No description available.

No description available.

<p>Future radars require more flexible and faster radar signal processing chain than commercial radars of today. This means that the demands on the processors in a radar signal system, and the desire to be able to compute larger amount of data in lesser time, is constantly increasing. This thesis focuses on commercial micro-processors of today that can be used for Active Electronically Scanned Array Antenna (AESA) based radar, their physical size, power consumption and performance must to be taken into consideration. The evaluation is based on theoretical comparisons among some of the latest processors provided by PACT, PicoChip, Intrinsity, Clearspeed and IBM. The project also includes a benchmark made on PowerPC G5 from IBM, which shows the calculation time for different Fast Fourier Transforms (FFTs). The benchmark on the PowerPC G5 shows that it is up to 5 times faster than its predecessor PowerPC G4 when it comes to calculate FFTs, but it only consumes twice the power. This is due to the fact that PowerPC G5 has a double word length and almost twice the frequency. Even if this seems as a good result, all the PowerPC´s that are needed to reach the performance for an AESA radar chain would consume too much power. The thesis ends up with a discussion about the traditional architectures and the new multi-core architectures. The future belongs with almost certainty to some kind of multicore processor concept, because of its higher performance per watt. But the traditional single core processor is probably the best choice for more moderate-performance systems of today, if you as developer looking for a traditional way of programing processors.</p>

Future radars require more flexible and faster radar signal processing chain than commercial radars of today. This means that the demands on the processors in a radar signal system, and the desire to be able to compute larger amount of data in lesser time, is constantly increasing. This thesis focuses on commercial micro-processors of today that can be used for Active Electronically Scanned Array Antenna (AESA) based radar, their physical size, power consumption and performance must to be taken into consideration. The evaluation is based on theoretical comparisons among some of the latest processors provided by PACT, PicoChip, Intrinsity, Clearspeed and IBM. The project also includes a benchmark made on PowerPC G5 from IBM, which shows the calculation time for different Fast Fourier Transforms (FFTs). The benchmark on the PowerPC G5 shows that it is up to 5 times faster than its predecessor PowerPC G4 when it comes to calculate FFTs, but it only consumes twice the power. This is due to the fact that PowerPC G5 has a double word length and almost twice the frequency. Even if this seems as a good result, all the PowerPC´s that are needed to reach the performance for an AESA radar chain would consume too much power. The thesis ends up with a discussion about the traditional architectures and the new multi-core architectures. The future belongs with almost certainty to some kind of multicore processor concept, because of its higher performance per watt. But the traditional single core processor is probably the best choice for more moderate-performance systems of today, if you as developer looking for a traditional way of programing processors.

Digital imaging systems for medical applications must bebased upon highly efficient detectors to ensure low patientdose. This is particularly important in screening mammographybecause of the large number healthy women that is examined. Amammography system must also provide high spatial and contrastresolution. Different approaches are compared in this thesis,and it is argued that a system based on photon countingdetectors in a scanned multi-slit geometry provides aperformance superior to established technologies. The system is realized using silicon strip detectorsirradiated at a small angle relative to the wafer surface,thereby offering large absorption depth. A linear pixelarray isscanned across the breast to obtain the complete image.Pulse-processing electronics rejecting all detector andelectronics noise count the number of photons that aredetected, forming the pixel values of the image. Optimization of the detector design is discussed in detail.The detector has been carefully simulated to investigate chargemotion and signal formation after photoninteraction. Based onthese simulations, the impact of the detector characteristicson the image quality has been evaluated. Detectors have been manufactured and evaluated both assingle components and as part of experimental imaging devicesincluding custom readout electronics. Presented in this thesisare the measured detector characteristics including a verifi-cation of charge collection efficiency and confirmation thatthe quantum efficiency is 90% for a typical mammographyspectrum. Modulation transfer functions and noise power spectrawere recorded and the detective quantum efficiency calculated.A prototype mammography system was also assembled and themodulation transfer function recorded. The interpretation ofthe modulation transfer function and detective quantumefficiency is discussed for digital systems in general and fora scanned multi-slit system in particular. <b>Keywords:</b>x-ray, imaging, silicon, detector, digital,mammography, scanning, photon counting.

