US RE41952 E1 Abstract A modulator, and demodulator, apparatus and method for use in a multiple sub-channel communication system is taught. A commutator is employed for fractionally sampling, or distributing, signals from, or to, a multiple channel polyphase filter. The filter is coupled with a discrete Fourier transform, or its inverse, such that the relationship between the base-band sampling rate of a plurality of sub-channel signals, the frequency spacing of the sub-channel signals, and the sampling rate of a composite signal can be related by any rational number, thereby freeing designers to optimize system design respecting channel spacing, bandwidth, and signaling rates. The advantages of the present invention are realized by adjusting the interpolation and decimation rates of the filter, and by adjusting the resolution and decimation rates of the transform.
Claims(25) 1. A modulator for receiving a plurality of sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and, for generating a composite signal, combining the plurality of sub-channel signals, that is sampled at a composite sampling frequency, comprising:
an inverse discrete Fourier transform coupled to receive and transform the plurality of sub-channel signals into a plurality of time domain signals;
a multiple channel polyphase filter coupled to receive said plurality of time domain signals and output a plurality of filter signals, and
a commutator operable to fractionally sample said plurality of filter signals at a rate defined by the ratio of said frequency spacing and a greatest common divisor between said composite sampling rate frequency and said frequency spacing.
2. The modulator in
3. The modulator in
4. The modulator in
5. The modulator in
6. A modulator for receiving a plurality of sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and, for generating a composite signal, combining the plurality of sub-channel signals, that is sampled at a composite sampling frequency, comprising:
an inverse discrete Fourier transform coupled to receive and transform the plurality of sub-channel signals into a plurality of time domain signals, said inverse discrete Fourier transform resolution transform having a resolution defined by the ratio of said composite sampling frequency and the greatest common divisor between said composite sampling frequency and said frequency spacing, and, said inverse discrete Fourier transform having a transform decimation rate defined by the ratio of said frequency spacing and a greatest common denominator between said composite sampling frequency and said frequency spacing;
a multiple channel polyphase filter coupled to receive said plurality of time domain signals and output a plurality of filter signals, said filter having an interpolation rate defined by a least common multiple between said base-band sampling frequency and said composite sample frequency divided by said base-band sampling frequency, said multiple channel polyphase filter having a filter decimation rate defined as said filter interpolation rate times said base-band sample frequency and divided by said composite sampling frequency, and
a commutator operable to fractionally sample said plurality of filter signals at a rate defined by the ratio of said frequency spacing and a greatest common divisor between said composite sampling rate frequency and said frequency spacing.
7. The modulator in
8. The modulator in
9. The modulator in
10. The modulator in
11. A method of modulating a plurality of sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, onto a composite signal that is sampled at a composite sampling frequency, comprisingthe steps of :
converting the plurality of sub-channel signals into a plurality of time domain signals by performing an inverse Fourier transform that has a resolution defined by the ratio of the composite sampling frequency and a greatest common divisor between the composite sampling frequency and the frequency spacing, and, a decimation rate defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling frequency and the frequency spacing;
filtering said plurality of time domain signals using a multiple channel polyphase filter to produce a plurality of filter signals, the multiple channel polyphase filter having an interpolation rate defined by a least common multiple between the base-band sampling frequency and the composite sample frequency divided by the base-band sampling frequency, the filter decimation rate defined as the filter interpolation rate times the base-band sample frequency and divided by the composite sampling frequency, and
fractionally sampling said plurality of filter signals at a rate defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling rate frequency and the frequency spacing.
