HK1031484A - Wideband channelization using subsampled discrete fourier transforms - Google Patents

Description

Wideband channelization with subsampled discrete fourier transform

Background

Field of the invention

The present invention relates to wideband channelization techniques, and more particularly, to a method for channelizing wideband signals using a sub-sampled discrete fourier transform filter bank.

Description of the related Art

A radio receiver that needs to receive multiple radio channels simultaneously requires many radio channels to be extracted from one wideband signal. Such receivers may include macro base stations, micro base stations, pico base stations, and the like. These types of receivers typically operate according to a frequency reuse scheme, effectively limiting each base station to a uniformly spaced subset of all available channels.

In one prior art implementation, each radio channel is extracted from one wideband signal using a DFT (discrete fourier transform) filter bank. A problem with existing DFT channelizers is extracting each channel from a wideband radio signal. Doing so requires a large number of arithmetic operations by the channelizer and increases the cost/complexity of the receiver. Since each base station utilizes only a uniformly spaced subset of all available channels. It is therefore desirable to have a more efficient, simpler method of extracting a radio channel from a wideband signal.

Summary of The Invention

The present invention overcomes the above and other problems with channelizers used in receivers to process wideband signals. The wideband signal is initially processed by a sub-sampling filter bank that extracts a selected number of evenly spaced channels from a plurality of channels within the received wideband signal. The subsampled DFT-channelizer consists of a bank of polyphase filters that extract all potential channels (M total channels) from the wideband signal. Then time-aliasing the polyphase filter output to produce a second sequence of signals equal in number to the selected number of uniformly spaced information (A desired channel).

ByThe inverse discrete Fourier transform of the point processes the signal of the second sequence to obtainA band pass signal. The inverse discrete fourier transformed coefficients are then mixed with a series of carrier signals to shift the bandpass signals to baseband, with the result being extracted from the wideband signalEach is provided withEvenly spaced channels. Such a system greatly reduces the total amount of processing required. In the system according to the invention, the number of arithmetic operations necessary to generate the desired channel is considerably lower than the number of arithmetic operations currently required to extract each channel.

Brief Description of Drawings

For a more complete understanding of the present invention, reference is made to the following detailed description, taken in conjunction with the accompanying drawings, in which:

fig. 1 is a block diagram of a general broadband receiver;

FIG. 2 is a functional diagram of a single branch in a DFT channelizer;

FIG. 3 is a diagram of a DFT channelizer; and

fig. 4 is a block diagram of a sub-sampling DFT channelizer.

Detailed Description

Referring now to the drawings and more particularly to fig. 1, there is shown a block diagram of a conventional wideband receiver in which a transmitted wideband signal is received at an antenna 5, processed to a desired frequency band by several stages of mixing and filtering (generally shown at 10), then down-mixed by a mixer 15 to a baseband signal x (t), and the analog wideband signal x (t) is converted by an analog-to-digital converter 20 to a digital wideband signal x (n) for input to a wideband analog-to-digital converter 20 with a relatively wide bandwidth, and then processed by a digital channelizer 25 to extract the individual radio channels 30. The prior art DFT channelizer 25 (shown in fig. 3) provides a computationally efficient way to extract each channel within the wideband signal x (n).

Referring now to fig. 2, there is shown a functional diagram of one branch of a DFT channelizer. Ho (w) represents a real low-pass FIR filter, each other filter in the filter bank being a modulated version of the low-pass prototype, and therefore,

i is more than or equal to 0 and less than or equal to M-1, and M is equal to the number of channels. Note that H_i(w) represents a frequency at discrete timeBand-pass complex-valued filters centred or, equivalently, around continuous-time frequenciesAs a center (F)_sIs the sampling frequency of the A/D converter, M is equal to the total number of channels between-Fs/2, + Fs/2). In other words, there are exactly M equal bandwidth filters in the filter bank, spaced at intervals ofThe DFT channelizer of fig. 3 is only valid if M is an integer multiple of the downsampling factor N (i.e., M ═ nxk, where K is some positive integer). The DFT channelizer, typically referred to as a DFT channelizer and shown in fig. 3, can be efficiently implemented using an Inverse Discrete Fourier Transform (IDFT) and polyphase decomposition of a low-pass prototype filter ho (n).

