Over the next decade, 5G networks will become more and more prevalent in everyday life. This will provide solutions to current limitations by allowing access to bands previously unavailable to civilian communication networks. However, this also provides new challenges primarily for the military operations. Radar bands have traditionally operated primarily in the sub-6 GHz region. In the past, these bands were off limits to civilian communications. However, that changed when they were opened up in the 2010's. With these bands now being forced to co-exist with commercial users, military operators need systems to identify the signals within a spectrum environment. In this thesis, we extend current research in the area of signal identification by using previous work in the area to construct a deep learning-based classifier that is able to classify a signal as either as a communication waveform (Single-Carrier (SC), Single-Carrier Frequency Division Multiple Access (SC-FDMA), Orthogonal Frequency Division Multiplexing (OFDM), Amplitude Modulation (AM), Frequency Modulation (FM)) or a radar waveform (Linear Frequency Modulation (LFM) or Phase-coded). However, the downside to this method is that the classifier is based on the assumption that all possible signals within the spectrum environment are within the training dataset. To account for this, we have proposed a novel classifier design for detection of unknown signals outside of the training dataset. This two-classifier system forms an open-set recognition (OSR) system that is used to provide more situational awareness for operators. / M.S. / Over the next decade, next-generation communications will become prevalent in everyday life providing solutions to limitation previously experienced by older networks. However, this also brings about new challenges. Bands in the electromagnetic spectrum that were reserved for military use are now being opened up to commercial users. This means that military and civilian networks now have a challenge of co-existence that must be addressed. One way to address this is being aware of what signals are operating in the bands such as either communication signals, radar signals, or both. In this thesis, we will developed a system that can do that task of identifying a signal as one of five communication waveforms or two radar waveforms by using machine learning techniques. We also develop a new technique for identifying unknown signals that might be operating within these bands to further help military and civilian operators monitor the spectrum.

Τα τελευταία χρόνια παρατηρήθηκε μια ραγδαία αύξηση στα ασύρματα συστήματα επικοινωνίας και τις σχετικές εφαρμογές. Μετά από αυτές τις εξελίξεις, το κλασικό σύστημα αδειοδότησης και κατόπιν αποκλειστικής χρήσης του ηλεκτρομαγνητικού φάσματος οδηγείται στα όριά του, καθώς πλέον πολύ λίγες περιοχές του φάσματος είναι ελεύθερες. Ωστόσο, αρκετές έρευνες που πραγματοποιήθηκαν από οργανισμούς όπως η Ομοσπονδιακή Επιτροπή Επικοινωνιών (Federal Communications Commission – FCC) στην Αμερική κατέδειξαν ότι μεγάλες περιοχές του ήδη αδειοδοτημένου φάσματος παραμένουν ανενεργές για σημαντικά χρονικά διαστήματα σε ορισμένες γεωγραφικές περιοχές. Μια νέα επαναστατική τεχνολογία που αποσκοπεί στην αποδοτικότερη χρησιμοποίηση του φάσματος είναι οι Γνωστικές Επικοινωνίες (Cognitive Radio). Η τεχνολογία αυτή θα υποστηρίζει “έξυπνα” τερματικά τα οποία θα είναι ενήμερα για το ασύρματο περιβάλλον τους και, ανάλογα με τις επικρατούσες συνθήκες και τις ανάγκες των χρηστών θα προσαρμόζουν κάποιες παραμέτρους της μετάδοσής τους, με πιο σημαντική την μπάντα μετάδοσης. Με άλλα λόγια, ένα Cognitive Radio θα ανιχνεύει το φάσμα και θα εντοπίζει φασματικές οπές (spectrum holes), περιοχές δηλαδή του φάσματος που τη δεδομένη στιγμή δεν χρησιμοποιούνται από τον πρωταρχικό χρήστη τους, και θα χρησιμοποιεί αυτές τις οπές για να μεταδώσει πληροφορία. Επιπλέον, το Cognitive Radio θα είναι ικανό να αναγνωρίζει ακριβώς τα συστήματα επικοινωνίας που υπάρχουν γύρω του (3G, WLAN,...) και θα μπορεί να συνδέεται σε αυτά, εφ' όσον ο χρήστης διαθέτει την κατάλληλη άδεια. Από τα παραπάνω γίνεται φανερό ότι ένα πολύ σημαντικό κομμάτι των γνωστικών επικοινωνιών είναι η ανίχνευση του φάσματος (spectrum sensing). Έχουν προταθεί αρκετοί αλγόριθμοι οι οποίοι είτε ανιχνεύουν την παρουσία πρωτεύοντος χρήστη, είτε κάνουν μια πιο λεπτομερή εκτίμηση του φάσματος αποσκοπώντας στην ταυτοποίηση του παρόντος τηλεπικοινωνιακού συστήματος. Επίσης ενδιαφέρον παρουσιάζει και η δυνατότητα συνεργασίας μεταξύ πολλών χρηστών κατά την ανίχνευση, η οποία έχει αποδειχθεί ότι παρέχει ανοσία σε φαινόμενα όπως multipath fading και shadowing. Σε αυτή την εργασία μελετάται και υλοποιείται μία τεχνική ανίχνευσης φάσματος και ταυτοποίησης σήματος, η οποία αξιοποιεί την a priori διαθέσιμη πληροφορία για τα πρωτεύοντα σήματα (εύρος ζώνης, κεντρική συχνότητα) για να αναγνωρίσει τον τύπο του σήματος. Η τεχνική εφαρμόζεται επίσης σε ένα συνεργατικό σενάριο, όπου πολλοί δευτερεύοντες χρήστες ανταλλάσσουν πληροφορία με στόχο την ακριβέστερη εκτίμηση του φάσματος. Διαπιστώνεται ότι η τεχνική καταφέρνει να διακρίνει μεταξύ διαφορετικών σημάτων, ακόμα και όταν αυτά επικαλύπτονται μερικώς στη συχνότητα. Επιπλέον, η συνεργασία οδηγεί σε μεγαλύτερη πιθανότητα ανίχνευσης και σε λιγότερα σφάλματα ταυτοποίησης. / In recent years, there has been a rapid increase in the number of wireless telecommunications systems and relevant applications. After these developments, the traditional system of licensing and exclusive use