Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Ayush_thesis.pdf (2.36 MB)

ETD Abstract Container

Abstract Header

Comparison of Statistical Signal Processing and Machine Learning Algorithms as Applied to Cognitive Radios

Tiwari, Ayush
http://rave.ohiolink.edu/etdc/view?acc_num=toledo1533218513862248

Abstract Details

2018, Master of Science, University of Toledo, Electrical Engineering.
Software defined radio (SDR) systems have attracted much attention recently for their affordability and simplicity for hands-on experimentation. They can be used for implementation of dynamic spectrum allocation (DSA) algorithms in cognitive radio (CR) platform. There has been a massive research in the DSA algorithms both in machine learning and signal processing paradigm, but, these CRs are still incapable to decide which algorithm suites for specific scenario. A comparison between the spectrum sensing algorithms using machine learning techniques and statistical signal processing techniques is needed in order to know which algorithm suits best for resource constrained environments for CRs and spectrum observatories. Two challenges; namely, multi-transmitter detection and automatic modulation classification (AMC) are chosen. Novel machine learning based and statistical signal processing based multi-transmitter detection algorithm are proposed and used in the comparison. After comparing accuracy, for multi-transmitter detection, machine learning algorithm has accuracy of 70% and 80% for 2 and 5 user system, respectively, whereas, the accuracy for statistical signal processing algorithm is 50% for 2 and 5 user system. For AMC, both signal processing and machine learning algorithm have a perfect accuracy beyond 10 dB for 100 test samples (64-QAM being an exception) but for 1000 test samples, the machine learning algorithm outperforms the signal processing algorithm. Time comparison showed that signal processing algorithms, in both cases, take fraction of the time required by machine learning algorithms. Hence, it is recommended to use machine learning techniques where accuracy is important and use signal processing approach where timing is important. The process of selecting the algorithms can be regarded as a tradeoff between accuracy and time.
Vijay Devabhaktuni (Committee Chair)
Harshavardan Chenji (Committee Co-Chair)
Ahmad Javaid (Committee Member)
77 p.
Electrical Engineering
Machine Learning, Signal Processing, Cognitive Radios, Software Defined Radios, Dynamic Spectrum Access, Automatic Modulation Classification,

Recommended Citations

Citations

  • Tiwari, A. (2018). Comparison of Statistical Signal Processing and Machine Learning Algorithms as Applied to Cognitive Radios [Master's thesis, University of Toledo]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=toledo1533218513862248

    APA Style (7th edition)

  • Tiwari, Ayush. Comparison of Statistical Signal Processing and Machine Learning Algorithms as Applied to Cognitive Radios. 2018. University of Toledo, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=toledo1533218513862248.

    MLA Style (8th edition)

  • Tiwari, Ayush. "Comparison of Statistical Signal Processing and Machine Learning Algorithms as Applied to Cognitive Radios." Master's thesis, University of Toledo, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=toledo1533218513862248

    Chicago Manual of Style (17th edition)

toledo1533218513862248
1,039
© 2018, some rights reserved.Creative Commons License Comparison of Statistical Signal Processing and Machine Learning Algorithms as Applied to Cognitive Radios by Ayush Tiwari is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by University of Toledo and OhioLINK.