KR101082464B1 - Method and Apparatus for the Digital Transmitter with Joint Pulse Shaping Filter and Modulator - Google Patents

Abstract

The method and apparatus for implementing a digital transmitter according to at least one embodiment of the present invention are performed by combining a pulse shaping filtering function and a modulation function that perform the most operations in a transmitter, thereby generating an oversample signal by the pulse shaping filtering function. Replace the existing two separate processes of modulating this with 'a process for generating an oversampled signal by the pulse shaping filtering function', and then generate an oversample signal with the 'pulse shaping filtering function' It is possible to obtain the same output result as the final output result of the existing two separate processes of modulating this, thereby dramatically reducing the amount of computation required for transmission, thereby realizing efficient transmission signal generation and consuming at the transmitter. Low power transmitters can be implemented with reduced power consumption.

Description

Method for implementing digital transmitter with pulse shaping filter and modulator {Method and Apparatus for the Digital Transmitter with Joint Pulse Shaping Filter and Modulator}

The present invention relates to a transmitter, and more particularly, to a digital transmitter for a wireless digital communication system for exchanging digital information by transmitting and receiving digital data through a wireless channel.

The digital transmitter maps to a symbol according to a modulation method using digital data transmitted to exchange digital information through a wireless channel, and inter-symbol interference (ISI) while limiting the bandwidth of the symbol signal. To prevent this from happening, the pulsed filter is passed, and the pulsed symbol signal is shifted from the baseband to the frequency band to be transmitted. The receiver receives a signal transmitted from the transmitter and performs demodulation to convert the passband signal into a baseband signal, and restores digital information transmitted using the baseband signal.

On the other hand, thanks to the rapid development of the computational performance of the signal processing processor, the signal processing processor has recently implemented a communication system using dedicated hardware such as a field programmable gate array (FPGA) and an application specific integrated circuit (ASIC). Software can be programmed and implemented.

In order to efficiently implement a digital communication system using a signal processing processor, it is necessary to minimize the amount of computation required for communication. The same applies to the implementation of a digital transmitter using a signal processing processor. Digital transmitters are largely divided into a digital signal processor (not shown) and an analog signal processor (not shown). The digital signal processor performs a function of mapping digital information data to a symbol signal according to a modulation method, and a pulse shaping filtering function to convert the symbol signal into a waveform signal. When digital modulation is performed according to the system configuration, the transmitter converts the baseband waveform signal of the pulse shaping filter into a passband signal of the carrier band by a digital signal processor and then performs an analog signal through a digital-to-analog converter (DAC). The digital signal processor outputs the pulse shaping filter by the digital signal processor when the analog signal processing unit converts the output signal of the DAC and transmits the analog signal through the antenna or the sensor. Is converted to a baseband analog signal through a DAC, and the analog signal processor converts the baseband signal, which is a DAC output signal, into a passband signal, and then amplifies the modulated signal and transmits the analog signal through an antenna or a sensor.

Such a series of processes that a conventional digital transmitter must perform requires a considerable amount of computation (particularly, a considerable amount of computation is required when performing pulse shaping filtering and modulation), and this considerable computation amount is an obstacle to efficient transmission signal generation. Increasing power consumption also makes it difficult to implement low-power transmitters.

An object of the present invention is to provide a digital transmitter having a high computational efficiency as a digital transmitter having a structure that performs pulse shaping filtering and digital modulation in the digital signal processor.

Another technical problem to be achieved by at least one embodiment of the present invention is a transmission method performed in a digital transmitter having a structure that performs pulse shaping filtering and digital modulation in a digital signal processing unit. To provide.

According to at least one embodiment of the present invention, a digital transmitter includes: an oversample filtering unit for oversampling P (where P is a natural number) for each symbol with respect to an input symbol signal; And interpolating and modulating the oversampled signal in parallel for each symbol, wherein Q is a natural number, and modulating a frequency band of the interpolation result.

The interpolation and modulation unit may include: an interpolation unit interpolating Q times the oversampled signal in parallel for each symbol; And a modulator for modulating the frequency band of the interpolated result. In this case, each of the interpolators operates simultaneously, and generates a corresponding q (where q is an integer of 0≤q≤ (Q-1)) sample signal for one of the oversampled signals. Q uncoupled interpolators; And a parallel-to-serial converter that sequentially aligns the Q sample signals generated from the Q uncoupled interpolators.