12. A system comprising:
a modulator to receive a first plurality of sub-channel signals that are sampled at a first base-band sampling frequency and separated by a first frequency spacing, and to generate a first composite signal, combining the first plurality of sub-channel signals, that is sampled at a first composite sampling frequency, comprising: an inverse discrete Fourier transform coupled to receive and transform the first plurality of sub-channel signals into a plurality of time domain signals; a first multiple channel polyphase filter coupled to receive said plurality of time domain signals and output a plurality of first filter signals, and a first commutator operable to fractionally sample said plurality of first filter signals at a rate defined by the ratio of said first frequency spacing and a greatest common divisor between said first composite sampling frequency and said first frequency spacing; and a demodulator to receive a second composite signal that is a combination of a second plurality of sub-channel signals and that is sampled at a second composite sampling frequency, and to discriminate the second plurality of sub-channel signals each at a second base-band sampling frequency and separated by a second frequency spacing, comprising: a second commutator operable to fractionally distribute said second composite signal to a plurality of filter input signals at a rate defined by the ratio of said second frequency spacing and a greatest common denominator between said second composite sampling frequency and said second frequency spacing; a second multiple channel polyphase filter having a plurality of filter inputs coupled to receive said plurality of filter input signals, said filter to output a plurality of second filter signals; and a discrete Fourier transform coupled to receive and transform said plurality of second filter signals and to output the second plurality of sub-channel signals. 13. The system of
14. The system of
15. The system of
16. A method of modulating a plurality of sub-
channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and generating a composite signal, combining the plurality of sub-channel signals, that is sampled at a composite sampling frequency, the method comprising: converting the plurality of sub-channel signals into a plurality of time-domain signals using an inverse discrete Fourier transform; filtering the plurality of time-domain signals using a multiple channel polyphase filter, to thereby output a plurality of filter signals; and fractionally sampling said plurality of filter signals at a rate defined by the ratio of said frequency spacing and a greatest common divisor between said composite sampling frequency and said frequency spacing. 17. The method of
18. The method of
mixing the composite signal with a wireless carrier for wireless transmission. 19. A non-
transitory processor-readable medium containing instructions that, when executed by a processor, cause the processor to implement a method of modulating a plurality of sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and generating a composite signal, combining the plurality of sub-channel signals, that is sampled at a composite sampling frequency, the method comprising: converting the plurality of sub-channel signals into a plurality of time-domain signals using an inverse discrete Fourier transform; filtering the plurality of time-domain signals using a multiple channel polyphase filter, to thereby output a plurality of filter signals; and 20. The medium of
21. The medium of
mixing the composite signal with a wireless carrier for wireless transmission. 22. A method of modulating a first plurality of sub-
channel signals that are sampled at a first base-band sampling frequency and separated by a first frequency spacing, and to generate a first composite signal, combining the first plurality of sub-channel signals, that is sampled at a first composite sampling frequency and of demodulating a second composite signal that is a combination of a second plurality of sub-channel signals and that is sampled at a second composite sampling frequency, and to discriminate the second plurality of sub-channel signals each at a second base-band sampling frequency and separated by a second frequency spacing, the method comprising: transforming the first plurality of sub-channel signals by using an inverse discrete Fourier transform, to obtain a plurality of time domain signals; filtering the plurality of time domain signals using a first multiple channel polyphase filter, to output a plurality of first filter signals; fractionally sampling said plurality of first filter signals at a rate defined by the ratio of said first frequency spacing and a greatest common divisor between said first composite sampling frequency and said first frequency spacing; fractionally distributing said second composite signal to a plurality of filter input signals at a rate defined by the ratio of said second frequency spacing and a greatest common denominator between said second composite sampling frequency and said second frequency spacing; filtering the plurality of filter input signals using a second multiple channel polyphase filter to output a plurality of second filter signals; and transforming said plurality of second filter signals using a discrete Fourier transform, to output the second plurality of sub-channel signals. 23. The method of
receiving a wireless carrier signal; and discriminating the second composite signal therefrom for wireless reception. 24. A non-
