FMCW radars are widely used in the process industry for range estimation, usu- ally for estimating the liquid level in a tank. Since the tank system, often is an automatically controlled system, reliable estimates of the surface level are re- quired, e.g. to avoid the tank from pouring over or become empty. The goal of this thesis is to investigate methods which can distinguish fre- quencies closer to each other than the FFT resolution limit. Two properties are of interest, the accuracy and the resolution performance. Three such methods have been evaluated: one that tries to compensate for the leakage and interference of close frequencies, one subspace-based method and one deconvolution method. The deconvolution is performed with the iterative Lucy Richardson algorithm. The methods are evaluated against each other and against a typical FFT based algorithm. The methods sensitivity to amplitude differences is examined together with the robustness against noise and disturbances which appear due to imperfections in the radar unit. The deconvolution algorithm is the one that performs the best. The subspace-based method SURE requires prior knowledge of the number of ingoing frequencies which is difficult to know for real data from an FMCW radar. The leakage compensation method main weakness is the influence of the phase difference between close frequencies. The deconvolution algorithm is evaluated on some real data, and it is proven that it has better resolution performance than the FFT. However, the accuracy of the estimates are dependent on the number of iterations used. With a large num- ber of iterations, the algorithm finds peaks with small amplitude nearby the large peaks and they will thus interact and hence contribute to a worse accuracy even in the undisturbed case. If too few iterations are used in the deconvolution algo- rithm the resolution performance is about the same as the FFT algorithm. With a suitable choice of iterations about 40–50 mm, extra of continuous measurements are achieved. However, the estimation error of the gained resolution can in the worst case be about 40–50 mm.

Asynchronous Transfer Mode (ATM) is fast emerging as the preferred information transfer technique for future Broadband Integrated Services Digital Networks (BISON), offering the advantages of both the simplicity of time division circuit switched techniques and the flexibility of packet switched techniques. ATM networks with their inherent rate flexibility offer new opportunities for the efficient transmission of real time Variable Bit Rate (VBR) services over such networks. Since most services are VBR in nature when efficiently coded, this could in turn lead to a more efficient utilisation of network resources through statistical multiplexing. Video communication is typical of such a service and could benefit significantly if supported with VBR video over ATM networks.

MRI image (Magnetic Resonance Imaging) as a universal body checkup method in modern medicine. It can help doctors to analyze the condition of patients as soon as possible. As the medical images, the MRI images have high quality and a large amount of data, which requires more transmission time and larger storage capacity. To reduce transmission time and storage capacity, the compression and decompression technology is applied. Now most MRI images are colour, but most theses still use gray MRI images to research. Compressed color MRI images is a new research area. In this thesis, some basic theories of the compression technoloy and medical technology were firstly introduced, then basic strcture and kernel algorithm of Huffman coding were explained in detail. Finally, Huffman coding was implemented in MATLAB to compress and decompress the colour MRI images.The result of the experiment shows that the Huffman coding in colour MRI image compression can get high compression ratio and coding efficient.

Kernel methods are able to exploit high-dimensional spaces for representational advantage, while only operating implicitly in such spaces, thus incurring none of the computational cost of doing so. They appear to have the potential to advance the state of the art in control and signal processing applications and are increasingly seeing adoption across these domains. Applications of kernel methods to fault detection and isolation (FDI) have been reported, but few in aerospace research, though they offer a promising way to perform or enhance fault detection. It is mostly in process monitoring, in the chemical processing industry for example, that these techniques have found broader application. This research work explores the use of kernel-based solutions in model-based fault diagnosis for aerospace systems. Specifically, it investigates the application of these techniques to the detection and isolation of IMU/INS sensor faults – a canonical open problem in the aerospace field. Kernel PCA, a kernelised non-linear extension of the well-known principal component analysis (PCA) algorithm, is implemented to tackle IMU fault monitoring. An isolation scheme is extrapolated based on the strong duality known to exist between probably the most widely practiced method of FDI in the aerospace domain – the parity space technique – and linear principal component analysis. The algorithm, termed partial kernel PCA, benefits from the isolation properties of the parity space method as well as the non-linear approximation ability of kernel PCA. Further, a number of unscented non-linear filters for FDI are implemented, equipped with data-driven transition models based on Gaussian processes - a non-parametric Bayesian kernel method. A distributed estimation architecture is proposed, which besides fault diagnosis can contemporaneously perform sensor fusion. It also allows for decoupling faulty sensors from the navigation solution.

