東華大學圖書館 |

Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions = = Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions =/
其他題名:	Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
其他題名:	Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
作者:	Santra, Avik.
面頁冊數:	1 online resource (100 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
Contained By:	Dissertations Abstracts International84-04B.
標題:	Mean square errors. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29311716click for full text (PQDT)
ISBN:	9798351488905

Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions = = Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
Santra, Avik.

Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions =Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen. - 1 online resource (100 pages)

Source: Dissertations Abstracts International, Volume: 84-04, Section: B.

Thesis (Ph.D.)--Friedrich-Alexander-Universitaet Erlangen-Nuernberg (Germany), 2022.

Includes bibliographical references

The topic of detection, tracking and classification of weak targets in interferencedominated environment using radars is studied in this thesis. The problem is approached from the perspective of both resource-friendly radar systems and resourcelimited radar systems. In the case of resource-friendly radar systems, cognitive architectures that can progressively sense the environment and adjust its operating waveform-receiver filters are analyzed. The study of joint optimal transmit waveform and receive filters such that they operate optimally in the presence of weak targets and interference has been of huge interest in both academia and industry. Recent advances in adaptive waveform synthesis have focused on joint design and implementation of knowledge-aided receiver signal processing techniques and adaptive transmit signals. This closed loop radar framework that mimics mammals' neurological capability to tune system parameters in response to cognition of the environment is commonly referred to as "cognitive radar" or "fully adaptive radar".In this thesis, the output signal to interference noise ratio maximizing jointly optimal transmit waveform and receive filter for a single-input, single-output radar design in the presence of extended target and colored interference is presented. The ambiguity function, processing gain and Cramer-Rao bound for such waveformfilters are derived. Apart from the optimal waveform dictated by the joint optimization strategy, it is desired that the radar transmit waveforms possess constant time envelope to drive the power amplifiers at saturation. This constraint requires reconstruction of constant envelope signals, which is addressed using proposed relaxed iterative error reduction algorithm. In general, iterative algorithms are sensitive to the initial seed, which is solved by deriving a closed-form solution making stationary phase assumption. In the case of multiple-input multiple-output (MIMO) radars, the interference between signals can significantly limit the radar's ability for observation of weak targets in presence of stronger targets and background clutter. For the multi-channel radars, in this thesis orthogonally coded Linear Frequency Modulated (LFM) waveforms is proposed, wherein consecutive complex LFM signals in a frame are coded by orthogonal codes, namely Golay complementary, Zadoff Chu, direct spread spectrum, space-time block coded, discrete Fourier transform and Costasbased sequences. The orthogonal codes to modulate the LFM across symbol form fixed library waveforms leading to partial adaptation instead of arbitrary waveform dictated by "fully adaptive radar". The ambiguity function for such orthogonallycoded MIMO radar is derived, and the waveforms are analyzed in terms of their ambiguity function and imaging performance.With the advancement in silicon and packaging technology, radars have evolved from high-end aerospace technology into relatively low-cost Human-Machine Interface (HMI) sensors. However, the sensor in such industrial and consumer setting should have a small form-factor and low-cost, thus they cannot sustain cognitive architectures to detect and classify weak human targets. To improve detection and classification performance for HMI applications, novel processing and learning algorithms are proposed. In practice, there are several challenges to learningbased solutions using low-cost radars particularly with respect to open set classification. In open set classification, the system needs to handle variations of the input data, alien operating environment and unknown classes. Conventional deep learning approaches use a simple softmax layer and evaluate the accuracy on known classes, thus on closed set classification. The softmax layer provides separability of classes but does not provide discriminative class boundaries. Hence, many unknown classes are erroneously predicted as one of the known classes with a high confidence, resulting in poor performance in real world environments. Other challenges arise due to the inconspicuous interclass difference between features from one class and other closely-related class and large intra-class variations in the radar data from the same classes. To address these challenges, novel representation learning algorithms along with novel loss functions are proposed in this thesis. Unlike conventional deep learning approaches using softmax that learns to classify, deep representation learning learns the process to classify by projecting the input feature images to an embedded space where similar classes are grouped together while dissimilar classes are far apart. Thus, deep representation learning approaches simultaneously learn separable inter-class difference and compact discriminative intraclass, essential for open set classification. Specifically, the proposed representation learning algorithms are evaluated in the context of gesture sensing, material classification, air-writing and kick sensing HMI applications.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798351488905Subjects--Topical Terms:

