A hardware based Ricean fading radio channel simulator

Abstract

As the signal propagates from the transmitter to the receiver, it is subjected to fading. One of the fading distributions is Ricean distribution that assumes a line of sight signal and an infinite number of multipath signals between the transmitter and the receiver. The key parameter in the Ricean distribution is the K factor which is a measure of the severity of fading; the lower the K value, the more severe the fading. The problem was to develop a hardware based fading simulator that is based on the Ricean distribution. The key component of the hardware is an eight channel beamformer board; each channel has a 6-bit attenuator and a 6-bit phase shifter controlled by a on board FPGA for beamforming purpose. To approximate the Ricean fading, one of the eight channels was designated the line of sight signal, and the seven other channels were used to emulate the multipath signals. Cables with different lengths were used as delay lines to emulate the actual delays that occur to the multipath signals. The signals were then normalized to account for the cable losses. The fade simulator was controlled using a MATLAB program that sent a serial bit stream to the attenuators and phase shifters on the beamformer board. The MATLAB program generated random attenuations and phase shifts according to the Ricean distribution for the inputted K value. For a constant amplitude input RF signal, as K was increased, the fluctuation in the signal at the output of the fade simulator decreased. The estimated K values from the collected data were close to the inputted K values. The probability density functions (PDFs), the cumulative density functions (CDFs) and the level crossing rates (LCRs) of the collected data compared closely to the theoretical Ricean fading distribution PDFs, CDFs and LCRs, which showed that the fade simulator provides an accurate simulation of Ricean fading. The fade simulator was used to evaluate the performance of Airspan and Harris radios. The throughput of the radios was tested under different K values and for different SNR. As K increased, the fluctuation in the signal decreased which led to higher throughput.