transitory processor-readable medium containing instructions that, when executed by a processor, cause the processor to implement a method of modulating a first plurality of sub-channel signals that are sampled at a first base-band sampling frequency and separated by a first frequency spacing, and to generate a first composite signal, combining the first plurality of sub-channel signals, that is sampled at a first composite sampling frequency and of demodulating a second composite signal that is a combination of a second plurality of sub-channel signals and that is sampled at a second composite sampling frequency, and to discriminate the second plurality of sub-channel signals each at a second base-band sampling frequency and separated by a second frequency spacing, the method comprising: transforming the first plurality of sub-channel signals by using an inverse discrete Fourier transform, to obtain a plurality of time domain signals; transforming said plurality of second filter signals using a discrete Fourier transform, to output the second plurality of sub-channel signals. 25. The medium of
receiving a wireless carrier signal; and discriminating the second composite signal therefrom for wireless reception. Description 1. Field of the Invention This invention relates in general to digital communication signal processing techniques, and more specifically to modulation and demodulation techniques in multiple sub-channel communications systems. 2. Description of the Related Art Multiple sub-channel modulation and demodulation engines, and corresponding transmitters and receivers, which are implemented with digital signaling techniques are known in the art. One application of such a device is to communicate several sub-channels of information within a single, broader bandwidth, channel. Such systems employ orthogonal frequency division multiplexed (“OFDM”) modulation, multi-carrier transmission (“MCT”) modulation, and others modulation techniques. The general modulation approach in such systems is to consolidate the multiple sub-channels as sub-carriers in a composite signal in the base-band, and then frequency-shift the consolidated signal to the allocated carrier signal band, usually centered about a carrier frequency. Although, it is also understood that such systems can operate in the base-band without frequency shifting, and that such systems can be utilized in other electromagnetic bands, such as infrared and visible light. Multiple sub-channel modulation and demodulation devices employ various topologies and architectures. Among these are a particular variety that employ a combination of inverse Fourier transforms (or Fourier transforms) with polyphase filters and a software commutator. This is typically accomplished within a digital signal processor, however software implementations in other kinds of computing devices are certainly available. The basic design parameters of such devices, that are used for the purpose of modulation, are characterized by a number of input sub-channel signals, or sub-carrier signals, that are sampled at an input, or base-band, sampling rate, and, that are spaced apart in frequency by an input frequency spacing that is ultimately translated through to a composite signal. In addition, such devices are characterized by the composite output signal that is sampled according to an output, or composite, sampling rate. Respecting demodulation, the same characteristics apply, but are related to the input sampling rate of a composite signal, and the output sampling rate and frequency spacing of plural output sub-channel signals. These characterizations of are of vital importance when considering an overall system design and architecture. The design of a polyphase filter for band limiting each sub-channel according to its pass-band and frequency spacing is critical to the design of an efficient data communications system. So too is the communications baud rate, which is directly affected by the filter design and sub-channel spacing, and which is important to the spectral response of the composite signal transmitted in the allocated channel. Previously, those skilled in the art have understood that there were various constraints placed on the relationship between the input sampling rate, the sub-channel frequency spacing, and the output sampling rate. These constraints have been considered problematic respecting optimum system design, especially where maximum performance in marginal signal conditions were sought. More particularly, polyphase filters operate by multiplying selected phases, or samples, of a filter impulse response with samples of one or more the aforementioned input signals. Prior-art multiple channel polyphase filters have synchronized the selected phases of the filter impulse response with the positions of a commutator of the filter. In prior-art multiple channel polyphase filters, a given position of the commutator has corresponded uniquely to a predetermined phase of the filter impulse response. Those of ordinary skill in the art of polyphase filter design have even accepted that in multiple sub-channel polyphase modulators, the sub-channel sampling rate must be an integral multiple of the input sample rate, which also implies that the channel bandwidth, or frequency spacing, must be an integer multiple of the input sample rate. The tight limitations on the relationship between input sampling rate, frequency spacing, and output sampling rate were partially alleviated by the teachings of the present inventor, McCoy, in a prior U.S. Pat. No. 6,134,268 to McCoy for APPARATUS FOR PERFORMING A NON-INTEGER SAMPLING RATE CHANGE IN A MULTIPLE CHANNEL POLYPHASE FILTER (the '268 patent), the contents of which are hereby incorporated by reference thereto. The '268 patent teaches a multiple channel polyphase filter that includes a processing system for accepting and processing ‘M’ input channels of data, each sampled at an input sampling rate, wherein ‘M’ is a positive integer. The processing system is programmed to provide a commutator for the multiple channel polyphase filter, wherein the position of the commutator is decoupled from the phase of a filter impulse response selected for the position, thereby allowing the multiple channel polyphase filter to be operated at a sampling rate that is a non-integer multiple of the input sampling rate. The processing system is further programmed to operate the multiple channel polyphase filter at the non-integer multiple of the input sampling rate to obtain a non-integer sampling rate change. Other embodiments and applications taught