3562318
Mean square errors.
Index Terms--Genre/Form:

542853
Electronic books.

Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions = = Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
LDR:12737nmm a2200385K 4500 001 2356129
005 20230612071825.5
006 m o d
007 cr mn ---uuuuu
008 241011s2022 xx obm 000 0 eng d
020 $a 9798351488905
035 $a (MiAaPQ)AAI29311716
035 $a (MiAaPQ)FAUErlangen19371
035 $a AAI29311716
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Santra, Avik. $3 3696599
245 1 0 $a Cognitive Architectures, Processing and Learning Algorithms for Intelligent Radar Solutions = $b Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
246 3 1 $a Kognitive Architekturen, Verarbeitungsund Lernalgorithmen fur Intelligente Radar-Losungen.
264 0 $c 2022
300 $a 1 online resource (100 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
500 $a Advisor: Huemer, Mario ; Weigel, Robert.
502 $a Thesis (Ph.D.)--Friedrich-Alexander-Universitaet Erlangen-Nuernberg (Germany), 2022.
504 $a Includes bibliographical references
520 $a The topic of detection, tracking and classification of weak targets in interferencedominated environment using radars is studied in this thesis. The problem is approached from the perspective of both resource-friendly radar systems and resourcelimited radar systems. In the case of resource-friendly radar systems, cognitive architectures that can progressively sense the environment and adjust its operating waveform-receiver filters are analyzed. The study of joint optimal transmit waveform and receive filters such that they operate optimally in the presence of weak targets and interference has been of huge interest in both academia and industry. Recent advances in adaptive waveform synthesis have focused on joint design and implementation of knowledge-aided receiver signal processing techniques and adaptive transmit signals. This closed loop radar framework that mimics mammals' neurological capability to tune system parameters in response to cognition of the environment is commonly referred to as "cognitive radar" or "fully adaptive radar".In this thesis, the output signal to interference noise ratio maximizing jointly optimal transmit waveform and receive filter for a single-input, single-output radar design in the presence of extended target and colored interference is presented. The ambiguity function, processing gain and Cramer-Rao bound for such waveformfilters are derived. Apart from the optimal waveform dictated by the joint optimization strategy, it is desired that the radar transmit waveforms possess constant time envelope to drive the power amplifiers at saturation. This constraint requires reconstruction of constant envelope signals, which is addressed using proposed relaxed iterative error reduction algorithm. In general, iterative algorithms are sensitive to the initial seed, which is solved by deriving a closed-form solution making stationary phase assumption. In the case of multiple-input multiple-output (MIMO) radars, the interference between signals can significantly limit the radar's ability for observation of weak targets in presence of stronger targets and background clutter. For the multi-channel radars, in this thesis orthogonally coded Linear Frequency Modulated (LFM) waveforms is proposed, wherein consecutive complex LFM signals in a frame are coded by orthogonal codes, namely Golay complementary, Zadoff Chu, direct spread spectrum, space-time block coded, discrete Fourier transform and Costasbased sequences. The orthogonal codes to modulate the LFM across symbol form fixed library waveforms leading to partial adaptation instead of arbitrary waveform dictated by "fully adaptive radar". The ambiguity function for such orthogonallycoded MIMO radar is derived, and the waveforms are analyzed in terms of their ambiguity function and imaging performance.With the advancement in silicon and packaging technology, radars have evolved from high-end aerospace technology into relatively low-cost Human-Machine Interface (HMI) sensors. However, the sensor in such industrial and consumer setting should have a small form-factor and low-cost, thus they cannot sustain cognitive architectures to detect and classify weak human targets. To improve detection and classification performance for HMI applications, novel processing and learning algorithms are