As link quality in 60GHz WLAN is significantly affected by moving human body shadowing, it is required to introduce shadowing countermeasures such as diversity and fast session transfer. For performance evaluation of these techniques in various environments, we need a deterministic shadowing model for ray-tracing simulations rather than a traditional empirical shadowing model based on experimental data. With this motivation, this paper proposes a twin cylinder model for moving human body shadowing in 60GHz WLAN, where three knife edges for diffraction loss calculation are derived from geometrical positions of a moving human body, a transmitter and a receiver. It is confirmed that calculated results using the proposed model agrees to measured results.

This paper describes performance evaluation results of SSB (Single Side Band) transmission of DFTs-OFDM (Discrete Fourier Transform Spreading Orthogonal Frequency Division Multiplexing) using multi-level BPSK (Binary Phase Shift Keying) through nonlinear HPA (High Power Amplifier). SSB-DFTs-OFDM is generated by eliminating USB (Upper Side Band) or LSB (Lower Side Band) spectrum in frequency domain after DFT spreading of multi-level BPSK signals, thus Hilbert transformer is not necessary to generate SSB signals. The simulation results show that SSB-DFTs-OFDM achieves almost the same PAPR (Peak to Average Power Ratio) and ACLP (adjacent channel power leakage) as conventional DFTs-OFDM and lower PAPR and ACLP than OFDM and required bandwidth can be reduced by half when BPSK modulation is employed. Furthermore, SSB-DFTs-OFDM using multi-level BPSK shows less error floor and/or Eb/N0 degradation compared with conventional DFTs-OFDM using QAM modulation in some cases when HPA OBO (Output Back Off) is small. SSB-DFTs-OFDM is an attractive alternative to DFTs-OFDM for WLAN and cellular applications.

This study considers blind channel shortening to mitigate the effects of inter-block interference and multiple-access interference in both uplink and downlink multicarrier code-division multiple-access systems. We derive the blind channel shortening methods by extending the blind channel equalization method previously proposed by Djebbar, Abed-Meraim, and Djebbari. The basic idea is to restore, at the frequency-domain equalizer output, the orthogonality of spreading codes, which is destroyed when interference is present. We show that the extended methods can perfectly shorten the channels and that the resulting effective channel is identical to the predetermined target channel thus avoiding channel estimation. Computer simulation results illustrate the effectiveness of the proposed methods.