Efficient Wideband Digital Front-End Transceivers for Software Radio Systems
Abu-Al-Saud, Wajih Abdul-Elah
MetadataShow full item record
Software radios (SWR) have been proposed for wireless communication systems to enable them to operate according to incompatible wireless communication standards by implementing most analog functions in the digital section on software-reprogrammable hardware. However, this significantly increases the required computations for SWR functionality, mainly because of the digital front-end computationally intensive filtering functions, such as sample rate conversion (SRC), channelization, and equalization. For increasing the computational efficiency of SWR systems, two new SRC methods with better performance than conventional SRC methods are presented. In the first SRC method, we modify the conventional CIC filters to enable them to perform SRC on slightly oversampled signals efficiently. We also describe a SRC method with high efficiency for SRC by factors greater than unity at which SRC in SWR systems may be computationally demanding. This SRC method efficiently increases the sample rate of wideband signals, especially in SWR base station transmitters, by applying Lagrange interpolation for evaluating output samples hierarchically using a low-rate signal that is computed with low cost from the input signal. A new channelizer/synthesizer is also developed for extracting/combining frequency multiplexed channels in SWR transceivers. The efficiency of this channelizer/synthesizer, which uses modulated perfect reconstruction (PR) filter banks, is higher than polyphase filter banks (when applicable) for processing few channels, and significantly higher than discrete filter banks for processing any number of variable-bandwidth channels where polyphase filter banks are inapplicable. Because the available methods for designing modulated PR filter banks are inapplicable due to the required number of subchannels and stopband attenuation of the prototype filters, a new design method for these filter banks is introduced. This method is reliable and significantly faster than the existing methods. Modulated PR filter banks are also considered for implementing a frequency-domain block blind equalizer capable of equalizing SWR signals transmitted though channels with long impulse responses and severe intersymbol interference (ISI). This blind equalizer adapts by using separate sets of weights to correct for the magnitude and phase distortion of the channel. The adaptation of this blind equalizer is significantly more reliable and its computational requirements increase at a lower rate compared to conventional time-domain equalizers making it efficient for equalizing long channels that exhibit severe ISI.