proposed. In practice, there are several challenges to learningbased solutions using low-cost radars particularly with respect to open set classification. In open set classification, the system needs to handle variations of the input data, alien operating environment and unknown classes. Conventional deep learning approaches use a simple softmax layer and evaluate the accuracy on known classes, thus on closed set classification. The softmax layer provides separability of classes but does not provide discriminative class boundaries. Hence, many unknown classes are erroneously predicted as one of the known classes with a high confidence, resulting in poor performance in real world environments. Other challenges arise due to the inconspicuous interclass difference between features from one class and other closely-related class and large intra-class variations in the radar data from the same classes. To address these challenges, novel representation learning algorithms along with novel loss functions are proposed in this thesis. Unlike conventional deep learning approaches using softmax that learns to classify, deep representation learning learns the process to classify by projecting the input feature images to an embedded space where similar classes are grouped together while dissimilar classes are far apart. Thus, deep representation learning approaches simultaneously learn separable inter-class difference and compact discriminative intraclass, essential for open set classification. Specifically, the proposed representation learning algorithms are evaluated in the context of gesture sensing, material classification, air-writing and kick sensing HMI applications.
520 $a In dieser Arbeit wird das Thema der Erkennung, Verfolgung und Klassifizierung von schwachen Zielen in einer von Interferenzen dominierten Umgebung mit Hilfe von Radarsystemen untersucht. Das Problem wird sowohl aus der Perspektive ressourcenschonender Radarsysteme als auch aus der Perspektive ressourcenbeschrankter Radarsysteme betrachtet. Im Fall von ressourcenschonenden Radarsystemen werden kognitive Architekturen analysiert, die schrittweise die Umgebung erfassen und ihre Sendesignalform und Empfangsfilter anpassen konnen. Die Untersuchung gemeinsamer optimaler Sendesignalformen und Empfangsfilter, die bei schwachen Zielen und Storungen optimal funktionieren, ist sowohl in der Wissenschaft als auch in der Industrie von grosem Interesse. Jungste Fortschritte in der adaptiven Signalformsynthese haben sich auf den gemeinsamen Entwurf und die Implementierung von wissensgestutzten Empfangssignalverarbeitungstechniken und adaptiven Sendesignalen konzentriert. Ein derartiges Radarsystem mit geschlossenem Regelkreis, das die neurologische Fahigkeit von Saugetieren nachahmt, die Systemparameter als Reaktion auf die Wahrnehmung der Umgebung einzustellen, wird gemeinhin als "kognitives Radar" oder "vollstandig adaptives Radar" bezeichnet.In dieser Arbeit werden eine optimale Signalform und der optimale Empfangsfilter, die das Signal-zu-Interferenz-und-Rausch-Verhaltnis maximieren, fur ein Radar mit einem Empfanger und einem Sender in Prasenz eines ausgedehnten Ziels und farbiger Interferenz vorgestellt. Die Mehrdeutigkeitsfunktion, die Verarbeitungsverstarkung und die Cramer-Rao-Schranke fur solche Signalform-Filter werden hergeleitet. Abgesehen von der optimalen Signalform, die durch die gemeinsame Optimierungsstrategie vorgegeben wird, ist es erwunscht, dass die Radar-Sendesignale eine konstante zeitliche Hullkurve besitzen, um die Leistungsverstarker in die Sattigung zu treiben. Diese Bedingung erfordert die Rekonstruktion von Signalen mit konstanter Hullkurve, was mit dem vorgeschlagenen iterativen Splitting-Verfahren zur Fehlerreduktion angegangen wird. Im Allgemeinen sind iterative Algorithmen empfindlich gegenuber ihrem Initialwert, welcher durch die Herleitung einer geschlossenen Losung unter der Annahme einer stationaren Phase bestimmt wird. Bei Mehrkanal-Radargeraten kann die Interferenz zwischen den Signalen die Fahigkeit des Radars zur Erkennung schwacher Ziele in Gegenwart von starkeren Zielen und Hintergrundstorungen erheblich einschranken. Fur Mehrkanal-Radargerate werden in dieser Arbeit orthogonal kodierte und linear-frequenzmodulierte Signalformen vorgestellt, bei denen Signale in einem Datenblock als orthogonal