by the '268 patent include a multiple channel polyphase filter, a multiple channel modulation engine, a corresponding multiple channel demodulation engine, a multiple channel transmitter, and a corresponding multiple channel receiver. While all of these embodiments teach the decoupling of the input sampling rate from the output sampling rate, each still suffers from a limitation defining the sub-channel frequency spacing by a fixed relationship between the input sampling rate and output sampling rate. Thus, even in view of the teachings of McCoy in the '268 patent, communication system designers are faced with a constraints on sub-channel spacing that is often times unable to deliver optimum performance in practical applications. Thus, there is a need in the art for a modulation and demodulation engine applicable to multiple sub-channel systems that allows for the arbitrary specification of sub-channel frequency spacing with respect to input and output sample rates. The need in the art is addressed by the apparatus and methods of the present invention. The present invention encompasses multiple channel modulation bank that includes a programmable processing system for accepting plural channels of input data at an arbitrary sampling rate, or baud, rate, forming a composite output signal at an arbitrary sampling rate. Also, a multiple channel demodulation bank that includes a programmable processing system for accepting a composite input signal at an arbitrary sampling rate and yielding plural channels of output data at an arbitrary sampling rate. The sub-channels of data are spaced at an arbitrary spacing. While the prior art has restricted the relationship between the input sampling rate, output sampling rate, and the frequency spacing in some way, the present invention allows any rational relationship between all of these. In particular, a modulator for receiving plural sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and, for generating a composite signal, combining the plurality of sub-channel signals, that is sampled at a composite sampling frequency is taught. The modulator includes an inverse discrete Fourier transform coupled to receive the plural sub-channel signals and transform them into plural time domain signals. It also includes a multiple channel polyphase filter that receives the plural time domain signals and outputs a plurality of filter signals to a commutator. The commutator fractionally samples the filter signals at a rate defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling rate and the frequency spacing. In a refinement of the foregoing modulator, the commutator fractional sampling ratio is derived as a ratio of integers thus allowing the relationship to be expressed as any rational number. In a further refinement, the modulator further adds a wireless modulator that mixes the composite signal with a wireless carrier for use in wireless transmission. In a further refinement, the transform, the filter, and the commutator are implemented with executable software on a processor. In a further refinement, the processor is a digital signal processor. The present invention also teaches a modulator for receiving plural sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, and, for generating a composite signal, combining the plural sub-channel signals, that is sampled at a composite sampling frequency. The modulator includes an inverse discrete Fourier transform that transforms the plural sub-channel signals into plural time domain signals. The transform resolution is defined by the ratio of the composite sampling frequency and the greatest common divisor between the composite sampling frequency and the frequency spacing. Also, the transform decimation rate is defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling frequency and the frequency spacing. The modulator also includes a multiple channel polyphase filter that receives the plural time domain signals and outputs plural filter signals. The filter has an interpolation rate defined by a least common multiple between the base-band sampling frequency and the composite sample frequency divided by the base-band sampling frequency. Also, the filter decimation rate is defined as the filter interpolation rate times the base-band sample frequency and divided by the composite sampling frequency. Finally, the modulator includes a commutator that fractionally samples the plural filter signals at a rate defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling rate and the frequency spacing. In a refinement of the foregoing modulator, the commutator fractional sampling ratio is defined by the decimation rate. In a further refinement, the modulator further adds a wireless modulator that mixes the composite signal with a wireless carrier for use in wireless transmission. In a further refinement, the transform, the filter, and the commutator are implemented with executable software on a processor. In a further refinement, the processor is a digital signal processor. The present invention also teaches a demodulator that receives a composite signal that is a combination of plural sub-channel signals and that is sampled at a composite sampling frequency, and, that discriminates the plural sub-channel signals each at a base-band sampling frequency and separated by a frequency spacing. The demodulator includes a commutator that fractionally distributes the composite signal into plural filter input signals at a rate defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling rate and the frequency spacing. The demodulator also includes a multiple channel polyphase filter that has plural filter inputs to receive the plural filter input signals. The filter outputs plural filter signals to a discrete Fourier transform. The transformtransforms the plural filter signals, and outputs the plural sub-channel signals. In a refinement to the foregoing demodulator, the commutator fractional sampling ratio is the decimation rate. In a further refinement, a wireless demodulator is added that is used to receive a wireless