermark by inverse transform is within 0.21 seconds. The proposed FDCT algorithm is shown more efficient than many previous works in computation.

Lane keeping systems for a keeping a vehicle in the desired lane is key to advanced driving assistance system in autonomous vehicles. This paper presents a cost-effective image sensor with efficient processing algorithm for lane detection and lane control applications to autonomous delivery systems. The algorithm includes (1) lane detection by inverse perspective mapping and random sample consensus parabola fitting and (2) lane control by pure pursuit steering controller and classical proportional integral speed controller based on a nonholonomic kinematic model. The image sensor experiments conducted on a 1/10 scale model car maneuvering in a straight&ndash, curve&ndash, straight lane validate the better processing performance before, during, and after the turning section over previous work. The image sensor with the processing algorithm achieves the average lane detection error within 5% and maximum cross-track error within 9% in real-time. The development shall pave the way to cost-effective autonomous delivery systems.

This paper proposes a model having both linear and nonlinear system dynamics by integrating both autoregressive integrated moving average (ARIMA) model and artificial neural network (ANN) model to simulate electrical energy supply inherent with strong seasonal and periodic characteristics in power system. Accurate electrical load forecast becomes possible by the integrated model for the ARIMA is effective to electricity load time series inherent with seasonal fluctuations as well as strong 7-day (per week) periodic characteristics. By using the input of historical daily electricity load data, weather data, and holiday effect variables, the integrated model is shown to be more accurate than the ANN model, the ARIMA model, the classical ARIMA-ANN model, and other well-known methods in the prediction and the forecast of electrical load in normal summer week, normal winter week, 3/4-day holiday week, long holiday week, and extreme weather week.

Safe path planning for obstacle avoidance in autonomous vehicles has been developed. Based on the Rapidly Exploring Random Trees (RRT) algorithm, an improved algorithm integrating path pruning, smoothing, and optimization with geometric collision detection is shown to improve planning efficiency. Path pruning, a prerequisite to path smoothing, is performed to remove the redundant points generated by the random trees for a new path, without colliding with the obstacles. Path smoothing is performed to modify the path so that it becomes continuously differentiable with curvature implementable by the vehicle. Optimization is performed to select a “near”-optimal path of the shortest distance among the feasible paths for motion efficiency. In the experimental verification, both a pure pursuit steering controller and a proportional–integral speed controller are applied to keep an autonomous vehicle tracking the planned path predicted by the improved RRT algorithm. It is shown that the vehicle can successfully track the path efficiently and reach the destination safely, with an average tracking control deviation of 5.2% of the vehicle width. The path planning is also applied to lane changes, and the average deviation from the lane during and after lane changes remains within 8.3% of the vehicle width.

Standard complementary-metal-oxide semi-conductor (CMOS) process for the advantages in batch production, low cost, and mostly importantly, scalability has been applied to thermoelectric generator (TEG) designs, where silicon materials are preferable for monolithic circuit integration. In this work, the BiCMOS process is found to be a perfect platform for TEG designs to achieve higher performance by the polysilicon-germanium thin film layer. Simulations of the TEG designs by 0.35 μm SiGe 3P3M and 0.18 μm SiGe 3P6M BiCMOS processes (TSMC) show that the power factors are respectively 0.242 and 0.125 μW/cm2 K2, and the voltage factors are 10.04 and 25.91 V/cm2 K. Both are shown to be superior to all micro TEGs by semiconductor process in the open literature. Experimental verifications confirm the simulation results and validate the effectiveness of TEG designs by the polysilicon-germanium thin film layer in BiCMOS process.

This work aims at integrating fuzzy logic and quality function deployment (QFD) technique to determine the critical chain in a project with uncertainties. The degree of fuzziness of every activity in a project is defined and calculated by QFD so as to (1) evaluate the effects of project uncertainties, (2) select adaptively the fuzzy logic model, and (3) calculate the buffer time of the critical chain. The example of a project with 11 activities facing 6 uncertainties: project time limit, activity begin time, budget, manpower, technical difficulty, and facility requirement is applied to demonstrate the effectiveness of the model integrating fuzzy logic and QFD. Analysis shows that project scheduling by this proposed integration is superior to that by Project Evaluation and Review Technique (PERT) and by the conventional critical chain method. Not only the scheduling is more efficient, but also the buffer time in critical chain can be systematically determined to accommodate the uncertainties.

