DSP

带限信号的抽样与恢复

2019-07-13 20:00发布

class="markdown_views prism-github-gist">

带限信号的抽样与恢复

  x(t)x(t)为带限信号,即对于fW,X(f)=0vert f vert geq W, X(f) = 0根据奈奎斯特抽样定律,只要抽样频率1Ts2Wfrac{1}{T_s} geq 2W即可无损恢复出信号x(t)x(t)。且 x(t)=2WTsn=x(nTs)sinc(2W(tnTs))x(t)=2W^{'}T_ssum_{n=-infty}^{infty} x(nT_s)sinc(2W^{'}(t-nT_s))
上式中,WW^{'}满足: WW1TsWWleq W^{'}leq frac{1}{T_s}-W
当以奈奎斯特频率进行抽样时,即1Ts=2Wfrac{1}{T_s}=2W时,x(t)x(t)可进一步化简为 x(t)=n=x(n2W)sinc(tTsn)x(t)=sum_{n=-infty}^{infty}x(frac{n}{2W})sinc(frac{t}{T_s}-n) 证明:
  当以1Tsfrac{1}{T_s}的频率对x(t)x(t)进行抽样,得到的抽样信号为xδ(t)x_{delta}(t),则: xδ(t)=n=x(nTs)δ(tnTs)=x(t)n=δ(tnTs) egin{aligned} x_delta(t) &=sum_{n=-infty}^{infty}x(nT_s) delta(t-nT_s) \ &=x(t)sum_{n=-infty}^{infty}delta(t-nT_s) end{aligned} 所以: F[xδ(t)]=F[x(t)n=δ(tnTs)]=X(f)F[n=δ(tnTs)] egin{aligned} F[x_{delta}(t)] &=F[x(t)sum_{n=-infty}^{infty}delta(t-nT_s)] \ &=X(f)*F[sum_{n=-infty}^{infty}delta(t-nT_s)] end{aligned}
F[n=δ(tnTs)]=1Tsn=δ(fnTs)F[sum_{n=-infty}^{infty}delta(t-nT_s)]=frac{1}{T_s}sum_{n=-infty}^{infty}delta(f-frac{n}{T_s})
所以 F[xδ(t)]=1Tsn=X(fnTs)F[x_{delta}(t)]=frac{1}{T_s}sum_{n=-infty}^{infty}X(f-frac{n}{T_s})
  要从xδ(t)x_{delta(t)}中恢复x(t)x(t),从频谱上看,只要通过一个在fWvert f vert leq W上增益为TsT_s