The basic form of conventional electro-optic sensors is described. The main drawback of these sensors is their inability to deal with the background radiation which usually accompanies the signal. This 'clutter' limits the sensors performance long before other noise such as 'shot' noise. Pre-detector signal processing using the complex amplitude of the light is introduced as a means to discriminate between the signal and 'clutter'. Further improvements to predetector signal processors can be made by the inclusion of optical fibres allowing radiation to be used with greater efficiency and enabling certain signal processing tasks to be carried out with an ease unequalled by any other method. The theory of optical waveguides and their application in sensors, interferometers, and signal processors is reviewed. Geometrical aspects of the formation of linear and circular interference fringes are described along with temporal and spatial coherence theory and their relationship to Michelson's visibility function. The requirements for efficient coupling of a source into singlemode and multimode fibres are given. We describe interference experiments between beams of light emitted from a few metres of two or more, singlemode or multimode, optical fibres. Fresnel's equation is used to obtain expressions for Fresnel and Fraunhofer diffraction patterns which enable electro-optic (E-0) sensors to be analysed by Fourier optics. Image formation is considered when the aperture plane of an E-0 sensor is illuminated with partially coherent light. This allows sensors to be designed using optical transfer functions which are sensitive to the spatial coherence of the illuminating light. Spatial coherence sensors which use gratings as aperture plane reticles are discussed. By using fibre arrays, spatial coherence processing enables E-0 sensors to discriminate between a spatially coherent source and an incoherent background. The sensors enable the position and wavelength of the source to be determined. Experiments are described which use optical fibre arrays as masks for correlation with spatial distributions of light in image planes of E-0 sensors. Correlations between laser light from different points in a scene is investigated by interfering the light emitted from an array of fibres, placed in the image plane of a sensor, with each other. Temporal signal processing experiments show that the visibility of interference fringes gives information about path differences in a scene or through an optical system. Most E-0 sensors employ wavelength filtering of the detected radiation to improve their discrimination and this is shown to be less selective than temporal coherence filtering which is sensitive to spectral bandwidth. Experiments using fibre interferometers to discriminate between red and blue laser light by their bandwidths are described. In most cases the path difference need only be a few tens of centimetres. We consider spatial and temporal coherence in fibres. We show that high visibility interference fringes can be produced by red and blue laser light transmitted through over 100 metres of singlemode or multimode fibre. The effect of detector size, relative to speckle size, is considered for fringes produced by multimode fibres. The effect of dispersion on the coherence of the light emitted from fibres is considered in terms of correlation and interference between modes. We describe experiments using a spatial light modulator called SIGHT-MOD. The device is used in various systems as a fibre optic switch and as a programmable aperture plane reticle. The contrast of the device is measured using red and green, HeNe, sources. Fourier transform images of patterns on the SIGHT-MOD are obtained and used to demonstrate the geometrical manipulation of images using 2D fibre arrays. Correlation of Fourier transform images of the SIGHT-MOD with 2D fibre arrays is demonstrated.

Transmission and reception of signals in wireless communication systems is affected by additive interference corrupting the signal. Traditionally, the interference is assumed to be AWGN and the system designs are usually based on that assumption. Modern military platforms consists of many electrical components and systems and as such the noise affecting the signals is often a product of interference between the components and systems. This type of noise tend to be very impulsive in nature. The standard AWGN model is not suited for impulsive noise which leaves an opportunity to investigate the performance of a demodulation scheme adapted to the current interference environment in order to increase the performance gain. To properly analyze the performance of an interference-adapted demodulator, knowledge about the characteristic parameters of the chosen noise model is required to perform the necessary calculations.  This project combines the aspect of adaptive demodulation with parameter estimation evaluation. Four different parameter estimation techniques specifically customized for Symmetric alpha-Stable distributed noise were implemented and examined. The four methods were the Empirical Characteristic Function (ECF) method, Fractional Lower-Order Moments (FLOM) method, Extreme-Order Statistics (EOS) method as well as the Quantiles method. The effectiveness and performance of the methods were investigated in two Symmetric alpha-Stable processes of varying level of impulsiveness as well as two Class A processes in order to monitor the performance in noise not specifically distributed according to the intended model, functioning as an arbitrary representation of non-Gaussian interference. The results were evaluated using the measure of Kullback-Leibler Divergence. The demodulator was designed for Symmetric alpha-Stable distributed noise and implemented using an LLR-algorithm. The simulations were performed using an LDPC-coding protocol and the experiment was conducted in both Class A and Symmetric alpha-Stable distributed noise. The modulation schemes were 4-QAM and BPSK. The simulations showed that ECF was the most consistent parameter estimation method overall, regardless of distribution model or number of available samples. FLOM performed well in alpha-Stable noise but struggled in Class A processes. EOS and Quantiles shared the struggles of fewer available samples. The experiments show that an alpha-Stable adapted demodulator coupled with a parameter estimation technique based on the empirical characteristic function (ECF) is a very competitive and viable option in impulsive interference environments regardless of the origin of the noise distribution. The performance gain vis-a-vis demodulation using the standard AWGN option exceeded thresholds of upwards 25 dB for impulsive noise processes.