Aim Despite the great advances and excellent outcomes of liver transplantation (LT), small-for-size (SFS) graft syndrome is a life-threatening complication that remains to be overcome. In the present study, we investigated the therapeutic effect of combined treatment with granulocyte colony-stimulating factor (G-CSF) and a dipeptidyl peptidase IV (DPP-IV) inhibitor on SFS liver graft syndrome. Methods The transplantation of small-sized Lewis donor livers into green fluorescent protein (GFP) transgenic Wistar rats was performed and the recipients were randomly assigned to one of four groups (without treatment, DPP-IV inhibitor treatment, G-CSF treatment and G-CSF/DPP-IV inhibitor combination). Recombinant human G-CSF was injected s.c. at a dose of 2 μg/kg per day starting 5 days prior to transplantation. G-CSF was combined with the p.o. administration of a DPP-IV inhibitor (2 mg/kg per day) after transplantation until the end of the observation period. Results The post-transplant survival and liver function of rats treated with G-CSF/DPP-IV inhibitor combination therapy were significantly improved with an increased number of recipient-derived GFP positive cells into the liver grafts. A confocal microscopy study showed cytokeratin (CK)-18 and GFP positive hepatic progenitor cells in the parenchyma of the liver allografts. Untreated rats and rats treated with either G-CSF or DPP-IV inhibitor did not exhibit the prolonged survival and had less GFP and CK-18 positive cells in the liver grafts after SFS LT. Conclusion Our results suggest that combined treatment with G-CSF and DPP-IV inhibitor may synergistically induce migration and differentiation of recipient-derived stem cells into the hepatic progenitor cells, resulting in the amelioration of SFS liver graft syndrome.

Fiber-reinforced composite structures can be tailored to desire mechanical properties and to embed microstrip antenna in aerospace applications. The electromagnetic characteristics of a microstrip antenna on isotropic and uniaxial substrates have been known, but that embedded in composite laminated substrates remain unavailable to date. This work aims at analyzing the performance of microstrip antenna embedded in composite laminated substrates by spectral domain analysis. Parameter studies are conducted to investigate the effects of the substrate's dielectric constants, the fiber directions the orientation between the antenna and the laminate layers, and the stacking sequence on antenna's resonant frequency and radiation pattern. The antenna size when embedded in composite laminated substrates is larger than that when attached on isotropic substrates, or conversely, the resonant frequency will deviate lower if assuming the substrate as isotropic. The far-field pattern in composite laminated substrates is more 'directional' than that in isotropic substrates. The antenna gain in the substrates of symmetric stacking with ±45i¾? fiber direction is 20dB better than that in isotropic substrates in some elevations. Copyright © 2014 John Wiley & Sons, Ltd.

Composite structures with desire mechanical properties can be tailored to embed microstrip antenna for reducing aerodynamic disturbance in aerospace systems. This work aims at developing a spectral domain model to analyze the effect of dielectric overlay on the performance of microstrip antenna embedded in composite laminated substrates. Analysis shows that antenna performance depends not only on the dielectric overlay but also on the laminate layer's electromagnetic property, fiber direction, and stacking sequence. The common belief that the antenna size will be reduced because of the increase of equivalent permittivity by the dielectric overlay is no longer valid in composite laminated substrates. Of the eight substrates studied, overlay can, in general, reduce antenna radiation efficiency; however, it is found to have favorable effect in substrates with some selected electromagnetic properties and stacking sequence. For example, removing the antenna overlay in [45/-45/-45/45/45/-45] substrate can be counterproductive. Copyright © 2014 John Wiley & Sons, Ltd.

Meander line antenna has been considered desirable on flight vehicles to reduce drag and minimize aerodynamic disturbance; however, the antenna design and performance analysis have made mostly by trial-and-error. An inductor model by simulating the meander line sections as electrical inductors and the interconnecting radiation elements as a quasi-monopole antenna is developed to analyze the antenna performance. Experimental verifications of the printed meander line antennas embedded in composite laminated substrates show that the inductor model is effective to design and analyze. Of the 4 antennas tested, the discrepancy of resonant frequency in simulation and experiment is within 4.6%.

With the trend toward taller and more flexible building structures, a mass-damper shaking table system has been considered as means for vibration suppression to external excitation and disturbances in recent years. However, there are few researches on the control of nonlinear structure using active mass damper (AMD) under earthquake excitation, especially for high-rise building. In this work, a multilayer feedforward neural network with the modified Newton method, similar to BFGS algorithm, is developed for vibration suppression of a building structure. The benchmark tests show that the modified Newton method is superior to many conventional ones: steepest descent, steepest descent with adaptive learning rate, conjugate gradient, and Newton-based methods and is suitable to small network in engineering applications. Experimental results show that an AMD system combined with the modified Newton method remains effective for building structure vibration suppression under free vibration, forced vibration and Ji-Ji Earthquake (Sep. 21, 1999) excitation.

Computer integrated manufacturing, CIM, has attracted many interests both in academic research and practical implementation. Design and implementation of a CIM project involves the massive modeling, simulation, and verification of production operations, machines, and operators. Overhead hoist transporter, OHT, is one of the effective tools for the automated material handling system that often applied in CIM projects to ultimate the utilization of footprint of a factory. This paper presents a case study for the OHT system design in a CIM project for automated food processing production line. The effectiveness of the number of OHTs installed for the production line and the proper operation scenarios for different production planning are studied. Modeling and simulation results are performed to suggest an efficient OHT system design for the CIM project.

