SOC dynamic power management using artificial neural network

Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic system by selectively shutting down idle components. In this article we try to introduce back propagation network and radial basis network into the research of the system-level power management policies. We proposed two PM policies-Back propagation Power Management (BPPM) and Radial Basis Function Power Management (RBFPM) which are based on Artificial Neural Networks (ANN). Our experiments show that the two power management policies greatly lowered the system-level power consumption and have higher performance than traditional Power Management(PM) techniques-BPPM is 1.09-competitive and RBFPM is 1.08-competitive vs. 1.79, 1.45, 1.18-competitive separately for traditional timeout PM, adaptive predictive PM and stochastic PM.

SOI-based 16x16 thermo-optic waveguide switch matrix

A 16 x 16 thermo-optic wavelenght switch matrix has been designed and febricated on silicon-on-insulator wafer. For reducing device lenght, blocking switch matrix configuration is chosen. The building block of a matix is a 2 x 2 cell with Mach-Zehnder interferometer configuration, where a multi-mode interferometer serves as splitters/combiners. Spot size converters and isolating grooves are integrated on the same chip to reduce loss and power consumption. Average power consumption of the switch cell is 220 mW. The switching time of a switch cell is less than 3 mu s.

Temperature and power-density-dependent inter-shell energy states in InAs/GaAs quantum dots

We have investigated the evolution of exciton state filling in InAs/GaAs quantum dot (QD) structures as a function of the excitation power density by using rnicro-photoluminescence spectroscopy at different temperatures. In addition to the emission bands of exciton recombination corresponding to the atom-like S, P and D, etc. shells of QDs, it was observed that some extra states V between the S and P shells, and D' between the P and D shells appear in the spectra with increasing number of excitons occupying the QDs at a certain temperature. The emergence of these inter-shell excitonic levels is power density and temperature dependent, which is an experimental demonstration of strong exciton-exciton exchange interaction, state hybridization, and coupling of a multi-exciton system in QDs. (c) 2006 Elsevier B.V. All rights reserved.

Getting high-efficiency photoluminescence from Si nanocrystals in SiO2 matrix

Silicon nanocrystals in SiO2 matrix are fabricated by plasma enhanced chemical vapor deposition followed by thermal annealing. The structure and photoluminescence (PL) of the resulting films is investigated as a function of deposition temperature. Drastic improvement of PL efficiency up to 12% is achieved when the deposition temperature is reduced from 250 degreesC to room temperature. Low-temperature deposition is found to result in a high quality final structure of the films in which the silicon nanocrystals are nearly strain-free, and the Si/SiO2 interface sharp. The demonstration of the superior structural and optical properties of the films represents an important step towards the development of silicon-based light emitters. (C) 2002 American Institute of Physics.

Dynamic power management approaches based on neural network

Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic system. by selectively shutting down idle components. In this article we try to introduce back propagation network and radial basis network into the research of the system-level policies. We proposed two PAY policies-Back propagation Power Management (BPPM) and Radial Basis Function Power management (RBFPM) which are based on Artificial Neural Networks (ANN). Our experiments show that the two power management policies greatly lowered the system-level power consumption and have higher performance than traditional Power Management(PM) techniques-BPPM is 1.09-competitive and RBFPM is 1.08-competitive vs. 1.79,145,1.18-competitive separately for traditional timeout PM, adaptive predictive PM and stochastic PM.

SOC dynamic power management using artificial neural network

Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic system by selectively shutting down idle components. In this article we try to introduce back propagation network and radial basis network into the research of the system-level power management policies. We proposed two PM policies-Back propagation Power Management (BPPM) and Radial Basis Function Power Management (RBFPM) which are based on Artificial Neural Networks (ANN). Our experiments show that the two power management policies greatly lowered the system-level power consumption and have higher performance than traditional Power Management(PM) techniques-BPPM is 1.09-competitive and RBFPM is 1.08-competitive vs. 1.79 . 1.45 . 1.18-competitive separately for traditional timeout PM . adaptive predictive PM and stochastic PM.

SOI-based thermo-optic waveguide switch matrix with spot size converters

A SOI-based thermo-optic waveguide switch matrix worked at 1.55 mu m, integrated with spot size converters is designed and fabricated for the first time. The insertion loss and polarization dependent loss are less than 13dB and 2dB, respectively. The extinction ratio is larger than 19dB. The response time is less than 5 mu s and the power consumption of the switch cell is about 200mW.

Temperature and power-density-dependent inter-shell energy states in InAs/GaAs quantum dots

We have investigated the evolution of exciton state filling in InAs/GaAs quantum dot (QD) structures as a function of the excitation power density by using rnicro-photoluminescence spectroscopy at different temperatures. In addition to the emission bands of exciton recombination corresponding to the atom-like S, P and D, etc. shells of QDs, it was observed that some extra states V between the S and P shells, and D' between the P and D shells appear in the spectra with increasing number of excitons occupying the QDs at a certain temperature. The emergence of these inter-shell excitonic levels is power density and temperature dependent, which is an experimental demonstration of strong exciton-exciton exchange interaction, state hybridization, and coupling of a multi-exciton system in QDs. (c) 2006 Elsevier B.V. All rights reserved.