This report is the result of a Master Thesis project which is a part of the Master programme in Electrical Engineering at Uppsala universitet. The purpose of this Master Thesis project is to develop an embedded platform for the design and implementation of optimal digital filters, in particular, the Kalman and the Wiener filter. In this project these filters are used for noise reduction on noisy signals. The project is a further development of a previous Master Thesis project where a Universal Filter Bank was developed. The Filter Bankis used for designing and implementing various linear digital filters such as lowpass, highpass, bandpass and bandstop. The Filter Bank is a hand held box with two input and two output connections and a human-device interface (HDI) including a Liquid Crystal Display (LCD) and a keypad. It contains anti-aliasing and reconstruction (analog) filters and an ARM 32-bit Microcontroller Unit (MCU) which is programmed in the C programming language. The HDI lets the user specify a desired digital filter. In this project the Kalman and Wiener filtering algorithms were first developed in MATLAB and tested with simulated autoregressive–movingaverage (ARMA) processes (signals) in additive white noise. Aftershown to work, they were implemented on the ARM 32-bit MCU development kit, and finally ported to the Filter Bank. A user interface specially for the specifications of the filters has been created. The Kalman and Wiener filtering algorithms have been tested using the same noisy ARMA processes and assessed in terms of the Normalized Mean Square Error (NMSE). The results have shown that both the Wiener and Kalman filters running on the development kit and the Filter Bank are successful in reducing noise. The Kalman filter is shown to perform better than the Wiener filter, which can be due to the extra information about the signal used in the Kalman filter. The performance of both algorithms are heavily dependent on the pre knowledge about the desired signal.

