This thesis proposes a distributed observer design for a class of nonlinear systems that arise in the application of model reduction techniques. Distributed observer design techniques have been proposed in the literature to address estimation problems over sensor networks. In large complex sensor networks, an efficient technique that minimizes the extent of the required communication is highly desirable. This is especially true when sensors have problems caused by physical limitations that result in incorrect information at the local level affecting the estimation of states globally. To address this problem, scalable algorithms for a suitable distributed observer have been developed. Most algorithms are focussed on large linear dynamical systems and they are not directly generalizable to nonlinear systems. In this thesis, scalable algorithms for distributed observers are proposed for a class of large scale observable nonlinear system. Distributed systems models multi-agent systems in which each agents attempts to accomplish local tasks. In order to achieve global objectives, there should be agreement regarding some commonly known variables that depend on the state of all agents. These variables are called consensus states. Once identified, such consensus states can be exploited in the development of distributed consensus algorithms. Consensus algorithms are used to develop information exchange protocols between agents such that global objectives are met through local action. In this thesis, a higher order observer is applied in the distributed sensor network system to design a distributed observer for a class nonlinear systems. Fusion of measurement and covariance information is applied to the higher order filter as the first method. The consensus filter is embedded in the local nonlinear observer for fusion of data. The second method is based on the communication of state estimates between neighbouring sensors rather than fusion of data measurement and covariance. The second method is found to reduce disagreement of the states estimation between each sensor. The performance of these new algorithms is demonstrated by simulation, and the second method is effectively applied over the first method. / Thesis (Master, Chemical Engineering) -- Queen's University, 2012-12-12 11:22:49.113

This thesis involves restoration, reconstruction, and enhancement of a digitized library of hand-written documents. Imaging systems that perform this digitization often degrade the quality of the original documents. Many techniques exist for reconstructing, restoring, and enhancing digital images; however, many require <i> a priori </i> knowledge of the imaging system. In this study, only partial <i> a priori </i> knowledge is available, and therefore unknown parameters must be estimated before restoration, reconstruction, or enhancement is possible. The imaging system used to digitize the documents library has degraded the images in several ways. First, it has introduced a ringing that is apparent around each stroke. Second, the system has eliminated strokes of narrow widths. To restore these images, the imaging system is modeled by estimating the point spread function from sample impulse responses, and the image noise is estimated in an attempt to apply standard linear restoration techniques. The applicability of these techniques is investigated in the first part of this thesis. Then nonlinear filters, structural techniques, and enhancement techniques are applied to obtain substantial improvements in image quality. / Master of Science

Algebraic attacks constitute an effective class of cryptanalytic attacks which have come up recently. In algebraic attacks, the relations between the input, output and the key are expressed as a system of equations and then solved for the key. The main idea is in obtaining a system of equations which is solvable using reasonable amount of resources. The new approaches proposed in this work and experimental studies on the existing algebraic attacks on stream ciphers will be presented. In the first attack on filter generator, the input-output relations are expressed in conjunctive normal form. The system of equations is then solved using modified Zakrevskij technique. This was one of the earliest algebraic attacks on the nonlinear filter generator. In the second attack, we relaxed the constraint on algebraic attack that the entire system description be known and the output sequence extension problem where the filter function is unknown is considered. We modeled the problem as a multivariate interpolation problem and solved it. An advantage of this approach is that it can be adapted to work for noisy output sequences where as the existing algebraic attacks expect the output sequence to be error free. Adding memory to filter/combiner function increases the degree of system of equations and finding low degree equations in this case is computeintensive. The method for computing low degree relations for combiners with memory was applied to the combiner in E0 stream cipher. We found that the relation given in literature [Armknecht and Krause] was incorrect. We obtained the correct equation and verified its correctness. A time-data size trade off attack for clock controlled filter generator was developed. The time complexity and the data requirements are in between the two approaches used in literature. A recent development of algebraic attacks - the Cube attack was studied. Cube attack on variants of Trivium were proposed by Dinur and Shamir where linear equations in key bits were obtained by combining equations for output bit for same key and a set of Initialization Vectors (IVs). We investigated the effectiveness of the cube attack on Trivium. We showed that the linear equations obtained were not general and hence the attack succeeds only for some specific values of IVs. A reason for the equations not being general is given and a modification to the linear equation finding step suggested.

