The application of adaptive antenna techniques to increase the channel capacity in mobile radio communication is discussed. Directional sensitivity is obtained by using an antenna array at the base station, possibly both in receiving and transmitting mode. A scheme for separating several signals at the same frequency is proposed. The method is based on high-resolution direction finding following by optimal combination of the antenna outputs. Comparisons to a method based on reference signals are made. Computer simulations are carried out to test the applicability of the technique to scattering scenarios that typically arise in urban areas. The proposed scheme is found to have great potential in rejecting cochannel interference, albeit at the expense of high computational requirements.
The use of adaptive antenna techniques to increase the channel capacity is discussed. Directional sensitivity is obtained by using an antenna array at the base station, possibly both in receiving and transmitting mode. A scheme for separating several signals at the same frequency is proposed. The method is based on high-resolution direction-finding followed by optimal combination of the antenna outputs. Comparison with a method based on reference signals is made. Computer simulations are carried out to test the applicability of the technique to scattering scenarios that typically arise in urban areas. The proposed scheme is found to have great potential in rejecting cochannel interference, albeit at the expense of high computational requirements.
In urban areas, cellular radio signals are subject to Rayleigh fading, which results in rapid changes in the channel gain. For TDMA-systems without the possibility of frequency hopping, this can cause severe problems for slowly moving cellular phones.
This thesis studies the possibility to use fast power control to counteract the problems introduced by Rayleigh fading in the up link. Simulations with two mobiles, with restrictions of power levels and rate of power change are made.
A two mode power control algorithm basing its control decisions on channel gain predictions is investigated. An important part of the thesis deals with the problem of modeling the gain of a Rayleigh fading channel using regression models, subspace models, and neural network models. Subspace models show the best performance in the noise free case but are very sensitive to noise. Combined with the power control algorithm, the choice of models seems to be of little importance when the noise level is low. The bit-error-rate in simulations is significantly reduced for low noise levels compared to a balanced strategy giving both mobiles equal Carrier-to-Interference-ratio. A comparison with the optimal solution, derived with dynamic programming under the power constraints, indicates that the algorithm is reasonably robust with respect to low noise levels. For higher noise levels, recursive autoregression models of low orders seem to be the most robust ones while subspace models and neural network models are the least robust.
An approach is described how to incorporate knowledge of symbolic/logic character into a conventional framework of noisy observations in dynamical systems. The idea is based on approximating the optimal solution that could theoretically be computed if a complete Bayesian framework were known (and infinite computational power were available). The nature of the approximations, the deviations from optimality and the sensitivity to ad hoc parameters are specifically addressed. This merging of logic and numerics is essential in many problems of adaptation in control and signal processing.
It is shown how to incorporate symbolic or logical knowledge into a conventional framework of noisy observations in dynamical systems. The idea is based on approximating the optimal solution that could, theoretically, be computed if a complete Bayesian framework were known (and infinite computational power were available). The nature of the approximations, the deviations from optimality and the sensitivity to ad hoc parameters are specifically addressed.
We propose a ground target recognition method based on 3-D laser radar data. The method handles general 3-D scattered data. It is based on the fact that man-made objects of complex shape can be decomposed to a set of rectangles. The ground target recognition method consists of four steps; 3-D size and orientation estimation, target segmentation into parts of approximately rectangular shape, identification of segments that represent the target's functional/main parts, and target matching with CAD models. The core in this approach is rectangle estimation. The performance of the rectangle estimation method is evaluated statistically using Monte Carlo simulations. A case study on tank recognition is shown, where 3-D data from four fundamentally different types of laser radar systems are used. Although the approach is tested on rather few examples, we believe that the approach is promising.
Over the years imaging laser radar systems have been developed for military and civilian applications. Among the applications we note collection of 3D data for terrain modeling and object recognition. One part of the object recognition process is to estimate the size and orientation of the object. This paper concerns a vehicle size and orientation estimation process based on scanning laser radar data. Methods for estimation of length and width of vehicles are proposed. The work is based on the assumption that from a top view most vehicles edges are approximately of rectangular shape. Thus, we have a rectangle fitting problem. The first step in the process is sorting of data into lists containing object data and data from the ground closest to the object. Then a rectangle with minimal area is estimated based on object data only. We propose an algorithm for estimation of the minimum rectangle area containing the convex hull of the object data. From the rectangle estimate, estimates of the length and width of the object can be retrieved. The first rectangle estimate is then improved using least squares methods based on both object and ground data. Both linear and nonlinear least squares methods are described. These improved estimates of the length and width are less biased compared to the initial estimates. The methods are applied to both simulated and real laser radar data. The use of the minimum rectangle estimator to retrieve initial parameters for fitting of more complex shapes is discussed.