carrier signal, and to discriminate the composite signal therefrom to enable wireless reception. In a further refinement, the commutator, the filter, and the transform are implemented with executable software on a processor. In a further refinement, the processor is a digital signal processor. The present invention also teaches another demodulator that receives a composite signal that is a combination of plural sub-channel signals, and that is sampled at a composite sampling frequency, and, that discriminates the plural sub-channel signals, each sampled at a base-band sampling frequency and separated by a frequency spacing. The demodulator includes a commutator that fractionally distributes the composite signal to plural filter input signals at a rate defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling rate and the frequency spacing. The demodulator also includes a multiple channel polyphase filter that receives the plural filter input signals and outputs plural filter signals. The filter has a decimation rate defined by a least common multiple between the base-band sampling frequency and the composite sample frequency divided by the base-band sampling frequency, and the filter has a interpolation rate defined as the filter interpolation rate times the base-band sample frequency and divided by the composite sampling frequency. The demodulator also includes a discrete Fourier transform that receives and transform the plural filter signals, and outputs the plural of sub-channel signals. The transform resolution is defined by the ratio of the composite sampling frequency and the greatest common denominator between the composite sampling frequency and the frequency spacing. Also, the transform decimation rate is defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling frequency and the frequency spacing. In a refinement to the foregoing demodulator, the commutator fractional sampling ratio is a ratio of integers that define any arbitrary rational number. In a further refinement, a wireless demodulator is added that is used to receive a wireless carrier signal, and to discriminate the composite signal therefrom to enable wireless reception. In a further refinement, the commutator, the filter, and the transform are implemented with executable software on a processor. In a further refinement, the processor is a digital signal processor. The present invention also teaches a method of modulating plural sub-channel signals that are sampled at a base-band sampling frequency and separated by a frequency spacing, onto a composite signal that is sampled at a composite sampling frequency. This method includes the steps of converting the plural sub-channel signals into plural time domain signals by performing an inverse Fourier transform that has a resolution defined by the ratio of the composite sampling frequency and the greatest common divisor between the composite sampling frequency and the frequency spacing. Also, the transform has a decimation rate defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling frequency and the frequency spacing. The next step is filtering the plural time domain signals using a multiple channel polyphase filter to produce plural filter signals. The filter has an interpolation rate defined by a least common multiple between the base-band sampling frequency and the composite sample frequency divided by the base-band sampling frequency. Also, the filter decimation rate defined as the filter interpolation rate times the base-band sample frequency and divided by the composite sampling frequency. The last step is fractionally sampling the plural filter signals at a rate defined by the ratio of the frequency spacing and a greatest common divisor between the composite sampling rate and the frequency spacing. There is also taught a corresponding method of demodulating a composite signal that is sampled at a composite sampling frequency into a plurality of sub-channel signals each sampled at a base-band sampling frequency and separated by a frequency spacing. This method includes the steps of fractionally distributing the composite signal to plural filter input signals at a rate defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling rate and the frequency spacing. Next, filtering the plurality of filter input signals using a multiple channel polyphase filter to produce plural filter signals. The filter has an decimation rate defined by a least common multiple between the base-band sampling frequency and the composite sample frequency divided by the base-band sampling frequency, and the filter has interpolation rate defined as the filter interpolation rate times the base-band sample frequency and divided by the composite sampling frequency. The next step is transforming the plural filter signals, using a discrete Fourier transform, into the plural sub-channel signals. The transform resolution is defined by the ratio of the composite sampling frequency and the greatest common denominator between the composite sampling frequency and the frequency spacing. Also, the transform decimation rate is defined by the ratio of the frequency spacing and a greatest common denominator between the composite sampling frequency and the frequency spacing. Illustrative embodiments and exemplary applications will now be described with reference to the accompanying drawings to disclose the advantageous teachings of the present invention. While the present invention is described herein with reference to illustrative embodiments for particular applications, it should be understood that the invention is not limited thereto. Those having ordinary skill in the art and access to the teachings provided herein will recognize additional modifications, applications, and embodiments within the scope thereof and additional fields in which the present invention would be of significant utility. An illustrative embodiment communications systems according to the present invention is illustrated in More particularly, the illustrative embodiment in In an illustrative embodiment, the channel