kodierte Sequenzen dargestellt werden. In der Arbeit werden hierfur Golay-Sequenzen, Zadoff-Chu-Sequenzen, Chipping-Sequenzen, Raum-Zeit blockkodierte Sequenzen, diskret Fourier-transformierte Sequenzen und Costas-Sequenzen verwendet. Die orthogonalen Codes, mit denen die lineare Frequenzmodulation symbolubergreifend moduliert wird, bilden fest vordefinierte Signalformen, die zu einer teilweisen Anpassung, anstatt einer willkurlichen Signalform, wie sie beim vollstandig adaptiven Radar erforderlich ist, fuhren. Die Mehrdeutigkeitsfunktion fur ein solches orthogonal codiertes MIMO-Radar wird hergeleitet, und die Signalformen werden hinsichtlich ihrer Mehrdeutigkeitsfunktion und bildgebenden Leistung analysiert.Mit den Fortschritten in der Silizium- und Gehausetechnologie haben sich Radare von High-End-Sensoren in der Luft- und Raumfahrt zu kostengunstigen Sensoren fur die Interaktion zwischen Mensch und Maschine entwickelt. Die Sensoren in solchen Industrie- und Verbraucheranwendungen sollten jedoch einen kleinen Formfaktor haben und kostengunstig sein, sodass sie keine kognitiven Architekturen zur Erkennung und Klassifizierung schwacher menschlicher Ziele unterstutzen konnen. Um die Erkennungs- und Klassifizierungsgenauigkeit fur Mensch-MaschineInteraktions-Anwendungen zu verbessern, werden neue Verarbeitungs- und Lernalgorithmen vorgeschlagen. In der Praxis gibt es mehrere Herausforderungen fur lernbasierte Losungen mit kostengunstigen Radargeraten, insbesondere im Hinblick auf die Klassifizierung offener Mengen. Bei der Klassifizierung offener Mengen muss das System mit Variationen der Eingabedaten, der fremden Betriebsumgebung und unbekannten Klassen umgehen konnen. Herkommliche mehrschichtige Lernansatze verwenden eine einfache Softmax-Schicht zur Klassifizierung und bewerten ihre Genauigkeit nur innerhalb bekannter Klassen, das heist bei der Klassifizierung geschlossener Mengen. Die Softmax-Schicht sorgt fur die Trennbarkeit von Klassen, bietet aber keine diskriminierenden Klassengrenzen. Daher werden viele unbekannte Klassen falschlicherweise als eine der bekannten Klassen mit hoher Sicherheit interpretiert, was zu einer schlechten Leistung in realen Umgebungen fuhrt. Andere Herausforderungen ergeben sich aus den geringen Unterschieden zwischen den Merkmalen einer Klasse und denen einer anderen, eng verwandten Klasse sowie aus einer grosen Varianz der Radardaten derselben Klasse. Um diese Herausforderungen zu bewaltigen, werden in dieser Arbeit neuartige Algorithmen fur das Lernen der Reprasentation der Daten zusammen mit neuartigen Verlustfunktionen vorgeschlagen. Im Gegensatz zu konventionellen mehrschichtigen Lernansatzen, die eine Softmax-Schicht zur Klassifizierung verwenden, wird beim mehrschichtigen Lernen der Reprasentation der Daten das Klassifizieren dadurch gelernt, dass Eingangsdaten so in einen eingebetteten Raum projiziert werden, dass ahnliche Klassen zusammen gruppiert und unahnliche Klassen weit separiert werden. So lernen mehrschichtige Ansatze zum Lernen der Reprasentation von Daten gleichzeitig trennbare Unterschiede zwischen den Klassen und kompakte diskriminierende Eigenschaften innerhalb der Klassen, welche fur die Klassifizierung offener Mengen unerlasslich sind. Die vorgeschlagenen Algorithmen zum Lernen der Reprasentation von Daten werden insbesondere im Kontext von Gestenerkennung, Materialklassifizierung, der Erkennung von in Luft geschriebenen Buchstaben und Zahlen und von Kick-Sensing-Anwendungen bewertet.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Mean square errors. $3 3562318
650 4 $a Design. $3 518875
650 4 $a Algorithms. $3 536374
650 4 $a Signal to noise ratio. $3 3681178
650 4 $a Target recognition. $3 3696600
650 4 $a Signal processing. $3 533904
650 4 $a Radar systems. $3 3694386
650 4 $a Neural networks. $3 677449
650 4 $a Parameter estimation. $3 567557
650 4 $a Classification. $3 595585
650 4 $a Computer science. $3 523869
650 4 $a Electrical engineering. $3 649834
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0389
690 $a 0800
690 $a 0984
690 $a 0544
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a Friedrich-Alexander-Universitaet Erlangen-Nuernberg (Germany). $3 3564652
773 0 $t Dissertations Abstracts International $g 84-04B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29311716 $z click for full text (PQDT)