Date of Award

9-16-2019

Publication Type

Doctoral Thesis

Degree Name

Ph.D.

Department

Electrical and Computer Engineering

Keywords

Detection Probability, Dynamic Spectrum Access, False Alarm Probability, Smart Spectrum Sharing, TVWS

Supervisor

Tepe, K.

Rights

info:eu-repo/semantics/openAccess

Abstract

The goal of Cognitive Radio (CR) is to facilitate efficient utilization of the electromagnetic spectrum. CR applies spectrum sensing techniques to detect unused channels and then allows opportunistic usage of such channels by secondary users, i.e., un-licensed users, without interfering with primary users, i.e., licensed users. In order to implement a complete TV White Spaces (TVWS) based CR system, at first a model is needed that can be used for identifying TVWS, which can then be exploited for dynamic spectrum access. This work is focused on proposing a sensing method and building a probabilistic model for identifying the occupancy of the electromagnetic spectrum within the UHF TV bands. It also develops a hardware prototype for demonstrating the performance of the proposed technique. It proposes simultaneous sensing both noise and noise-contaminated user's signal (composite signal) for detecting spectrum occupancy minimizing errors. The proposed sensing technique combines energy detection, pilot detection, and information obtained from an external source in order to reduce missed-detection probability. In addition to pre-defined threshold levels, the proposed probabilistic model considers parameters like probability of false alarm and probability of detection for measuring detection accuracy. Finally, a mobile sensing station is designed and implemented using off-the-shelf components to verify the developed technique for TVWS spectrum sensing. Using this mobile station, the UHF TV channels within the spectrum band of 500MHz-698MHz (Channel #19 to Channel 51) are scanned. Covering the total bandwidth of 198MHz, over 8 million data samples are collected through repeated scanning, ensuring possible spatio-temporal variations are taken into account. Results show that the availability of TVWS changes quite significantly with spatial variations. But, even in the most crowded spectrum locations, 28% of UHF channels were identified as TVWS. The model demonstrates about 10% improvement in detecting accuracy compared to other existing models.

Download

Share

COinS