allocations are 25 kHz wide, and the aforementioned FDM channelization sub-divides each 25 kHz channel into four sub-channels, which are spaced 4.8 kHz apart. The TDMA channelization further divides each sub-band into two communications time slots with multiplexed data packets that are arranged in time to accomplish simplex, half-duplex, and duplex communications in real time. The illustrative embodiment system operates through the transmission and reception of four bit symbols at a 4 kHz symbol rate. The symbols are modulated and demodulated to the RF band according to a 16-QAM modulation scheme with a 16-point constellation in phase and amplitude, as is understood by those of ordinary skill in the art. The present invention teaches novel approaches to the base-band modulation and demodulation of such multiple sub-channelized systems. An exemplary receiver In several of the following illustrative examples, a multiple sub-carrier modulation system, developed by Com-Space Corporation and known in the marketplace by the “DCMA” trademark, is used to exemplify the prior art and the advantageous teaching of the present invention. The DCMA systems modulation approach transmits and receives a single or multiple sub-carriers within a FCC allocated radio channels that are spaced at 25 kHz centers. Such channel allocations exist in the United States in the 800 MHz and 900 MHz bands. However, the present invention is in no way limited by these examples, and, those of ordinary skill in the art will appreciate that the teachings herein are applicable to a vast array of communications systems and encompass a vast range of communications bandwidths. The exemplary DCMA systems modulates each sub-channel with a frequency offset that takes on one of four possible values with respect to the FCC allocated channel center frequency, and these values are; −7200 Hz, −2400 Hz, +2400 Hz, and +7200 Hz. Modulations is by 16-QAM comprised of 4-bit symbols transmitted at a symbol rate of 4000 baud. According to the aforementioned frequency offsets, the sub-carrier channel spacing is 4800 Hz. The output sampling rate of the combined sub-carriers ultimately transmitted over the allocated channel is 260 kilo samples per second (“ksps”). Reference is directed to Mathematically, these operations can be expressed as follows.
Where D is the polyphase filter decimation rate, I is the polyphase filter interpolation rate, m is the input signal time at the input signal sampling rate, and n is the output time at the upsampled output rate. And, s Given that the upsample rate, establishes a modulo ratio, the expression becomes:
And since any realizable digital filter must be truncated in time, the summation is limited to the number of taps per phase of the filter, the expression becomes:
Adding the frequency shift and summing the plural singles, the output y(n) One of the advantages of the present invention is the reduction in processor operation required to achieve the desired modulation and demodulation of the plural sub-channels. Respecting the prior art modulation engine of Where p is the equivalent number of multiplies required in order to generate the complex sinusoid, and f Which is algebraically manipulated to yield Equation 8.
Use the substitution, n=r+sM, where r is an indicator of the commutator cycle position in time, and s is an indicator of overall time, to yield a modified summation.
Equation 9 will be readily understood by those possessing ordinary skill in the art. An important advancement in the art made by the present invention occurs when the constraint that f Which is algebraically manipulated in the same fashion as between Equations (1) and (2) to yield:
The commutator position s Now, considering the forgoing advancement in view of the illustrative embodiment DCMA system, which operation parameters were enumerated herein before, it will be appreciated that the foregoing modulation model can be applied to the modulation bank for the downlink. The downlink is the communications path from the repeater systems to the terminal units. As the repeater systems must communicate simultaneously with a plurality of terminal units, it must do so on each of the sub-channels simultaneously as well. However, since the output sampling rate in DCMA is fixed at 260 ksps, it is necessary to contemplate the more complex sampling rate change this implies (4 ksps to 260 ksps). Consider, generally, a process where the DFT is greatly over sampled in order to accommodate a difficult rate change. Such an operation can be mathematically constructed as follows.
The first summation generally defining the filtering operation and the second generally defining the transformation. Again applying an advancement of the present invention, let f Since f Again, applying the same practical application of a finite time filter, Equation (14) becomes:
Where N Respecting the question of processor overhead burden of the DCMA illustrative embodiment of the present invention, in the worst case, the computational complexity of using this filter bank is as follows.
Where CDFT is the number of processor multiplies required in order to compute the IDFT. Again, for the DCMA illustrative embodiment, the following is an approximation of the complexity numbers.
The question of how many processor multiplies will it take to implement the corresponding DFT. In this case there are far fewer inputs than total output points to be computed. Assuming the input is complex, the number of computations becomes:
It is important to note in this case that the number of operations required for the mixing operation can be increased above the required amount for the prior art implementation. This situation can arise if the number of sub-carriers is much smaller than the DFT size. However, in such situations, the utilized bandwidth in ratio to the output sampling rate will be small. If that is the case, the modulation bank can be implemented at a lower output sampling rate followed by an upsampler. The computational complexity of such an implementation in this case is as follows.