We developed a model that integrates fuzzy logic, the analytic hierarchy process (AHP), and quality function deployment (QFD) to evaluate the importance weighting of the technical factors in aviation safety, as defined by the International Air Transport Association. AHP evaluates the technical factors using pairwise comparisons, and QFD categorizes the relationship between the criteria and subcriteria using expert knowledge. Based on a questionnaire given to aviation professionals, the model shows that the importance weightings of technical factors “Extensive engine failure, uncontained engine fire,” “Design, manufacture,” and “Engine overheat, propeller failure” are most significant. The result also shows that the over-valued and/or under-valued factors predicted by the conventional AHP model can be better described by the integrated fuzzy AHP and QFD models.

Microstrip antennas have the advantages of light weight, low profile, and conformal to the attached surface with antenna feed line. This work presents the design of transparent microstrip antennas by In2O3:Sn thin film on glass and polycarbonate substrates. The transparent conducting thin films of 21–300 nm thickness deposited by magnetron sputtering are measured by X-ray diffraction for microstructure, 4-point probe for electrical resistance, and spectrometer for optical transmittance. Analyses show that the 2.4 GHz antenna can achieve 5.1 and 4.7 dB antenna gain on glass (1.2 mm) and polycarbonate (0.7 mm) substrate, respectively. Experimental verification on glass substrate shows that the antenna achieves 3.1 dB gain and 86% optical transmittance on 550 nm wave length.

The development of neuro-fuzzy systems by integrating neural networks and fuzzy systems is desired because such systems can adjust fuzzy membership functions and produce fuzzy inference rules by case-learning without the need for experts or experiments. It has been applied to various fields, but there has been no detailed study of the various neuro-fuzzy models applicable to rule generation. In this paper, an experimentally verified five-layer and three-phase network is presented, which shows the effectiveness with which the neuro-fuzzy system automatically determines membership functions and selects activation fuzzy rules using both system identification and vibration control examples in engineering applications.

Noise estimation is an important process in digital imaging systems. Many noise reduction algorithms require their parameters to be adjusted based on the noise level. Filter-based approaches of image noise estimation usually were more efficient but had difficulty on separating noise from images. Block-based approaches could provide more accurate results but usually required higher computation complexity. In this work, a design framework for combining the strengths of filter-based and block-based approaches is presented. Different homogeneity analyzers for identifying the homogeneous blocks are discussed and their performances are compared. Then, two well-known filters, the bilateral and the non-local mean, are reviewed and their parameter settings are investigated. A new bilateral filter with edge enhancement is proposed. A modified non-local mean filter with much less complexity is also present. Compared to the original non-local mean filter, the complexity is dramatically reduced by 75% and yet the image quality is maintained.

A two-stage algorithm combining the advantages of adaptive genetic algorithm and modified Newton method is developed for effective training in feedforward neural networks. The genetic algorithm with adaptive reproduction, crossover, and mutation operators is to search for initial weight and bias of the neural network, while the modified Newton method, similar to BFGS algorithm, is to increase network training performance. The benchmark tests show that the two-stage algorithm is superior to many conventional ones: steepest descent, steepest descent with adaptive learning rate, conjugate gradient, and Newton-based methods and is suitable to small network in engineering applications. In addition to numerical simulation, the effectiveness of the two-stage algorithm is validated by experiments of system identification and vibration suppression.

A 3000 μm × 3000 μm × 500 μm in-plane, gap overlap, electrostatic comb-drive operating in charge-constrained cycle is developed to harvest ambient vibration. This is the smallest vibration harvester resonating at 105 Hz, implementable by standard CMOS process and compatible to System-on-Chip (SoC) design. The 2000 μm × 2200 μm × 450 μm (4.9 mg) vibrating mass is made of the CMOS handle layer so as to achieve low frequency operation. Design verifications show that, with the excitation input of 10 μm amplitude at 105 Hz, the harvester can generate an average output power of 0.0924 μW, or equivalent output density 0.021 μW/mm 3 . The design is the most area-efficient CMOS vibration energy harvester reported to date.

This work aims at improving the energy harvester performance by using low-dimensional thermoelectric materials. A micro-thermoelectric generator (μTEG) with quantum well-like thermocouples is developed by state-of-the-art CMOS (Complementary metal-oxide semiconductor) process. A relaxation-time model is applied to analyze the characteristic length of silicon germanium quantum well, and a thermal model is also applied to calculate the thermocouple size for optimal performance by matching the thermal/electrical resistance. Analysis based on TSMC 0.35 μm 3P3M (3-poly and 3-metal layers) BiCMOS process shows that the maximum power factor and voltage factor of a μTEG is 0.241 μW/cm 2 K 2 and 10.442 V/cm 2 K. Design implementation validates that the μTEG with 60 μm × 4 μm quantum well-like thermocouples (0.05 μm Si 0.9 Ge 0.1 quantum well on 0.300 μm P-thermoleg and 0.280 μm N-thermoleg) has the best performance compared with those reported in the literature.