This thesis describes the implementation of an ELINT algorithm for the detection and classification of Low Probability of Intercept (LPI) signals. The algorithm was coded in the C programming language and executed on a Field Programmable Gate Array based reconfigurable computer; the SRC-6 manufactured by SRC Computers, Inc. Specifically, this thesis focuses on the preprocessing stage of an LPI signal processing algorithm. This stage receives a detected signal that has been run through a Quadrature Mirror Filter Bank and outputs the preprocessed signal for classification by a neural network. A major value of this study comes from comparing the performance of the reconfigurable computer to that of supercomputers and embedded systems that are currently used to solve the signal processing needs of the United States Navy. / US Navy (USN) author.

No description available.

An increasing number of LPI radars are integrated into integrated air defense systems (IADS) and modern platforms and weapons, such as anti-ship missiles, and littoral weapon systems. These LPI radars create a requirement for modern armed forces to develop new techniques, strategies, and equipment. The primary objective of this thesis is to investigate methods and means to counter LPI radar threats integrated into a modern platforms and weapons and focus on the related techniques, strategies, and technology. To accomplish this objective both platform centric and network centric approaches will be examined thoroughly.

Technological innovation is inescapable if civilisation is to continue in the face of population growth, rising expectations and resource exhaustion. Unfortunately, major innovations, confidently thought to be safe, occasionally fail catastrophically. The fears so engendered are impeding technical progress generally and that of nuclear power in particular. Attempts to allay disquiet about these disastrous Low Probability Events (LPEs) by exhaustive studies of nuclear power plant designs have, so far, been less than successful. The New Treatment adopts instead an approach that, after examination of the LPE in its historical and societal settings, combines theoretical design analysis with construction site and operational realities in pragmatic engineering, the quality of which can be assured by accountable inspection. The LPE is envisaged as a singularity in a stream of largely mundane, but untoward incidents, described as 'Event-noise'. Predictions of the likelihood of plant LPEs by frequency-theory probability are illusory because the LPE is unique and not part of a stable distribution. Again, noise analysis seems to lead to intractable mathematical expressions. While theoretical LPE prognostications depend on the identification of fault sequences in design that can either be designed-out or reduced to plausibly negligible probabilities, the reality of LPE prevention lies with the plant in operation. As absolute safety is unattainable, the approach aims at ensuring that the perceived residual nuclear risk is societally tolerable. An adaption of elementary Catastrophe theory to model the prospective Event-noise field to be experienced by the plant is proposed whereby potential, credible LPEs could be more readily discerned and avoided. In this milieu of increasing sophistication in technology when management in the traditional administrative mold is proving inadequate, the engineer emerges as the proper central decision-maker. The special intellectual capability needed is acquired during his training and experience, a claim that can draw support from new studies in neuropsychology. The Nuclear Installation Inspectorate is cited as an exemplar of a body practising the kind of engineering inspection needed to apprehend those human fallibilities to which most catastrophic failures of technology are due. Nevertheless, such regulatory systems lack accountability and, as Goedel's theorem suggests, cannot assess their own efficiency. Independent appraisal by Signal Detection Theory is suggested as a remedy.

Thesis (M.S. in Systems Engineering and M.S. in Engineering Science (Electrical Engineering))--Naval Postgraduate School, September 2003. / Thesis advisor(s): Phillip E. Pace, D. Curtis Schleher. Includes bibliographical references (p. 125-126). Also available online.

Today, there is a demand for reliable ways to perform automatic modulation recognition of Low Probability of Intercept (LPI) radar signals, not least in the defense industry. This study explores the possibility of performing automatic modulation recognition on these signals through clustering and more specifically how to learn representations of input signals for this task. A semi-supervised approach using a bootstrapped convolutional neural network classifier for representation learning is proposed. A comparison is made between training the representation learner on raw time-series and on spectral representations of the input signals. It is concluded that, overall, the system trained on spectral representations performs better, though both approaches show promise and should be explored further. The proposed system is tested both on known modulation types and on previously unseen modulation types in the task of novelty detection. The results show that the system can successfully identify known modulation types with adjusted mutual information of 0.86 for signal-to-noise ratios ranging from -10 dB to 10 dB. When introducing previously unseen modulations, up to six modulations can be identified with adjusted mutual information above 0.85. Furthermore, it is shown that the system can learn to separate LPI radar signals from telecom signals which are present in most signal environments. / Idag finns ett behov av pålitlig automatiserad modulationsigenkänning (AMR) av Low Probability of Inercept (LPI)-radarsignaler, inte minst hos försvarsindustrin. Denna studie utforskar möjligheten att utföra AMR av dessa signaler genom klustring och mer specifikt hur man bör lära in representationer av signalerna i detta syfte. En halvövervakad inlärningsmetod som använder en klassificerare baserad på faltningsnätverk föreslås. En jämförelse görs mellan ett system som tränar för representationsinlärning på råa tidsserier och ett system som tränar på spektrala representationer av signalerna. Resultaten visar att systemet tränat på spektrala representationer på det stora hela presterar bättre, men båda metoderna visar lovande resultat och bör utforskas vidare. Systemet testas på signaler från både kända och för systemet tidigare okända modulationer i syfte att pröva förmågan att upptäcka nya typer av modulationer. Systemet identifierar kända modulationer med adjusted mutual information på 0.86 i brusnivåer från -10 dB till 10 dB. När tidigare okända modulationer introduceras till systemet ligger adjusted mutual information över 0.85 för upp till sex modulationer. Studien visar dessutom att systemet kan lära sig skilja LPI-radarsignaler från telekommunikationssignaler som är vanliga i de flesta signalmiljöer.