The development of computer supported control education at Linköping University is presented. A review of the different phases of development of computer support is given. The experiences from the introduction of computer supported exams are discussed.
This book provides signal processing exercises and can with advantage be used together with the text book Signal Processing by Fredrik Gustafsson, Lennart Ljung and Mille Millnert. The chapters of the books are aligned, which means that there are matching exercises to each theory chapter. The first part of the book treats classical digital signal processing based on transforms and filters, while model based digital processing is in focus in the second part. Some exercises are more theoretical and solved by hand, while others are intended for Matlab on a computer. The book material is inspired by real problems, and so are the exercises. This is emphasized by the use of data sets, both simulated and real. Most exercises have complete solutions, and a section with hints provides guidance to some exercises. Selected exercises also result in a Matlab function corresponding to specific signal processing algorithms. These functions are used to solve other exercises. Thereby, the reader gradually build up a signal processing toolbox during the studies of the material. The book homepage contains more information and links to access the matlab functions, data sets and examples used in the book. Main book Signal Processing
Dans ce papier une méthode pour l'estimation de l'entrée (ou déconvolution) est présentée. La méthode est basée principalement sur l'utilisation d'une certaine paramétrization du modèle du signal d'entrée. Pour utiliser cette méthode, nous devons être capable d'exprimer le signal d'entrée en fonction de quelques paramètres inconnues et du temps. L'algorithme est conçu pour estimer, simultanément, les paramètres du signal d'entrée et ceux de la fonction de transfert du système. On se limite à l'étude des systèmes dont la fonction de transfert ne comportant que des pôles (c.à.d modèles ARX). La méthode peut être étendue pour consider aussi les zéros de la fonction de transfert. Il est évident que ceci entraîne une augmentation de la charge numérique. L'algorithme est basé sur des méthodes numériques efficaces comme par exemple la factorisation QR utilisant les transformations de Householder. L'application d'un tel algorithme au codage de la parole est présentée. It est à noter que la qualité du signal synthétisé de la parole, peut être nettement améliorée si un modèle plus détaillé est utilisé pour décrire, le modèle du mouvement des cordes vocal plutôt qu'un train d'impulsion. On montre aussi que la méthode envisagée peut être utilisée pour estimer les paramètres du système vocales et ceux du modèle du mouvement des cordes vocales simultanément.
In this paper computational algorithms for inverse glottal filtering are studied. The objective of inverse glottal filtering is to estimate the driving source. A good model for the glottal pulse is useful for, e.g., speech synthesis, speech recognition and speaker diagnostics. One common approach is to use a parameterized model of the input signal, i.e., the glottal pulses. The algorithm presented enables simultaneous estimation of the parameters of the input signal and the parameters of the system transfer function, the vocal tract model. The presentation here is restricted to transfer functions of all-pole type, i.e., AR-models. The method can be extended to handle zeros in the transfer function. The computational burden would, however, increase significantly. The algorithm uses efficient numerical methods, as, for instance, QR-factorization through Householder transformations.
A way to model systems with abruptly changing dynamics is suggested. The parameters of the system are described as realizations of a finite-state Markov chain. It is further discussed how to perform recursive parameter identification for this type of system. A crucial part in the identification algorithm is to estimate the present state of the Markov chain. The effects of some typical rules to do this estimation are examined. Also a procedure which reduces the need for a priori information is given.
Many different recursive identification methods for time varying systems have been suggested in the literature. An assumption that the variations in the system parameters are slow is common for almost all the methods. When using the methods on systems with faster variations one is forced to compromise between alertness to parameter variations on one hand and noise sensitivity on the other. The topic ofthe thesis is to investigate how such a compromise can be avoided fora certain class of systems.
The systems considered are such that their dynamic changes between some different typical modes. As an example one can think of the different "flight cases" of a supersonic aircraft. The philosophy behind the approach taken in the thesis is that the observations of such a system can be separated into different sets corresponding to the different states of the system. The parameters ofthe different modes can then be estimated from the separated data sets.
