12.07.2015 Views

a non-linear scheme for pmepr reduction in mc-cdma system - IJCSET

a non-linear scheme for pmepr reduction in mc-cdma system - IJCSET

a non-linear scheme for pmepr reduction in mc-cdma system - IJCSET

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

S.Tamilselvan et al./ International Journal of Computer Science & Eng<strong>in</strong>eer<strong>in</strong>g Technology (<strong>IJCSET</strong>) ; , ,… (1)• The Inverse Discrete Fourier Trans<strong>for</strong>m (IDFT) of this frequency-doma<strong>in</strong> block is computed, lead<strong>in</strong>g to thetime doma<strong>in</strong> block { ; n=0,1,...N-1} with is given Eqn.(2). ∑ (2) • Each sample is submitted to a <strong>non</strong><strong>l<strong>in</strong>ear</strong> operation(an envelope clipp<strong>in</strong>g)accord<strong>in</strong>g to, lead<strong>in</strong>g to themodified block { ; n=0,1,…,N’-1}, where is given Eqn.(3). | | (3)• A discrete Fourier trans<strong>for</strong>m (DFT) br<strong>in</strong>gs the <strong>non</strong><strong>l<strong>in</strong>ear</strong>ly modified block back to the frequency doma<strong>in</strong>,where a shap<strong>in</strong>g operation is per<strong>for</strong>med by a multiplier bank with selected coefficients, {Gk; k=0,1,…,N’-1}, so as to obta<strong>in</strong> the block , with is given by the Eqn.(4). ;,,…. (4)• The f<strong>in</strong>al frequency-doma<strong>in</strong> block results from by remov<strong>in</strong>g the zeros, where is shown <strong>in</strong>Eqn.(5). ;,,…….. (5)• Append<strong>in</strong>g zeros to each <strong>in</strong>put block prior to comput<strong>in</strong>g the required IDFT is a well-known OFDMimplementation technique, which is equivalent to oversampl<strong>in</strong>g, by a factor given <strong>in</strong> Eqn.(6) below, theideal MC-CDMA burst. (6)• The subsequent <strong>non</strong><strong>l<strong>in</strong>ear</strong> operation is crucial <strong>for</strong> reduc<strong>in</strong>g the envelope fluctuations, whereas the frequencydoma<strong>in</strong>filter<strong>in</strong>g us<strong>in</strong>g the set given <strong>in</strong> Eqn.(7) can reduce the result<strong>in</strong>g spectral spread<strong>in</strong>g (of course, withsome regrowth of the envelope fluctuations). ;,,,……, (7)The removal of subcarriers with zero amplitude reduces the computational ef<strong>for</strong>t and corresponds todecimation <strong>in</strong> the time doma<strong>in</strong>. For a given and a careful selection of , the <strong>non</strong><strong>l<strong>in</strong>ear</strong> characteristic (<strong>for</strong> a given<strong>in</strong>put level) and the set can {Gk} ensure small envelope fluctuations while ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g low out-of-bandradiation and <strong>in</strong>-band self-<strong>in</strong>terference levels. When the <strong>non</strong><strong>l<strong>in</strong>ear</strong> operation is chosen to be an ideal envelopeclipp<strong>in</strong>g, with clipp<strong>in</strong>g. , , (8)It should be mentioned that this wide class of signal-process<strong>in</strong>g <strong>scheme</strong>s <strong>in</strong>cludes, as specific cases, <strong>scheme</strong>sproposed so far. where the same clipp<strong>in</strong>g is adopted when assum<strong>in</strong>g ,the condition <strong>in</strong> Eqn.(8) with <strong>for</strong> the <strong>in</strong>bandsubcarriers, and out-of-band. It should also be mentioned that, <strong>for</strong> an ideal envelope clipp<strong>in</strong>g, the proposedsignal-process<strong>in</strong>g <strong>scheme</strong> can be shown to be equivalent to the peak cancellation method.A more sophisticated technique, allow<strong>in</strong>g improved PMEPR reduc<strong>in</strong>g results, could be simply developed onthe basis of the signal-process<strong>in</strong>g approach described above. Such a technique consists of repeatedly us<strong>in</strong>g, <strong>in</strong> aniterative way, the signal-process<strong>in</strong>g cha<strong>in</strong> which leads from to <strong>in</strong> Fig.1. The technique proposed <strong>in</strong> [17]corresponds to the particular case where the <strong>non</strong><strong>l<strong>in</strong>ear</strong> operation is an envelope clipp<strong>in</strong>g, and the frequencydoma<strong>in</strong>filter<strong>in</strong>g is characterized by <strong>for</strong> the Gk=0 <strong>for</strong> the out-of-band subcarriers.ISSN : 2229-3345 Vol. 3 No. 3 March 2012 16

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!