of the radio spectrum is driven to its limits, since very few regions of the spectrum are free anymore. However, a number of measurements performed by organizations such as the Federal Communications Commission (FCC) in the USA have shown that large regions of licensed spectrum remain idle for significant portions of time, in certain geographic areas. Cognitive Radio is a new, revolutionary technology that aims in more efficient use of the spectrum. This technology supports “intelligent” terminals which are aware of their wireless environment and, depending on present circumstances and user needs they can adjust certain parameters of their transmissions, mainly the transmission band. In other words, a Cognitive Radio senses the radio spectrum and detects spectrum holes, i.e. regions of the spectrum that are currently not used by their primary user, and uses these holes to transmit. In addition, Cognitive Radio is expected to be able to identify the communication systems in its environment and connect to them, as long as the user has proper authorization. It then becomes obvious that spectrum sensing is a very important part of Cognitive Radio. A number of algorithms have been proposed that either detect the presence of a primary user, or perform a more detailed estimation of the spectrum in order to accurately identify the current communication standard. The possibility of cooperation between many users during sensing has also attracted interest, since it has proven to provide immunity against channel effects such as multipath fading and shadowing. In this work, a spectrum sensing and signal identification technique is studied and implemented that takes advantage of a priori information available about the primary systems (signal bandwidth, center frequency), in order to characterize the signal type. The technique is also applied to a collaborative scenario, where many secondary users exchange information to more accurately estimate the spectrum. It is seen that this technique is able to distinguish different signals, even when they partially overlap in frequency. Furthermore, it is shown that cooperation leads to a greater probability of detection and a lower identification error rate.

Wireless communications systems underwent an evolution as the voice oriented applications evolved to data and multimedia based services. Furthermore, current wireless technologies, regulations and the un- derstanding of the technology are insufficient for the requirements of future wireless systems. Along with the rapid rise at the number of users, increasing demand for more communications capacity to deploy multimedia applications entail effective utilization of communications resources. Therefore, there is a need for effective spectrum allocation, adaptive and complex modulation, error recovery, channel estimation, diversity and code design techniques to allow high data rates while maintaining desired quality of service, and reconfigurable and flexible air interface technologies for better interference and fading management. However, traditional communications system design is based on allocating fixed amounts of resources to the user and does not consider adaptive spectrum utilization. Technologies which will lead to adaptive, intelligent, and aware wireless communications systems are expected to come up with consistent methodologies to provide solutions for the capacity, interference, and reliability problems of the wireless networks. Spectrum sensing feature of cognitive radio systems are a step forward to better recognize the problems and to achieve efficient spectrum allocation. On the other hand, even though spectrum sensing can constitute a solid base to achieve the reconfigurability and awareness goals of next generation networks, a new perspective is required to benefit from the whole dimensions of the available electro hyperspace. Therefore, spectrum sensing should evolve to a more general and comprehensive awareness providing a mechanism, not only as a part of CR systems which provide channel occupancy information but also as a communication environment awareness component of dynamic spectrum access paradigm which can adapt sensing parameters autonomously to ensure robust identification and parameter estimation for the signals over the monitored spectrum. Such an approach will lead to recognition of communications opportunities in different dimensions of spectrum hyperspace, and provide necessary information about the air interfaces, access techniques and waveforms that are deployed over the monitored spectrum to accomplish adaptive resource management and spectrum access. We define multidimensional signal analysis as a methodology, which not only provides the information that the spectrum hyperspace dimension in interest is occupied or not, but also reveals the underlaying information regarding to the parameters, such as employed channel access methods, duplexing techniques and other parameters related to the air interfaces of the signals accessing to the monitored channels and more. To achieve multidimensional signal analysis, a comprehensive sensing, classification, and a detection approach is required at the initial stage. In this thesis, we propose the multidimensional signal analysis procedures under signal identification algorithms in time, frequency. Moreover, an angle of arrival estimation system for wireless signals, and a spectrum usage modeling and prediction method are proposed as multidimensional signal analysis functionalities.

