A Novel Approach to Modulation Classification in Cognitive Radios

Open Access
Author:
Azarmanesh, Okhtay
Graduate Program:
Electrical Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
May 09, 2011
Committee Members:
  • Sven G Bilen, Dissertation Advisor
  • Sven G Bilen, Committee Chair
  • Julio Urbina, Committee Member
  • James Kenneth Breakall, Committee Member
  • John F Doherty, Committee Member
  • John D Mitchel, Committee Member
  • Kyusun Choi, Committee Member
Keywords:
  • modulation classification
  • cognitive radios
  • Software-defined radios
  • OFDM
Abstract:
Thesis Statement: In cognitive radios, Blind Modulation Classification is an important intermediate step between signal detection and demodulation. There has been an increasing need and thus search for optimal modulation classifiers due to ever increasing variety of digital modulations. The modulation classification technique discussed here is being designed for a real-time Software-Defined Radio (SDR) system to be implemented on SDR development boards and it is robust and efficient with a processing time overhead low enough to allow the software radio to maintain its real-time operating objectives. We are investigating classification of digital single-carrier modulations as well as multi-carrier modulations. The method is to use the waveform's In-phase--Quadrature (I--Q) diagrams and, by employing clustering algorithms on them, determine the type of modulation being transmitted. For classifying single-/multi-carrier modulations, we further study existing methods to find the appropriate Gaussianity test to classify single-carrier signals from multi-carrier ones. This technique may require a lot of processing power; however, with today's technology, it is feasible. Thus, we try to find the test with the best error rate and least amount of processing time. We also include this algorithm with methods to extract the features of an Orthogonal Frequency Division Multiplexing (OFDM) signal in the case of multi-carrier modulations. We also utilize the clustering algorithms to propose a new method in correcting frequency imbalances and I--Q offsets in OFDM signals as a mean of signal synchronization. This new Modulation Classification method will be capable of determining the type of modulation scheme among different PAM, PSK, QAM, and OFDM modulations and can be further expanded to include any new modulation scheme.