<div> <div> <div> <p>Speech quality is a demand in voice commanded systems and in telephony. The voice communication system in real time often suffers from audible echoes. In order to cancel echoes, an acoustic echo cancellation system is designed and applied to increase speech quality both subjectively and objectively. </p> <p>In this research we develop various nonlinear adaptive filters wielding the new channel sparsity-aware recursive least squares (RLS) algorithms using a sequential update. The developed nonlinear adaptive filters using the sparse sequential RLS (S-SEQ-RLS) algorithm apply a discard function to disregard the coefficients which are not significant or close to zero in the weight vector for each channel in order to reduce the computational load and improve the algorithm convergence rate. The channel sparsity-aware algorithm is first derived for nonlinear system modeling or system identification, and then modified for application of echo cancellation. Simulation results demonstrate that by selecting a proper threshold value in the discard function, the proposed nonlinear adaptive filters using the RLS (S-SEQ-RLS) algorithm can achieve the similar performance as the nonlinear filters using the sequential RLS (SEQ-RLS) algorithm in which the channel weight vectors are sequentially updated. Furthermore, the proposed channel sparsity-aware RLS algorithms require a lower computational load in comparison with the non-sequential and non-sparsity algorithms. The computational load for the sparse algorithms can further be reduced by using data-selective strategies. </p> </div> </div> </div>

Ačkoliv můžeme v literatuře nalézt řadu příkladů prezentujících evoluční návrh jakožto zajímavou a slibnou alternativu k tradičním návrhovým technikám používaným v oblasti číslicových obvodů, praktické nasazení je často problematické zejména v důsledku tzv. problému škálovatelnosti, který se projevuje např. tak, že evoluční algoritmus je schopen poskytovat uspokojivé výsledky pouze pro malé instance řešeného problému. Vážný problém představuje tzv. problém škálovatelnosti evaluace fitness funkce, který je markantní zejména v oblasti syntézy kombinačních obvodů, kde doba potřebná pro ohodnocení kandidátního řešení typicky roste exponenciálně se zvyšujícím se počtem primárních vstupů. Tato disertační práce se zabývá návrhem několika metod umožňujících redukovat problem škálovatelnosti evaluace v oblasti evolučního návrhu a optimalizace číslicových systémů. Cílem je pomocí několika případových studií ukázat, že s využitím vhodných akceleračních technik jsou evoluční techniky schopny automaticky navrhovat inovativní/kompetitivní řešení praktických problémů. Aby bylo možné redukovat problém škálovatelnosti v oblasti evolučního návrhu číslicových filtrů, byl navržen doménově specifický akcelerátor na bázi FPGA. Tato problematika reprezentuje případ, kdy je nutné ohodnotit velké množství trénovacích dat a současně provést mnoho generací. Pomocí navrženého akcelerátoru se podařilo objevit efektivní implementace různých nelineárních obrazových filtrů. S využitím evolučně navržených filtrů byl vytvořen robustní nelineární filtr implusního šumu, který je chráněn užitným vzorem. Navržený filtr vykazuje v porovnání s konvenčními řešeními vysokou kvalitu filtrace a nízkou implementační cenu. Spojením evolučního návrhu a technik známých z oblasti formální verifikace se podařilo vytvořit systém umožňující výrazně redukovat problém škálovatelnosti evoluční syntézy kombinačních obvodů na úrovni hradel. Navržená metoda dovoluje produkovat komplexní a přesto kvalitní řešení, která jsou schopna konkurovat komerčním nástrojům pro logickou syntézu. Navržený algoritmus byl experimentálně ověřen na sadě několika benchmarkových obvodů včetně tzv. obtížně syntetizovatelných obvodů, kde dosahoval v průměru o 25% lepších výsledků než dostupné akademické i komerční nástroje. Poslední doménou, kterou se práce zabývá, je akcelerace evolučního návrhu lineárních systémů. Na příkladu evolučního návrhu násobiček s vícenásobnými konstantními koeficienty bylo ukázáno, že čas potřebný k evaluaci kandidátního řešení lze výrazně redukovat (defacto na ohodocení jediného testovacího vektoru), je-li brán v potaz charakter řešeného problému (v tomto případě linearita).