<p>Digital imaging systems for medical applications must bebased upon highly efficient detectors to ensure low patientdose. This is particularly important in screening mammographybecause of the large number healthy women that is examined. Amammography system must also provide high spatial and contrastresolution. Different approaches are compared in this thesis,and it is argued that a system based on photon countingdetectors in a scanned multi-slit geometry provides aperformance superior to established technologies.</p><p>The system is realized using silicon strip detectorsirradiated at a small angle relative to the wafer surface,thereby offering large absorption depth. A linear pixelarray isscanned across the breast to obtain the complete image.Pulse-processing electronics rejecting all detector andelectronics noise count the number of photons that aredetected, forming the pixel values of the image.</p><p>Optimization of the detector design is discussed in detail.The detector has been carefully simulated to investigate chargemotion and signal formation after photoninteraction. Based onthese simulations, the impact of the detector characteristicson the image quality has been evaluated.</p><p>Detectors have been manufactured and evaluated both assingle components and as part of experimental imaging devicesincluding custom readout electronics. Presented in this thesisare the measured detector characteristics including a verifi-cation of charge collection efficiency and confirmation thatthe quantum efficiency is 90% for a typical mammographyspectrum. Modulation transfer functions and noise power spectrawere recorded and the detective quantum efficiency calculated.A prototype mammography system was also assembled and themodulation transfer function recorded. The interpretation ofthe modulation transfer function and detective quantumefficiency is discussed for digital systems in general and fora scanned multi-slit system in particular.</p><p><b>Keywords:</b>x-ray, imaging, silicon, detector, digital,mammography, scanning, photon counting.</p>

The present thesis is focused on the development and application of narrow scanned high energy photon beam for radiation therapy. The introduction of physically and biologically optimized intensity modulated radiation therapy (IMRT) requires a flexible and accurate dose delivery method to maximize the treatment outcome. Narrow scanned photon beams is a fast option for IMRT since it is not dependent on mechanically moving heavy collimator leafs and largely independent of the complexity of the desired dose distribution. Scanned photon beams can be produced by scanning an electron beam of low emittance, incident on a thin bremsstrahlung target of low atomic number. The large fraction of high energy electrons that are transmitted through the target has to be removed by a strong purging magnet. In the thesis a strong purging magnet, coupled with a magnetic scanning magnet, is presented for an intrinsic electron energy of 50 - 75 MeV and a source to isocenter distance of 75 cm. The available scan area at isocenter can be as large as 43 x 40 cm2 for an incident electron energy of 50 MeV and 28 x 40 cm2 at 75 MeV. By modifying the existing treatment head of the racetrack microtron MM50, it was possible to experimentally produce relevant dose distributions with interesting properties from 50 MV scanned narrow photon beams while deflecting the transmitted electrons onto a simplified electron stopper. The deflection of the transmitted electrons was studied both experimentally and by the Monte Carlo method. With high energy photons, treatment verification is possible through PET-CT imaging of the positron annihilations induced by photonuclear reactions in the patient. Narrow scanned high energy photon beams is the ideal beam quality since the activation efficiency and the effective photon energy will be more uniform than the filtered photon beam from a full range bremsstrahlung target. The induced 11C activity 50 MV by scanned narrow photon beams was measured using PET-CT imaging and compared with Monte Carlo simulations. The combination of fast flexible dose delivery with treatment verification using PET-CT imaging makes narrow high energy scanned photon beams a very interesting treatment modality for biologically optimized adaptive radiation therapy. / At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 3: Submitted. Paper 4: Submitted.

Skogliga inventeringar utförs för att ge information om skogen som kan användas i olika sammanhang. De ger data om bland annat trädhöjd, ålder, grundyta, trädslagsfördelning och volym. De har historiskt utförts som fältinventeringar. På senare tid har den tekniska utvecklingen öppnat upp för möjligheten att utföra inventeringarna via fjärranalys. En metod för fjärranalys som används inom skogsbruket är flygburen laserskanning. Genom att kombinera flygburen laserskanning med Riksskogstaxeringens fältinventeringar har Skogsstyrelsen skapat produkten Skogliga grunddata. Denna produkt visar bland annat skattad volym för bestånd som ritas upp i program för geografiska informationssystem, GIS. Nätraälven Skog använder Skogliga grunddata och de har märkt att volymen inte alltid stämmer när de jämför mot inmätt volym. Felkällor som kan påverka noggrannheten i Skogliga grunddata är t.ex. fel i kartmodellerna, fel mellan systemen som skannar terrängen, mätfel vid fältinventeringarna eller fel vid beräkningar.Syftet med detta arbete är att ta reda på varför det finns en stor differens i volymen när Nätraälven skog jämför mellan Skogliga grunddata och inmätt volym på industri. Dessutom är syftet att urskilja parametrar som påverkar volymen och är generaliserbara på samtliga bestånd inom Nätraälvens upptagningsområde. Arbetet har som mål att efter identifiering av parametrar som påverkar noggrannheten bygga en beräkningsmodell som är applicerbar på bestånden inom Nätraälvens upptagningsområde.Arbetet utfördes som en kvantitativ studie där data från slutavverkningar användes. Jämförelser av volymskillnad och parametrar utfördes som regressionsanalyser. Fältinventeringar utfördes som kvantitativa mätningar i fält. Avverkningstrakter letades upp i ett GIS-program och ritades därefter in för att beräkna volymen via Skogliga grunddata. Denna volym jämfördes sedan mot inmätt volym. Regressionsanalyser genomfördes mellan trakternas tillgängliga parametrar och volymskillnaden mellan Skogliga grunddata och inmätt volym.I medel visade Skogliga grunddata 95% av inmätt volym, som lägst visade den 76% och som högst 128%. I 58% av trakterna låg volymskillnaden i ett intervall på 10%.Totalt har tio parametrar använts vid regressionsanalys av volymskillnaden mellan Skogliga grunddata och inmätt volym. Volymskillnaden har varit väl spridd oavsett parametrarnas storlek. Vad som framkommit i studien är att ingen av parametrarna: trädhöjd, areal, höjd över havet, ålder, stamantal, procentuell andel tall, procentuell andel gran, procentuell andel löv, medelstam eller stammar/hektar enskilt påverkar noggrannheten i Skogliga grunddata. Arbetet har också visat stor spridning i volymskillnad mellan Skogliga grunddata och inmätt volym för olika trakter. Medelskillnaden mellan volymen i Skogliga grunddata och inmätt volym var visserligen fullt acceptabel även om skillnaden trakter emellan kunde vara väldigt stor. / Remote analyses can be a good tool to facilitate silvicultural estimates. They can however contain sources of error, which can affect their accuracy. The purpose of this study was to find out why differences in volume occurred between “forest basic data”, obtained from Swedish forest agency, and measured volume in a certain area in mid-northern Sweden. The study was conducted as a quantitative study, using regression analyzes to compare the volume difference to different parameters, such as tree height, altitude, area, age and species mix. Felling areas was drawn in a GIS-program, that calculated the volume from “Forest basic data”. The volume was then compared to the measured volume obtained from sawmill and pulp companies. The average volume from “Forest basic data” was 95% of measured volume, ranging between 76% at the lowest and 128% at the highest. The study showed that the volume difference was well spread regardless of forest parameter, no specific parameter could explain the differences. A model was created to calculate the volume from “Forest basic data”, and that model improved the volume from “Forest basic data” when compared to measured volume.