Where N Respecting an implementation of the present invention modulation engine illustrated in The taps of the filter phase, are generated as follows.
The data pointer position is:
The commutator position is given by:
The foregoing equation set does not specify the implementation of the IDFT. The IDFT has four inputs and 65 outputs. There are four active sub-carriers. The sub-carriers should are located at −7200 Hz, −2400 Hz, +2400 Hz and +7200 Hz. In order to accommodate these frequencies, the basic IDFT is modified as follows.
This additional shift is not expensive in terms of processor overhead because a direct implementation of the IDFT is used. The composite operation of the filter bank is summarized by the following steps: - STEP 1: Compute the IDFT of one time-slice of symbols according to Equation (24). This will account for 65 separate filter histories, each at the symbol rate, (The original symbols will never be needed again.).
- STEP 2: Identify the filter phase according to Equation (20).
- STEP 3: Identify the data pointer position according to Equation (22).
- STEP 4; Identify the commutator position according to Equation (23).
- STEP 5: Compute the dot product of the filter phase and the data vector according to the following equation:
$\begin{array}{cc}y\left(n\right)=\sum _{k=0}^{{N}_{\mathrm{tapsNyq}}-1}\text{\hspace{1em}}h\left(\varphi \left(n\right)+65k\right)\xb7{s}_{\chi \left(n\right)}\left(\delta \left(n\right)-k\right)& \left(25\right)\end{array}$
This operation is followed by a direct interpolation by 5 to achieve the design output sampling rate of 260 ksps, as discussed above. The interpolation by 5 operation is specified as follows.
The taps of the filter phase are generated as follows.
The data pointer position is:
The output is computed as follows:
The teachings of the present invention are also applicable to the demodulation of multiple sub-carrier signals in a similar fashion to the previously discussed modulation aspects. For a more detailed understanding of the demodulation aspects of the teachings of the present invention, it is useful to consider the prior art, and especially the prior art DCMA systems. By way of review, DCMA modulation consists of a single (in the case of a terminal unit operating on a single sub-channel) or multiple sub-carriers with a frequency offset that takes on four possible values, −7200, −2400, +2400, or +7200 Hz from the allocated carrier frequency. The basic parameters of the system are a symbol rate is 4000 baud; the sub-carrier spacing is 4800 Hz, and an output sampling rate is 260 ksps. The prior art demodulation approach is diagrammed in Mathematically, the processes illustrated in Where x Which is algebraically equivalent to:
In the flowing analysis, the inner summation, indicating the filter response, is separated out, and, the modulo count remainder of the decimation/interpolation process incorporated for thoroughness.
Using the substitution,
The operation of commutation is expressed by, y Equation 36 is the form of an ordinary polyphase filtering operation, with the slight modification of the rI term. This can be readily implemented in a digital signal processor, as will be appreciated by those of ordinary skill in the art. Finally, substitution back in to Equation (33) yields the following complete expression of the output channels.
The discrete Fourier transform (“DFT”) is not necessarily an ordinary DFT, although it can be in certain cases, depending on the relationship between Δf and T Respecting the illustrative embodiment DCMA communications system, the input sampling rate is 260 ksps with 4 sub-carriers spaced at 4800 Hz having a baud rate of 4 ksps. The calculations yield values as follows; M=325, f Respecting the processor demand for this illustrative embodiment application of the present invention, the number of processor multiplications required in order to implement such an operation is approximated in the worst case as follows.
The number of computations required for a 10 sub-carrier system assuming a 11 symbol delay is 11 MIPS. The same system with a conventional implementation would require 70 MIPS. On the receive side at least 3 taps of the filter must be sampled. A fine timing offset (early or late) may be realized by changing the phase of the filter. These additional outputs scale the complexity requirements. For 3 output taps, the complexity requirements triple, which is true of both the prior art implementation and the illustrative embodiment implementation. Thus, the present invention has been described herein with reference to a particular embodiment for a particular application. It is therefore intended by the appended claims to cover any and all such applications, modifications and embodiments within the scope of the present invention. Accordingly, Patent Citations
Classifications
Legal Events
Rotate |