On Wavelet Based Spectrum Estimation for Dynamic Spectrum Access

On Wavelet Based Spectrum Estimation for Dynamic Spectrum Access

Ariananda, D.D.

Lakshmanan, M.K. (mentor)
Nikookar, H. (mentor)

Electrical Engineering, Mathematics and Computer Science

Telecommunications

2009-08-19

One of the important functionalities of Dynamic Spectrum Access is spectrum estimation. Accuracy and speed of estimation are the key indicators to select the appropriate spectrum estimation technique. In this thesis work, the possibility of employing wavelet packet decomposition as a basis for a new spectrum estimation approach is investigated. Once the new approach is developed, four types of sources, namely partial band, single tone, multi-tones, and swept tone, are used to investigate the performance of the proposed wavelet based approach. Preliminary comparative analysis between the performance of wavelet based approach with conventional techniques, such as Periodogram and Welch technique has also been conducted. The studies show that the wavelet based approach offers great flexibility, reconfigurability and adaptability. Key to the successful operation of the wavelet based spectrum estimation is the choice of the wavelet used. Commonly known wavelets are not suitable for spectrum estimation because they result in estimates with poor frequency resolution. To alleviate this problem, we design and develop a family of wavelets that are maximally frequency selective in nature as our second contribution in this thesis work. To this end, the design constraints are first enlisted. Then the problem, originally non-convex, is reformulated into a convex optimization problem and solved using Semi Definite Programming (SDP) tools. Through simulation studies the benefits of the newly designed wavelets are demonstrated. The next contribution of this thesis work is to combine the existing wavelet packet multi-carrier modulation (WPMCM) technique with our wavelet based spectrum estimator in order to form a wavelet packet transceiver for a dynamic spectrum access environment. To enable the wavelet packet transceiver cognitive radio (CR) system to co-exist with other Licensed Users (LU), a common spectrum pool is maintained and the WPMCM transmission waveform characteristics are shaped to communicate in the idle time-frequency gaps of the licensed user. This is achieved by dynamically vacating wavelet packet carriers in and near the region of the licensed user spectrum. The spectrum estimation unit is tagged to the WPMCM transceiver structure by exploiting the filter bank infrastructure used for Discrete Wavelet Packet Transform implementation. Thus spectrum analysis is done at no additional cost. In the studies, four types of LUs are employed, namely, partial band, single tone, swept and multiple tone. The simulation results show that in the presence of an LU, the proposed spectrum adaptation method offers significant BER improvements allowing the CR to operate invisibly to the LU.

spectrum estimation
Wavelet Packet Transform
frequency selectivity
dynamic spectrum access
filter bank

http://resolver.tudelft.nl/uuid:008a3e63-00cf-40d6-95f8-200762568b35

2009-08-28

Student theses

master thesis

(c) 2009 Ariananda, D.D.