On detection wavelength and electron-hole wave function overlap of type II InAs/In-x Ga1-x Sb superlattice infrared photodetector

In this paper, the detection wavelength and the electron-hole wave function overlap of InAs/IrxGa1-xSb type II superlattice photodetectors are numerically calculated by using the envelope function and the transfer matrix methods. The band offset is dealt with by employing the model solid theory, which already takes into account the lattice mismatch between InAs and InxGa1-xSb layers. Firstly, the detection wavelength and the wave function overlap are investigated in dependence on the InAs and InxGa1-xSb layer thicknesses, the In mole fraction, and the periodic number. The results indicate that the detection wavelength increases with increasing In mole fraction, InAs and InxGa1-xSb layer thicknesses, respectively. When increasing the periodic number, the detection wavelength first increases distinctly for small periodic numbers then increases very slightly for large period numbers. Secondly, the wave function overlap diminishes with increasing InAs and InxGa1-xSb layer thicknesses, while it enhances with increasing In mole fraction. The dependence of the wave function overlap on the periodic number shows the same trend as that of the detection wavelength on the periodic number. Moreover, for a constant detection wavelength, the wave function overlap becomes greater when the thickness ratio of the InAs over InxGa1-xSb is larger.

Deterministic Column-Based Matrix Decomposition

In this paper, we propose a deterministic column-based matrix decomposition method. Conventional column-based matrix decomposition (CX) computes the columns by randomly sampling columns of the data matrix. Instead, the newly proposed method (termed as CX_D) selects columns in a deterministic manner, which well approximates singular value decomposition. The experimental results well demonstrate the power and the advantages of the proposed method upon three real-world data sets.

The scattering matrix method for quantum waveguides

A scattering matrix method for investigating the electron transport in quantum waveguides is presented. By dividing the structure into a number of transverse slices, the global scattering matrix is obtained by the composition of the individual scattering matrices associated with each interface. Complicated geometries and inhomogeneous external potentials are included in the formulation. It is shown that the proposed scattering matrix method possesses many advantages over the traditional mode-matching and transfer matrix methods, especially in treating the electron wave propagation in complicated geometries. Justification for the method is provided by the unitarity of the calculated scattering matrix, and the consistency of the results with those obtained by the recursive Green's function method.

A transfer matrix approach to conductance in quantum waveguides

A transfer matrix approach is presented for the study of electron conduction in an arbitrarily shaped cavity structure embedded in a quantum wire. Using the boundary conditions for wave functions, the transfer matrix at an interface with a discontinuous potential boundary is obtained for the first time. The total transfer matrix is calculated by multiplication of the transfer matrix for each segment of the structure as well as numerical integration of coupled second-order differential equations. The proposed method is applied to the evaluation of the conductance and the electron probability density in several typical cavity structures. The effect of the geometrical features on the electron transmission is discussed in detail. In the numerical calculations, the method is found to be more efficient than most of the other methods in the literature and the results are found to be in excellent agreement with those obtained by the recursive Green's function method.

Design and Fabrication of Thermo-Optic 4×4 Switching Matrix in Silicon-on-Insulator

A rearrangeable nonblocking thermo-optic 4×4 switching matrix,which consists of five 2×2 multimode interference-based Mach-Zehnder interferometer(MMI-MZI) switch elements，is designed and fabricated.The minimum and maximum excess loss for the matrix are 6.6 and 10.4dB,respectively.The crosstalk in the matrix is measured to be between －12 and －19.8dB.The switching speed of the matrix is less than 30μs.The power consumption for the single switch element is about 330mW.

Optimal distance from high power laser diode to cylindrical microlens in a coupling system

In a practical coupling system, a cylindrical microlens is used to collimate the emission of a high powerlaser diode (LD) in the dimension perpendicular to the junction plane. Using passive alignment, the LD isplaced in the focus of the cylindrical microlens generally, regardless of the performance of the multimodeoptical fiber and the LD. In this paper, a more complete analysis is arrived at by ray-tracing technique,by which the angle θ of the ray after refraction is computed as a function of the angle θo of the ray beforerefraction. The focus of the cylindrical microlens is not always the optimal position of the LD. In fact, inorder to achieve a higher coupling efficiency, the optimal distance from the LD to the cylindrical microlensis dependent on not only the radius R and the index of refraction n of the cylindrical microlens, but alsothe divergence angle of the LD in the dimension perpendicular to the junction plane and the numericalaperture (NA) of the multimode optical fiber. The results of this discussion are in good agreement withexperimental results.

Computer simulation of the damage evolution of fiber-reinforced polypropylene-matrix composites with matrix defects

The damage evolution of fiber-reinforced polypropylene-matrix composites with matrix defects was studied via a Monte Carlo technique combined with a finite element method. A finite element model was constructed to predict the effects of various matrix defect shapes on the stress distributions. The results indicated that a small matrix defect had almost no effect on fiber stress distributions other than interfacial shear stress distributions. Then, a finite element model with a statistical distribution of the fiber strength was constructed to investigate the influences of the spatial distribution and the volume fraction of matrix defects on composite failure. The results showed that it was accurate to use the shear-lag models and Green's function methods to predict the tensile strength of composites even though the axial stresses in the matrix were neglected.