Ημερολόγιο Εκδηλώσεων

Περιγραφή:
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών Πρόγραμμα Προπτυχιακών Σπουδών ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ Εμμανουήλ Φουντουλάκη με θέμα Παρακολούθηση Υποχώρου για 'Ενθετες Συστοιχίες κεραιών Subspace Tracking for Nested Arrays Εξεταστική Επιτροπή Αναπληρωτής Καθηγητής Γεώργιος Ν. Καρυστινός (επιβλέπων) Αναπληρωτής Καθηγητής 'Αγγελος Μπλέτσας Καθηγητής Βασίλειος Διγαλάκης Abstract In radar, sonar, and mobile communications, the estimation of the directions from which multiple signals arrive at a point is called the direction of arrival (DoA) estimation problem and, over the past decades, has been performed often through uniform linear arrays (ULAs) in conjunction with high-resolution subspace-based algorithms. Such techniques, however, have limited capability of the number of directions they can estimate; if the ULA consists of N antenna elements, then high-resolution subspace-based algorithms can estimate the directions of up to N-1 signals. To increase this number, a novel structure that consists of N antenna elements and enables the estimation of O(N^2) signal directions has been developed recently. It lies on specific nonuniform-array structures which are called nested arrays. In this thesis, we first overviewed the structure and properties of nested arrays and evaluated their performance through computer simulations. We observed that the nested arrays with conventional subspace-based signal-processing algorithms can offer high performance but require high complexity, which makes them impractical for real-time applications. Then, we developed novel subspace tracking techniques for nested arrays that have lower complexity, are proven to converge to the optimal, subspace-based estimator, and are capable of tracking changes in the directions of the arriving signals (for example, when a source is moving with respect to the array receiver).