Referring now to fig. 3, there is shown a block diagram of a DFT channelizer, in fig. 3, E_i(z) s represents the multiphase component of Ho (z). Therefore, the temperature of the molten metal is controlled,wherein: e.g. of the type_i(n)＝h_i[nM+i] 0≤i≤M-1

The main limitation of the prior art DFT channelizer is that it is frequency range specificChannelizing each channel; even though only a subset of these channels may actually be needed. For example, in most cellular systems utilizing the 7/21 frequency reuse plan, each base station utilizes only one out of every seven radio channels. Thus, the receiver only needs to channelize every seventh channel.

Referring now to fig. 4, there is shown a block diagram of a sub-sampling DFT channelizer of the present invention. For the subsampled DFT channelizer, it is assumed that only every lth output channel has to be calculated and that the total number of channels M is an integer multiple of L, and therefore,

M＝L×r

where r is some positive integer.

From the discrete wideband signal x (n), the subsampled DFT-channelizer computes only the desired channel.

{c₀[n]，c_L[n]，c_2L[n]，…，c_M-L[n]}

Comparing fig. 4 and fig. 3, we see the time for the subsampled DFT-channelizerAliasing block sumPoint IDFT instead of M-point IDFT, time-aliased block sumThe complexity of the combination of point IDFTs is much less than the complexity of M-point IDFTs.

The output of the time aliasing block is formed from the output of the polyphase filter according to,i is not less than 0 and not more than Q-1, whereinAndK.ltoreq.0.ltoreq.Q-1 in the subsampled DFT channelizer of FIG. 5The Q output of the point IDFT is

{r₀[n]，r_L[n]，r_2L[n]，…，r_m-L[n](i.e., the lth output of the IDFT block in fig. 3).

Also, the final output of the subsampled DFT channelizer in FIG. 5 is

{c₀[n]，c_L[n]，c_2L[n]，…，c_m-L[n](i.e., every lth last output of the DFT channelizer in fig. 3).

For example, let us consider an approximate 10MH_ZAnd let us assume that each radio channel complies with the D-AMPS standard. In particular, the channel spacing is f_cs＝30KH_Z. Let us assume that 7/21 frequency multiplexing is used, and therefore, only every seventh channel needs to be extracted from x (n), i.e., L is 7.

If the sampling frequency of the A/D converter is set to F_s＝34.02MH_ZThen the entire DFT channelizer of fig. 3 can be used for every 30K in x (n)H_ZThe frequency is decimated once. In this case, the total number of channels is. IDFT of size 1134 requires every other IDFT by the DFT channelizerOnce per second. Because 1134 is a high non-prime number, the Dooley-Tukey FFT algorithm can be used to efficiently compute the IDFT.

Alternatively, the subsampled DFT-channelizer of fig. 4 may be used to extract from x (n) only once every 7 th channel (i.e., 34.02MH if the sampling frequency of the a/D converter is set to Fs ═ 34.02MH_ZAnd L ═ 7). In this case, a 162-point IDFT requires every other pass through the subsampled DFT-channelizerIs executed once per second (because) The complexity of the 1134-point IDFT is about seven times that of the 162-point IDFT.

Referring back to FIG. 4, the discrete wideband signal X [ n ]]Sampled and filtered by a polyphase filter bank 100 to produce a sequence S_i[n]。S_i[n]Each branch of the signal is time-aliased by L at 105 to produce a new sequence z_i[n]。 The point IDFT110 is taken from the sequence z_i[n]Obtaining the sequence r_i[n]The sequence and the carrier signal sequence e^jWrNnMixing at mixer 115; whereinThe selected channel is derived from the wideband signal.

In a subsampled DFT channelizerThe point IDFT may be calculated using any known fast algorithm for calculating the DFT/IDFT. These algorithms include radix-2 FFT algorithm, Cooley-TukeyFFT algorithm, Wionogard prime number lengthFFT algorithm, and prime factor FFT algorithm, depending onA particular algorithm for calculating the IDFT may be more efficient. Thus, the free parameters (e.g., Fs and M) of the subsampled DFT channelizer may be chosen such that the resulting IDFT may be more efficiently calculated using some specific FFT/IFFT algorithm, in other words, the parameters may be chosen to obtain a size of IDFT that may be efficiently calculated.

