Recently, the evolutions of wireless communication systems are growing rapidly to satisfy the personal communication requirements. Compact, small size, low cost, easy fabrication, and multi-function are the major developing trends among these modern wireless communication devices. In this research, high quality Al2O3 ceramic materials are used as a substrate to fabricate modifiable dual-mode bandpass filters with the advantages of light weight, small size, and high performance. Dual mode cavity and dielectric resonator filters are the mainstay of satellite communications. In this research, a class generation of planar dual mode filters are introduced which significantly offers the advantages of small size, light weight, and low cost. The proposed bandpass filters are first simulated on Al2O3 (er=9.4) substrate for microwave frequency, and the proposed structures consist one square-typed patch, one circle-typed patch, or one meander-typed patch. Each pattern is connected with two 50£[ microstrip line, one for input and another for output. Two 50£[ microstrip lines are connected to the central of the patch with two notches, and the patch will act as the elliptic-function bandpass filters. Before the fabrication of a bandpass filter, IE3D microwave simulation software is used to simulate the designed filters. With a metal mask, the filter patterns are fabricated on the Al2O3 substrate by using evaporation method. It is found that as the length of notches increase, the three of structures band-pass filters reveal better filtering effect, i.e., the smaller insertion loss (S21), and larger return loss (S11), and the perturbation in the patch facilitates coupling between two orthogonal modes within the resonator, i.e., the perturbation size can modify bandwidth and resonator frequency. After the measurement of the fabricated filters, it is found that the measured results will match the simulated results.

In recent years there has been increased research activity in the area of Func- tional Data Analysis. Methodology from finite dimensional multivariate analysis has been extended to the functional data setting giving birth to Functional ANOVA, Functional Principal Components Analysis, etc. In particular, some studies have pro- posed inferential techniques for various functional models that have connections to well known areas such as mixed-effects models or spline smoothing. The methodol- ogy used in these cases is computationally intensive since it involves the estimation of coefficients in linear models, adaptive selection of smoothing parameters, estimation of variances components, etc. This dissertation proposes a wide-ranging modeling framework that includes many functional linear models as special cases. Three widely used tools are con- sidered: mixed-effects models, penalized least squares, and Bayesian prediction. We show that, in certain important cases, the same numerical answer is obtained for these seemingly different techniques. In addition, under certain assumptions, an applica- tion of a Kalman filter algorithm is shown to improve the order of computations, by two orders of magnitude, for point and interval estimates (with n being the sample size). A functional data analysis setting is used to exemplify our results.

A polarization independent electrooptic add/drop tunable filter fabricated on a LiNbO3 substrate with a narrow -3 dB bandwidth (FWHM) of 1.56 nm operating in the 1.55 μm wavelength regime has been developed to meet the demands of fast tuning speed and increased channel capacity for dense wavelength division multiplexed (WDM) networks. The operation of the filter is based on passive polarization beam splitters and strain-induced phase-matched TE↔TM polarization mode converters. Extinction ratios as high as 20 dB for polarization beam splitters were achieved using zero-gap two-mode interference directional couplers with an opening angle of 0.55°. A tunable TE↔TM polarization mode converter with 98.2 % conversion efficiency was obtained using a strain-induced refractive index grating consisting of 765 parallel strips 10.5 μm wide in a strained SiO2 surface film with a spatial period of 21 μm. Thermal and electrooptic tuning of the polarization mode converters were examined. A polarization independent electrooptic add/drop tunable filter in which the fabrication parameters of the splitter and the mode converter were optimized was produced. Fiber-to-fiber insertion loss less than 6.3 dB was measured on a 62 mm long filter device. The spectral characteristics reveal a -3 dB bandwidth of 1.56 nm and nearest sidelobes about 12 dB below the center peak. A thermal tuning rate of -0.903 nm/°C was realized. Electrooptic tuning was also demonstrated. A tuning range of 14.08 nm with applied DC voltages from -80 V to +80 V was achieved indicating an electrooptic tuning rate of 0.086 nm/V. The filter performance for both TE and TM modes was examined and polarization independence of the spectral characteristics was confirmed.