Context and rationale – This work originates from policy priorities established within Defra to manage exotic animal diseases (EAD); specifically to understand the causes of low probability events, and to establish contingencies to manage outbreak incidents. Outbreaks of exotic animal diseases, e.g. FMD, CSF and HPAI, can cause economic and social impacts of catastrophic proportions. The UK’s government develops and implements policies and controls to prevent EAD and thus minimise these impacts. Control policies to achieve this are designed to address the vulnerabilities within the control systems. However, data are limited for both the introduction of an EAD as well as its resurgence following the disposal of infected carcasses, i.e. the pre-outbreak and post-outbreak phases of an EAD event. These lack of data compromises the development of policy interventions to improve protection. To overcome these data limitations, predictive models are used to predict system vulnerabilities. Cont/d.

Thesis (M.App.Sc.) - Carleton University, 2007. / Includes bibliographical references (p. 83-86). Also available in electronic format on the Internet.

Approved for public release; distribution is unlimited. / In military operations, covertness of operation is of paramount importance. The transmission power of the data link must be kept to the minimum to maintain a low probability of detection (LPD) from the adversary. However, a reduction in the transmitted power implies a reduction in the operating range, though the detection range by the enemy is also reduced. Therefore, to reduce the enemy’s detection range while maintaining operating distance, this thesis explores strategies to discriminate gain against an adversary’s sensor. The strategies involve using processing gain, directional antennas, polarization and the natural environment as a transmission shield. The processing gain strategy analyzed in this thesis uses a diversity technique called Maximal Ratio Combining (MRC) applied to an IEEE 802.16a link. Sinclair D. Smith carried out a study on the possible processing gain derivable from this technique and this thesis will bring his results to practical applications via link analyses. In the event that the link is detected and the enemy decides to carry out jamming, the thesis explores a possible anti-jamming (AJ) strategy by using MRC and a directional antenna. Daniel P. Zastrow carried out a study on the AJ capability of MRC and this thesis brings his results to practical applications via link analyses. / Major, Republic of Singapore Airforce

Cette thèse étudie l’assurabilité des risques catastrophiques sous différents angles. Le premier chapitre intéresse aux risques de très faibles probabilités. Nous montrons comment des instruments financiers hybrides, tels que les obligations catastrophes, peuvent être utilisés pour étendre le domaine d’assurabilité des risques catastrophiques. Notre application au cas du nucléaire en France révèle que, malgré des prix plus élevés pour la réassurance des événements de faibles probabilités, il est possible, et vraisemblablement souhaitable d’organiser un système d’assurance plus important que ce que prévoit la loi française. Le second chapitre s’attaque à la question du prix de la réassurance des risques de faibles probabilités. Nous montrons que les risques catastrophiques, ayant une composante systémique, donnent lieu à une prime de risque dont le montant décroit avec la probabilité de la catastrophe moins rapidement que la disposition à payer d’un assuré typique. Cela explique pourquoi les risques systémiques de faibles probabilités sont difficiles à assurer. Le troisième chapitre étudie le rôle des contrats mutuels et participatifs pour améliorer l’assurabilité des risques catastrophiques. De tels contrats permettent aux assurés d’ajuster au mieux leur demande d’assurance, en prenant en compte la dimension systémique des risques auxquels ils sont exposés. Enfin, le quatrième chapitre étudie l’utilisation d’obligations catastrophes pour assurer le risque de variations du prix des matières premières agricoles consécutives à des aléas climatiques extrêmes. En émettant une obligation catastrophe, l’entreprise qui s’approvisionne en matières premières emprunte un capital qu’elle peut conserver en cas de catastrophe, lorsque ses coûts d’approvisionnement sont élevés. Cette solution présente deux avantages par rapport à une couverture par achats de contrats à terme. D’une part,elle permet de réduire la facture d’assurance par effet de diversification. D’autre part, elle permet d’ajuster la couverture aux contraintes logistiques et stratégiques particulières de l’entreprise. / This thesis addresses several aspects of the insurability of catastrophic risks. In a first chapter, we focus on very low probability events and we show how hybrid financial instruments can be used to extend the domain of insurablerisks. Our application to the case of nuclear accidents using cat-bonds data in France shows that despite the higherprice of reinsurance for low probability events, it is advisable to insure more than is currently provided for by the Frenchlaw. The second chapter takes on the issue of why reinsurance is more costly for low probability events. We show thatbecause catastrophic risks have a systemic component, they give rise to a risk premium in equilibrium which decreasesat a lower pace than the willingness to pay for insurance. We use this finding to explain why systemic low probability catastrophes are hard to insure. The third chapter investigates the role of mutual and participating contracts to improveinsurability. Such contracts are necessary for people to adjust their demand for insurance when individual losses are correlated. Finally, the fourth chapter investigates the use of cat-bonds to hedge the risk of extreme agricultural suppliesprice variations. By issuing a cat-bond, the firm that purchases supplies borrows a capital that can be retained in caseof catastrophe. Such a solution would combine the advantage of risk-pooling, to lower the price of insurance, with lowerbasis risk compared to more traditional hedging strategies such as future purchases.