Technically, this parallel modelling is achieved by describing the system parameters as the realizations of a Markov-chain. An estimation algorithm for time varying systems based on this parallel modet approach is given in the thesis.
The behaviour of the algorithm is analysed and problems connected to it are illustrated through simulations. The analysis and the simulations show that a major problem is the initialization of the algorithm without sufficient a priori information. Based on the analysis a procedure is given that makes it possible to use the algorithm with a minimum of a priori information.
It is further shown how this recursive identification algorithm can be utilized for adaptive control. As an illustration of this, the method is used for adaptive control of a mode! of a cold rolling mill.
Many different recursive identification methods for time-varying systems have been suggested in the literature. An assumption that the variations in the system parameters are slow is common for all the methods. When using the methods on systems with faster variations, one is forced to compromise between alertness to parameter variations on one hand and noise sensitivity on the other. The topic of this paper is to investigate if this compromise can be avoided for a special class of systems. The systems considered are such that their dynamic changes between some different typical modes. The philosophy behind the approach taken in the paper is to separate the observations into different sets corresponding to the different modes. The parameters of the different modes can then be estimated using the separated data sets. Technically, this parallel modelling is achieved by describing the system parameters as the realizations of a Markov chain. A parameter-identification algorithm for time-varying ARX models is then given in the paper. The behaviour of the algorithm is then investigated using simulations and some analysis. The analysis and the simulations show that a major problem is the initialization of the algorithm. Based on the analysis, modifications are made to the algorithm that improve the convergence properties.
A coherent laser radar system based on semiconductor laser technology has been designed and built. The compact design and the absence of adjustments makes the system mechanically robust and easy to use. The present system has an output power of 50 mW and a line width of 280 kHz (HWHM). The laser radar system has been used in vibrometry measurements. For vibrometry of moving objects, adaptive signal processing is required in order to obtain the vibration signature. Especially for unresolved objects, interference between different vibrating parts will complicate the analysis. Modelbased estimation techniques are used to obtain the parameters which determine the dynamics of the reflecting object.
Recently a novel approach to nonlinear function approximation using hinging hyperplanes, was reported by L. Breiman [3]. In this contribution we have combined smooth hinging hyperplanes and the efficient initialization procedure existing for hinging hyperplanes in [3], with a Gauss-Newton procedure, see [4], to perform the final adjustment of the smooth hinging hyperplanes. This combination uses the property of the hinge functions that makes them effective, namely that there is a simple and computationally efficient method for locating hinges. The result of the hinge finding procedure is then used as an initial value to the Gauss-Newton procedure applied on the smoothed hinges. The smooth hinging hyperplanes and neural networks are related, but the significant problem of choosing initial parameters of neural networks, in this case, is circumvented. Further, the influence of the choice of initial value of the "smoothness parameter" on the final approximating function estimate, is investigated. A recommendation on how to choose an initial value is given.
Three examples of techniques that can be used for state order estimation of hidden Markov models are given. The methods are also exemplified using real laser range data, and the computational burden of the three methods is discussed. Two techniques, maximum description length and maximum a posteriori estimate, are shown to be very similar under certain circumstances. The third technique, predictive least squares, is novel in this context.
In this contribution three examples of techniques that can be used for state order estimation of hidden Markov models are given The methods are also exem plied using real laser range data and the computa tional burden of the three methods is discussed Two techniques Maximum Description Length and Maximum a Posteriori Estimate are shown to be very sim ilar under certain circumstances The third technique Predictive Least Squares is novel in this context
Segmentation is a first step towards successful tracking and object recognition in 2-D pictures. Mostly the pictures are segmented with respect to quantities as range, intensity etc. Here a method is presented for segmentation of 2-D laser range pictures with respect to both range and variance simultaneously. This is very useful since man-made objects differ from the background in the terrain by their smoothness. The approach is based on modeling horizontal scans of the terrain as piecewise constant functions. Since the environment has a complicated and irregular structure we use multiple models for modeling different segments in the laser range image. The switching between different models, i.e., ranges belonging to different segments in a horizontal scan, are modeled by a hidden Markov model. The method is of relatively low computational complexity and the maximal complexity can be controlled by the user. Real data is used for illustration of the method.