In this thesis, a post-processing deblocking filter is proposed to reduce the blocking effects. This proposed algorithm is suitable for the current block-based video standards without any modification. The proposed algorithm uses Sobel operator and wavelet transform to accurately detect blocking effects at 4¡Ñ4 or 8¡Ñ8 blocking boundaries automatically. In the filtering stage, the proposed algorithm provides four filter modes to eliminate blocking effects at different color regions according to human color vision and color psychology analysis. Experimental results show that the proposed algorithm has better subjective and objective qualities for H.264/AVC reconstructed video sequences compared with several existing methods.

Abstract The purpose of this thesis is to design and fabricate optical waveguide reflectors by using vertical mirror reflector. In order to reflect the incident optical mode of multi-mode interference coupler, we fabricate the etched vertical reflective mirror surface with dry etch processes. We used the design of etched vertical mirror surfaces to bend the incident optical mode, and changed the length of multi-mode interference coupler for the power splitting ratio and high power output. In order to make total internal reflection in vertical reflective mirror, we used ICP-RIE dry etching process to fabricate vertical and smooth mirror. In this part of design the semiconductor optical filter, we design the 90o vertical mirror reflector and the length of MMI by using a 2¡Ñ2 multi-mode interference coupler to get the power splitter with coupling coefficient. We comprise the ring resonator by two vertical mirror reflector and bended waveguide to displace the cleaved of Fabry-Perot resonator in tradition. In fabrication process, we using dry etching (ICP-RIE) process that is the plasma system high density in all part of this paper. first, we defined the device pattern by using photo-lithography technique in our sample . Second, we etched ridge waveguide by using dry etching ICP-RIE method. In order to decrease the scattering loss, we deep etched bended waveguide and the reflector. Then, we etched the reflector mirror by ICP-RIE dry etching process to get smoother surface. Finally, we used polyimide to flatten the sides of the ridge waveguide and evaporated metal electrode. In the device characteristic, we get the waveguide loss in 35.68dB by Fabry-Perot resonator and a power splitter with 85 percent and 15 percent output. Finally, we could get a transmission frequency about 81GHz from optical transmission spectrum in the filter device.

This thesis presents an automatic tuning circuit that it is focused on compensation for the filter¡¦s frequency error resulting from the variation of fabrication process, supply voltage and temperature. We utilize a tunable operational transconductance amplifier and a capacitor to form a single-time constant circuit (STC). When we input a reference signal to this circuit, the output of STC circuit can produce a controllable delay time clock. The tuning circuit uses the constant delay time to tune the frequency of the filter. The design of the STC circuit is simple and it has less chip area. All circuits are designed by using the parameters of TSMC 0.35um mixed signal process, and the supply voltage is 3V. The simulation result shows that the filter¡¦s 3-dB frequency error can be controlled by less than 7% as the filter is under the condition of over a range of supply voltages(¡Ó10¢H), operating temperatures(-20 ¢Jto 70¢J ) and five models of SPICE model.

A laboratory-scale biotrickling filter (BTF, inner size 14cm ¡Ñ 14cm ¡Ñ 120cmH, packed with fern chips to a height of 100 cm and volume of 19.6 L) was used to test the feasibility of removing odorous compounds emitted from rubber-processing operations. In addition, granule activated carbon (GAC) adsorption and permanganate solution scrubbing were also tested in order to further reduce the odor intensity of vented gas from the BTF. Results indicated that with the operation conditions of EBRT (empty bed retention time) of 20-40 s for the gas through the packed space and VOCs of 5-50 ppm (as methane) for the influent gas to the BTF, approximately 96-97% of ketones (acetone and methyl ethyl ketone) and 50% of toluene and carbon disulfide in the influent gas could be removed. However, there was no significant removal for alkanes by the BTF. The overall VOC and odor removals were both around 80% at an EBRT of 23 s. Performances of GAC adsorption of the residual chemicals in the vented gas from the BTF were better than those of permanganate solution scrubbing. The overall VOC removal by the BTF-GAC system was around 90% and the overall odor (expressed as the dilution to threshold D/T value) removal was even better than that of VOC. A test indicated that D/T were 1303, 733, and 23, respectively, for the influent, BTF effluent, and GAC effluent, and the overall efficiency for the odor removal was 98.2%. It was estimated that the cost is around NT$ 70 for treating 1,000 m3 of the teat gas by the GAC. Efforts should be made by decreasing the cost by other alternative technologies.

