An Efficient Packet Detection Method with a Verification Mode in MB-OFDM UWB Systems

Abstract

In this paper, we propose a novel packet detection method using a cross-correlation scheme with a verification mode in ultra wide-band (UWB) systems adopting multi-band OFDM (MB-OFDM) techniques. In the packet detection process, a cross-correlation scheme has the advantage of performing the acquisition of transmitted time-frequency code (TFC) for frequency hopping as well as packet detection with low hardware costs; however, it suffers from performance degradation caused by an imperfect decision process using a threshold value compared with an auto-correlation method. By using a verification mode, the proposed method improves detection performance of a cross-correlation technique considerably. This is because the verification mode can reduce a false alarm probability on packet detection by reconfirming whether or not the desired signal is received and can effectively reject temporal small correlation peaks due to the noise or sub-path signals having smaller energies. For the verification of proposed algorithm, we employ CM1 and CM4 channel models given in IEEE802.15.3a and compare the detection performance between conventional algorithms as well as the proposed one by using computer simulations. Experimental results show that the proposed scheme can reduce a packet detection error probability (PDER) to about 10⁻³ in CM1 channel and about 10⁻² in CM4 channel at SNR = 9 dB compared with conventional methods having more than 10⁻¹ PDER. From the experimental results, we can find that the proposed algorithm can compensate for performance degradation under high SNR condition caused by using a fixed single threshold value. We also estimate the power consumption of algorithms by utilizing both the total number of multiplications and additions employed in the algorithms and the consuming power of hardware elements under 0.13 um CMOS technology. Next, we discover that the proposed method can reduce computational complexity by a minimum of 24.5% compared with conventional cross-correlation algorithms. From the experimental and power estimation results, we can see that the proposed method is very useful for packet detection in UWB MB-OFDM systems, in which low-power implementation is an important issue.