In the bus industry, it is well-known that bus doors often cause buses to stand at a stop after either a door could not be opened or closed. This causes huge costs for the bus owner. Therefore, it is desirable to be able to predict whether a door's behaviour begins to become too close to causing such an event. Based on this background, this study has been completed at Scania CV AB in Södertälje in order to find out which signals from a door system that can be used to determine if a door is at risk of failing. The main problems that this study is based on are: - Which existing sensors can be used to measure the most common errors before causing a bus that cannot drive? - Is it possible to find a model that can indicate how to adjust speed and attenuation to get a "perfect" adjusted door? These questions were studied by making measurements on a Scania City Wide bus with double inswing door leafs where measurements were made only on the rear door leaf and for its closing behaviour. To make measurements on the door an oscilloscope was used which measured the voltage across a potentiometer with a voltage that varied depending on how open the door was. Another measurement was made on the embedded control system in the bus body. The noise-corrupted measurement data was filtered with a simple type of filter, called moving average filter. In order to find different methods that predict if there were any errors on the door's settings or components, curves were examined over the doors' position during a closing event. One thing that was studied was which position value that was present when the door couldn’t get closed more. Another thing was the curve's slope in different segments of the closing. The slope of the curve was calculated by using the least squares method, which is a mathematical method in linear algebra. Based on the slope in some curve segments of a closing event, a model was created that could guess the door closing time as well as the damping. This model functioned in a convincing way for measurements from the test bus that was used when developing the model. Other models that the study resulted in was one that could warn if the potentiometer values began to approach a limit when the door could no longer be closed, as well as a model that calculated the closing and opening times of the door. Through these models, the idea was that a warning signal could be sent to a system that warned that the door should be adjusted. When some measurements for the guess model were made on another bus, it gave indications that the model did not work as well as the test bus. But one conclusion that could be drawn was that there was an indication that the door was misaligned, even though the model guessed wrong for the present settings. Another conclusion was that it is likely possible to proceed further with some research on this model to find a new model that is somewhat more general for all individual bus doors. / Inom bussbranschen är det välkänt att bussdörrarna alltför ofta orsakar att bussar blir stående vid en hållplats efter att en dörr antingen inte kunde öppnas eller stängas. Detta orsakar stora kostnader för bussägaren. Därför är det önskvärt att kunna förutsäga om en dörrs beteende börjar bli farligt nära att orsaka en sådan händelse. Utifrån denna bakgrund har denna studie utförts vid Scania CV AB i Södertälje för att försöka hitta vilka signaler från ett dörrsystem som kan användas för att kunna avgöra om en dörr riskerar att sluta fungera. De huvudsakliga frågeställningarna som studien utgått ifrån är:  ▪ Vilka befintliga sensorer kan användas för att mäta de vanligaste felen innan de orsakar en buss som inte kan köra vidare?  ▪ Är det möjligt att hitta en modell som kan indikera hur man bör justera hastighet och dämpning för att få en "perfekt" justerad dörr?  Dessa frågor studerades genom att göra mätningar på en Scania City Wide buss med dubbla inåtgående dörrblad där mätningar endast gjordes på det bakre dörrbladet och för dess stängningsbeteende. För att göra mätningarna på dörren användes ett oscilloskop som mätte spänningen över en potentiometer som varierade en spänningen beroende på hur öppen dörren var. Ytterligare en mätning gjordes på det inbyggda styrsystemet i busskarossen. Mätdata filtrerades med enkel typ av filter, glidande medelvärdsfilter, som tog bort brus som smutsat ner mätningarna. För att komma fram till olika metoder att förutsäga om det var något fel på dörrens inställningar eller komponenter så studerades kurvor över dörrarnas position vid en stängning. Det som studerades var för vilka positionsvärden dörren inte kunde stängas mer samt kurvans lutning i olika segment av stängningen. Lutningen på kurvan togs fram genom användning av minsta kvadratmetoden som är en matematisk metod inom linjär algebra. Utifrån lutningen inom vissa kurvsegment av en stängning skapades en modell som kunde gissa en dörrs stängningstid samt dämpningsinställning. Denna modell fungerade på ett övertygande sätt för mätningar på den testbuss som användes vid framtagningen av modellen. Andra modeller som studien mynnade ut i var en som kunde förvarna om potentiometervärdena började närma sig en gräns då dörren inte längre gick att stänga samt en modell som beräknade stängnings- och öppningstiderna för dörren. Genom dessa modeller var tanken att en varningssignal skulle kunna skickas till berörd och varna att dörren bör justeras. När några mätningar för gissningsmodellen gjordes på en annan buss så gav det indikationer på att modellen inte fungerade lika väl som för testbussen. Men en slutsats som man kunde dra var att man fick en indikation att dörren var feljusterad även om modellen gissade fel på hur inställningarna var. Ytterligare en slutsats var att det troligtvis går att utifrån denna modell forska vidare för att finna en ny modell som är något mer generell för alla bussdörrsindivider.

Finite-length impulse response (FIR) filters are one of the most commonly used digital signal processing algorithms used nowadays where a FPGA is the device used to implement it. The continued development of the FPGA device through the insertion of dedicated blocks raised the need to study the advantages offered by different FPGA families. The work presented in this thesis study the special features offered by FPGAs for FIR filters and introduce a cost model of resource utilization. The used method consist of several stages including reading, classification of features and generating coefficients. The results show that FPGAs have common features but also specific differences in features as well as resource utilization. It has been shown that there is misconception when dealing with FPGAs when it comes to FIR filter as compared to ASICs.

One of the fundamental problems of sonar systems is the determination of the bearings of underwater sources/targets. The classical solution to this problem, the 'Conventional Beamformer', uses the outputs from the individual sensors of an acoustic array to form a beam which is swept across the search sector. The resolution of this method is limited by the beam width and narrowing this beam to enhance the resolution may have some practical problems, especially in low frequency sonar, because of the physical size of the array needed. During the past two decades an enormous amount of work has been done to develop new algorithms for resolution enhancements beyond that of the Conventional Beamformer. However, most of these methods have been based on computer simulations and very little has been published on the practical implementation of these algorithms. One of the main reasons for this has been the lack of hardware that can handle the relatively heavy computational load of these algorithms. However, there have been great advances in semiconductor and computer technologies in the last few years which have led to the availability of more powerful computational and storage devices. These devices have opened the door to the possibility of implementing these high-resolution Direction Finding (DF) algorithms in real sonar systems. The work presented in this thesis describes a practical implementation of some of the high-resolution DF algorithms in a simple sonar system that has been designed and built for this purpose.