National Technical University of Athens (NTUA)
Σύνθεση στοχαστικών διαφορικών εξισώσεων 1ης τάξης για τη μοντελοποίηση στάσιμων τυχαίων διαδικασιών με έμφαση στην απόσβεση λόγω βροχής που επιβαρύνει τις μικροκυματικές ζεύξεις
Abstract
dc:descriptionThis thesis focuses on the Synthesis of 1st Order Stochastic Differential Equations for the simulation of Actual, Stationary Baseband Random Processes such as Rain Attenuation which affects microwave links operating at frequencies over 10GHz. Hereafter, the term '‘Actual Random Process” will denote Random Processes of the aforementioned kind. The term “Baseband” implies that the Spectral Density Function of the particular Stationary Random Process gets its maximum value at zero frequency and it is defined within a frequency window which is symmetrical around the origin. The established Synthesis procedure is modified so that 1st Order Stochastic Differential Equations simulate efficiently not only the 1st and 2nd order statistics but the statistical characteristics of Hitting Times as well, all three belonging to an Actual Random Process. Note that the 1st and 2nd Order Statistics imply respectively the Probability Density Function and the Autocorrelation Function or equivalently, the Spectral Density Function while Hitting Times are Random Variables which are similar to Fade Duration. As a consequence of the modified procedure, the Stochastic Differential Equation reproduces better than previously the generation mechanism of the Actual Process. The new condition added, clarifies in more detail how the generation mechanism of the particular Actual Random Process is “diffracted” through the simulating Stochastic Differential Equation and this in turn allows one to choose more realistically the values for the input parameters so that the “deformation” on the generation mechanism is minimized. In other words, the more the performance of the simulation is printed in detail, the more it becomes feasible to focus on the simulation of particular statistical characteristics such as the Hitting Times around a critical (attenuation) threshold. This is highly beneficial since the simulation becomes more selective and accurate. All these ideas are used in practice to construct three dynamical models which in turn are based on three separate Stochastic Differential Equations of 1st order. These models aim at simulating attenuation due to rain which in the long-term is statistically described by one of the following three Probability Density Functions: the Lognormal, the Weibull or the Gamma. All three models are fully developed in the sense that all necessary quantities are calculated within the thesis and are provided in user-friendly Tables and Appendices so that one can use them without necessarily having to work on cumbersome calculations and mathematical notions. The aforementioned three models are employed within the thesis, for the simulation of rain attenuation which affects an actual microwave link. Rain attenuation time series concerning that particular link are derived by the Synthetic Storm Technique and thus they are considered to be nearly experimental. The aforementioned Technique is quite popular among engineers working in the E/M propagation area. It accepts as “input” rain rate time series measured by an appropriate piece of equipment planted near the receiver antenna (earth-station in case of satellite links) or somewhere else along the trace of the radio path on the ground and it returns as “output” the simulated rain attenuation time series due to the measured rainfall. Rain rate time series were collected by a disdrometer during the period 1st of January 2000 until 31st of May 2002. The collected time series were shown to be representative of the actual phenomenon on an annual time period. Further, they were inserted as “input” in the Synthetic Storm Technique and then the simulated rain attenuation time series extracted, were statistically elaborated in order to obtain the 1st, 2nd and Hitting Time statistics, all three belonging to rain attenuation affecting the particular link. Out of the three dynamic models, the one corresponding to the Lognormal Distribution was shown to be more reliable for the simulation of the rain attenuation affecting the link.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2007
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kanellopoulos, S.
- Κανελλόπουλος, Σωτήριος
Subjects
dc:subject × 17- Διαφορικές εξισώσεις, Στοχαστικές
- Σύνθεση
- Εξίσωση διάχυσης Fοkker - Planck
- Μικροκυματικές ζεύξεις μεγάλων αποστάσεων
- Τεχνικές άμβλυνσης διαλείψεων
- Αυτόματος έλεγχος
- Differential equations
- Synthesis
- Fokker planck equation
- Attenuation due to rain
- Microwave links, Long distance
- Fade mitigation techniques
- Automatic control
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/16866