A wavelet-based architecture for ultra wideband system with spectrum sensing

Abstract

Ultra Wideband (UWB) system possesses attractive features such as high data rate, low cost, and low power consumption. However, UWB wide range of frequency band causes mutual interference with other narrowband systems. The main goal of the work is to develop UWB system that can mitigate interference. In this thesis a Wavelet Cognitive UWB System has been designed that can minimize interference and enhance system performance. The proposed system is composed of Wavelet Packet Multicarrier Modulation (WPMCM), wavelet spectrum sensing, and spectrum and power allocation. WPMCM is deployed since it is flexible and inherently robust against interference mitigation. The WPMCM employs Inverse Wavelet Packet Transform (IWPT) and WPT engine to play the same role as Inverse Fast Fourier Transform (IFFT) and FFT in Orthogonal Frequency Division Multiplexing (OFDM) system. Wavelet Spectrum Sensing is designed to minimise interference by sensing the whole spectrum band and make decision to deactivate the occupied subcarriers. The proposed wavelet spectrum sensing has the capability to detect interfering occupied subcarriers. It uses energy detection for spectrum analysis and decision. Enhanced Forward Consecutive Mean Excision (E-FCME) algorithm is used to provide decision threshold whether to use the subcarriers. Spectrum and power allocation has been developed to allocate the transmission power to the selected subcarriers according to channel estimation gain. In addition, an efficient pilot pattern strategy is employed in the channel estimation to enhance the estimated channel gain. The optimum power allocation is derived by Lagrange multiplier method to minimize the Bit Error Rate (BER) at the constraint of UWB power limit. As a benchmark, the proposed system is compared to the conventional FFT based system in different UWB channel models (CM1-CM4) using MATLAB simulation. Simulation results show enhancement in side-lobes suppression of WPMCM about 20 . The result proves significant improvements, while primary and secondary links are subjected to multipath fading and noise. The probability of detection average is increased from 0.87 to 0.99, and probability of false alarm is reduced and controlled around 0.02 in the spectrum sensing phase. At =25 , the achieved average BER is reduced from 4.4×10 to 3.95×10 for the proposed spectrum and power allocation. In general, the numerical results verify that the proposed system outperforms the traditional system with various metrics of performance analysis.