For example, ifIs a high non-prime number, the Cooley-Tukey FFt algorithm can be used to efficiently compute the resulting IDFT. Otherwise, ifIs a prime number, the Winograd prime length FFT algorithm can be used to efficiently calculate the resulting IDFT. Finally, ifIs a power of 4, radix-4 FFT algorithm can be used to efficiently calculate the resulting IDFT.

Although a preferred embodiment of the method and apparatus of the present invention has been illustrated in the accompanying drawings and described in the foregoing detailed description, it will be understood that the invention is not limited to the embodiment disclosed, but is capable of numerous rearrangements, modifications and substitutions without departing from the spirit of the invention as set forth and defined by the following claims.

Claims (14)

1. A receiver, comprising:

means for converting the received signal to an analog baseband signal;

an analog-to-digital converter for converting the analog baseband signal into a digital baseband signal; and

a sub-sampled DFT channelizer extracts a plurality of selected, uniformly spaced channels from a digital baseband signal.

2. The receiver of claim 1, wherein the subsampled DFT-channelizer comprises:

a plurality of polyphase filters for decimating the first sequence of signals;

means for time-aliasing the first sequence signal to produce a second sequence signal;

for calculating IDFT coefficients for a plurality of selected, uniformly spaced channels from the second sequence signalPoint inverse discrete Fourier transform; and

means for combining the IDFT coefficients with a carrier signal sequence to provide a plurality of selected, evenly spaced channels.

3. The receiver of claim 2, wherein the first sequence of signals includes each channel within the digital baseband signal.

4. A receiver as claimed in claim 2, wherein the second sequence of signals comprises only a number of signals equal to the selected, evenly spaced number of channels.

5. The receiver of claim 1, wherein the subsampled DFT-channelizer extracts any first channel from the digital baseband signal.

6. A channelizer for processing a wideband signal, comprising:

means for receiving a broadband signal;

a subsampled DFT-channelizer that extracts a number of selected, uniformly spaced channels from a plurality of channels within the received wideband signal; and

means for outputting the selected, evenly spaced channels.

7. The channelizer according to claim 6, wherein the sub-sampling DFT-channelizer comprises:

a plurality of polyphase filters for decimating the first sequence of signals;

means for time-aliasing the first sequence signal to produce a second sequence signal;

for calculating IDFT coefficients for a plurality of uniformly spaced channels from the second sequence signalPoint discrete Fourier transform; and

means for combining the IDFT coefficients with a carrier signal sequence to provide a plurality of selected, evenly spaced channels.

8. A channelizer according to claim 7 wherein the first sequence signal includes each channel within the wideband signal.

9. A channelizer according to claim 7 wherein the second sequence signal includes only a number of signals equal to a plurality of selected, uniformly spaced channels.

10. The receiver of claim 1, wherein the subsampled DFT-channelizer extracts any first channel from the wideband signal.

11. A method for processing a wideband signal containing a plurality of channels, comprising the steps of:

receiving a wideband signal comprising a plurality of channels;

extracting a selected number of evenly spaced channels from a plurality of channels within the wideband signal; and

outputting the extracted evenly spaced channels.

12. The method of claim 11, wherein the step of extracting further comprises the steps of:

a plurality of channel subsamples DFT are channelized within the wideband signal to extract a selected number of evenly spaced channels.

13. The method of claim 11, wherein the step of extracting further comprises the steps of:

filtering the wideband signal to extract each of a plurality of channels within the wideband signal;

time-aliasing a plurality of channels to obtain a sequence signal equal in number to the number of selected uniformly spaced channels;

according toPoint inverse discrete fourier transform processing the second sequence of signals to obtain IDFT coefficients for a selected number of evenly spaced channels; and

the IDFT coefficients are mixed with the carrier signal sequence to obtain a selected number of evenly spaced channels.

14. The method of claim 11, wherein the step of extracting further comprises the steps of:

filtering the wideband signal to extract each of a plurality of channels;

determining IDFT coefficients for a selected number of evenly spaced channels from the extracted plurality of channels; and

the IDFT coefficients are mixed with the carrier signal sequence to obtain a selected number of evenly spaced channels.