La thèse propose un système embarqué hybride pour l'acquisition de données. Le système se compose d'un "Smartphone" couplé à un micro-contrôleur de type MBED doté d'une interface bus CAN et d'une carte mémoire SD. Selon les besoins de la recherche,on peut ajouter des capteurs ad-hoc, installés sur le véhicule par exemple, en plus des capteurs présents dans les "smartphones" récents. Le postulat est que l'on peut bénéficier des capteurs présents dans les "Smartphones" pour réduire la complexité et le coût de l'instrumentation tout en obtenant une précision de mesure acceptable, et ainsi permettre un déploiement à large échelle du système d'instrumentation. Un tel instrument de mesure a pour objectif de permettre des applications variées dans le domaine des transports routiers (étude des comportements de conduite, contrôle des flux, ...). Une méthode a été implémentée pour identification des performances des capteurs embarqués dans divers smartphones. Des travaux ont été conduit pour la détection "en ligne" des défaillances de capteurs, et la reconnaissance "hors ligne" de manœuvres réalisées par le conducteur, l'objectif étant, à terme, de reconnaître automatiquement des manœuvres typiques telles que : la prise de virages, la prise de rond-points; les manœuvres d'évitement. / In this thesis, we propose a new technic to identify a hybrid system for Data Acquisition,by using ad-hoc sensors on the vehicle, the sensors in the recent smartphones, MBED and CAN-BUS. The assumption is that the Smartphone's sensors will reduce the complexity and the high cost of these instrumentations. The objective is obtaining acceptable measurement accuracy of the collected trajectories and enable for a large-scale deployment of the system's instrumentation, such as a helpful system in the domain of transport. Weshow in this thesis how to build a hybrid system by depending on the properties of the used sensors in both the smartphones and in the vehicles to identify several situation like a failure sensor, accident situation and Rider's behaviour. This system is tested and evaluated on several real time on line - off- line including the used mode and method.

This project studies the signal detection and identification problem in high-dimensional noisy data and the possibility of using it on microbiome data. An extensive simulation study was performed on generated data using as well as a microbiome dataset collected on patients with Parkinson's disease, using Donoho and Jin's Higher criticism, Jager and Wellner's phi-divergence-based goodness-of-fit-test and Stepanova and Pavlenko's CsCsHM statistic . We present some novel approaches based on established theory that perform better than existing methods and show that it is possible to use the signal identification framework to detect differentially abundant features in microbiome data. Although the novel approaches produce good results, they lack substantial mathematical foundations and should be avoided if theoretical rigour is needed. We also conclude that while we have found that it is possible to use signal identification methods to find abundant features in microbiome data, further refinement is necessary before it can be properly used in research. / Detta projekt studerar signaldetekterings- och identifieringsproblemet i högdimensionell brusig data och möjligheten att använda det på mikrobiomdata från människor. En omfattande simuleringsstudie utfördes på genererad data samt ett mikrobiomdataset som samlats in på patienter med Parkinsons sjukdom, med hjälp av ett antal goodness-of-fit-metoder: Donoho och Jins Higher criticis , Jager och Wellners phi-divergenser och Stepanova och Pavelenkos CsCsHM. Vi presenterar några nya tillvägagångssätt baserade på vedertagen teori som visar sig fungera bättre än befintliga metoder och visar att det är möjligt att använda signalidentifiering för att upptäcka olika funktioner i mikrobiomdata. Även om de nya metoderna ger goda resultat saknar de betydande matematiska grunder och bör undvikas om teoretisk formalism är nödvändigt. Vi drar också slutsatsen att medan vi har funnit att det är möjligt att använda signalidentifieringsmetoder för att hitta information i mikrobiomdata, är ytterligare experiment nödvändiga innan de kan användas på ett korrekt sätt i forskning.