The growing demand of portable electronic equipment and system-on-a-chip has been pushing the industry to design circuits with very low power supply voltage and low power consumption. The Hard Disk drive industry is looking for developments in the read channel chip to push the data rates to higher speed, along with a low voltage and low cost solution. Read channel requires high-speed linear phase filters to meet these objectives. The primary objective of this project is to design, layout, and characterize a 4th-order continuous-time equiripple linear phase filter with automatic tuning system. The main requirements for design are high speed, low group delay variations, good linearity and power efficiency. This filter features wide cut-off frequency 500MHz, which is far beyond the current state-of-the-art. The linear phase filter is based on Gm-C biquadratics. Higher speed has been achieved by minimizing the parasitics and a complementary input stage OTA. A common mode feedback (CMFB), which ensures stability at such high frequencies, has also been designed. The inaccuracies of the filter are compensated by using a simple automatic tuning system. The design is fabricated in 0.35 um TSMC CMOS process technology. The design was simulated in Cadence using SPICE models provided by MOSIS for the 0.35 um TSMC process in the presence of parasitic capacitance and transistor non-idealities. Cut-off frequency of 500 MHz was achieved along with a 9% variation in the group delay.

Carbon filled columns were built and tested as prototype radon filters. The length of time taken for a pulse of radon to travel from one end of the filter to the other was measured using nitrogen, and then air as the carrier gas. Its dependency on two variables were investigated: temperature of the tube and the flow-rate of carrier gas through the column. Using Height Equivalent to a Theoretical Stage (HETS) the carbon was found to have a adsorption constant (ka) of 9.7 ± 1.7 L/g at STP, and a binding energy value (Eb) of (30±1) kJ/mol. The radon concentration in the carrier gas was measured using a silicon detector housed in a metal chamber with a charged collector plate. Most filter tests measured Rn concentrations by measuring levels of the daughter atom 218Po. A computer simulation was created for aspects of both the working detector chamber and a prototype detector chamber. The charged fractions of 218Po and 214Po were calculated by comparing the Monte Carlo simulation to data obtained from the detectors. The positively charged fraction of 218Po was found to be (80 ± 9)% and 214Po was (110 ± 180)%. / Thesis (Master, Physics, Engineering Physics and Astronomy) -- Queen's University, 2008-08-29 16:38:33.442

The Kalman filter algorithm can be applied as a recursive estimator of the state of a dynamic system described by a linear difference equation. Given discrete measurements linearly related to the state of the system, but corrupted by white Gaussian noise, the Kalman filter estimate of the system is statistically optimal with respect to a quadratic function of the estimate error. The first objective of this paper is to give deep enough insight into the mathematics of the Kalman filter algorithm to be able to choose the correct type of algorithm and to set all the parameters correctly in a basic application. This description also includes several examples of different approaches to derive and to explain the Kalman filter algorithm. In addition to the mathematical description of the Kalman filter algorithm this paper also provides an implementation written in MATLAB. The objective of this part is to correctly replicate the target tracker used in the surveillance radar PS-90. The result of the implementation is evaluated using a simulated target programmed to have an aircraft-like behaviour and done without access to the actual source code of the tracker in the PS-90 radar