This dissertation explores the expansion of the Electrically Steerable Parasitic Array Radiator (ESPAR) technology to arrays using microstrip patch elements. Scanning arrays of two and three closely-coupled rectangular patch elements are presented, which incorporate no phase shifters. These arrays achieve directive radiation patterns and scanning of up to 26° with maintained impedance match. The scanning is effected by tunable reactive loads which are used to control the mutual coupling between the elements, as well as additional loads which compensate to maintain the appropriate resonant frequency. The design incorporates theoretical analysis of the system of coupled antennas with full-wave simulation. A prototype of the threeelement array at 1 GHz is fabricated and measured to exhibit a maximum gain of 7.4 dBi with an efficiency of 79.1%. Further, the microstrip ESPAR is thoroughly compared to uniformlyilluminated arrays of similar size. To satisfy the need for higher directivity antennas with inexpensive electronic scanning, the microstrip ESPAR is then integrated as a subarray. The three-element subcell fabrication is simplified to a single layer with an inverted-Y groove in the ground plane, allowing for DC biasing without the need for the radial biasing stubs or tuning stubs found in the two-layer design. The 1 GHz ESPAR array employs a corporate feed network consisting of a Wilkinson power divider with switchable delay line phase shifts, ring hybrid couplers, and achieves a gain of 12.1 dBi at boresight with ±20° scanning and low side lobes. This array successfully illustrates the cost savings associated with ESPAR subarray scanning and the associated reduction in required number of phase shifters in the RF front end.

The Electronically Steered Phased Array is one of the most versatile antennas used in radars applications. Some of the advantages of electronic steering are faster scan, no moving parts and higher reliability. However, the cost of phased arrays has always been prohibitive - in the order of $1M per square meter. The cost of a phased array is largely impacted by the cost of the high frequency electronics at each element and the cost of packaging. Advances in IC integration will allow incorporating multiple elements such as low noise amplifier, power amplifier, phase shifters and up/down-conversion into one or two ICs. Even though the cost for large quantities of ICs (both Silicon and GaAs) has lowered, the high cost of IC packaging and the array backplane still make the use of phase arrays for radar applications costly. The focus of this research is on techniques that reduce the packaging and the backplane cost of large electronically steered arrays. These techniques are based on simplified signal distributions schemes, reduction of layers in the backplane and use of inexpensive materials. Two architectures designed based on these techniques, as well as a novel BGA active antenna package for dual polarized phased arrays are presented. The first architecture, called the series fed row-column architecture, focuses on the reduction of phase shifters and control signals used in the backplane of the array. The second architecture, called the parallel plate feed architecture, is based on a simplified scheme for distribution of the local oscillator signal. A prototype making use of each one of these architectures is presented. Analysis of advantages and disadvantages of each of these architectures is described. The necessity of cost reduction is a factor that can possibly impact the polarization performance of the antenna. This factor is a motivation to study and develop calibration techniques that reduce the cross-polarization of electronically steered phased arrays. Advances on Interleaving Sparse Arrays, a beam forming technique for polarization improvement/correction in phased arrays, are also presented.