The interpolation and modulator may be implemented as an interpolator having a combined interpolator coefficient as an interpolator coefficient, and the interpolator may complete the interpolation and modulation function by performing Q times interpolation of the oversampled signal.

In this case, the interpolator may be a single interpolator having the combined interpolator coefficients for generating different Q sample signals for one sample signal of the oversampled signal as an interpolator coefficient.

In this case, the interpolators operate simultaneously with each other, and generate corresponding q (where q is an integer of 0≤q≤ (Q-1)) sample signal for one sample signal of the oversampled signal. The interpolation and modulator may sequentially output the Q interpolators and the sample signals generated from the combined interpolators to output a result to be obtained by performing interpolation and modulation functions.

In this case, the combined interpolator coefficient may be a first interpolator coefficient and a second interpolator coefficient determined according to the period of the modulation signal. For example, in the real axis, the combined interpolator coefficient of the combined interpolator corresponding to q may be a result of multiplying cos (2qπ / Q) by the first interpolator coefficient of the combined interpolator corresponding to q. Similarly, in the case of the sewer axis, the combined interpolator coefficient of the combined interpolator corresponding to q is the result of multiplying '-sin (2qπ / Q)' by the first interpolator coefficient of the combined interpolator corresponding to q Can be.

In order to achieve the above object, the digital transmission method according to at least one embodiment of the present invention comprises the steps of: (a) over-sampling P (where P is a natural number) for each symbol with respect to the input symbol signal; And (b) interpolating Q (where Q is a natural number) multiple times in parallel with the oversampled signal for each symbol, and modulating a frequency band of the interpolation result.

Wherein step (b) comprises: (b1) interpolating the oversampled signal by Q times in parallel for each symbol; And (b2) modulating the frequency band of the interpolated result. In this case, step (b1) may include simultaneously generating different Q sample signals for one sample signal of the oversampled signal; And sequentially arranging the generated Q sample signals.

Here, the step (b) is performed by an interpolator having a combined interpolator coefficient as the interpolator coefficient, and the step (b) completes the interpolation and modulation functions by interpolating the oversampled signal by Q times. can do.

In this case, the interpolator may be a single interpolator having the combined interpolator coefficients for generating different Q sample signals for one sample signal of the oversampled signal as an interpolator coefficient.

In this case, step (b) may include simultaneously generating different Q sample signals for one sample signal of the oversampled signal; And (b) outputting a result to be obtained by performing interpolation and modulation functions by sequentially outputting the generated Q sample signals.

In this case, the combined interpolator coefficient may be a first interpolator coefficient and a second interpolator coefficient determined according to the period of the modulation signal.

In order to achieve the above object, the computer-readable recording medium according to at least one embodiment of the present invention comprises the steps of filtering the P (where P is a natural number) multiple times per symbol for the input symbol signal; And Q (where Q is a natural number) multiplex interpolation of the oversampled signal in parallel for each symbol, and store a computer program for causing a computer to modulate the frequency band of the interpolation result.

The digital transmitter and method according to at least one embodiment of the present invention can perform an interpolation operation for generating an oversampled signal in parallel to generate an oversampled signal more quickly, thereby implementing a transmitter having a high computational efficiency. have.

In particular, the digital transmission apparatus and method according to at least one embodiment of the present invention, by performing a combination of the pulse shaping filtering function and the modulation function that performs the most operation in the transmitter 'generating the over-sample signal by the pulse shaping filtering function And replace the existing two separate processes of modulating it with 'a process for generating an oversampled signal by the pulse shaping filtering function', and then generate an oversample signal with the 'pulse shaping filtering function' through one process. It can obtain the same output result as the final output result of the existing two separate processes of modulating it, thereby significantly reducing the amount of computation required for transmission to implement efficient transmission signal generation, Low power transmitters can be realized by reducing power consumption. In addition, according to at least one embodiment of the present invention, it is possible to configure a system having a standardized structure to implement the transmitter using a signal processing processor or to implement the transmitter using a device having a dedicated computing function Can be increased.

In order to fully understand the present invention, the operational advantages of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings that illustrate preferred embodiments of the present invention and the accompanying drawings.

According to at least one embodiment of the present invention, a digital transmitter is a digital transmitter using 'pulse shaping filtering' and 'modulation operation requiring modulation' by over-sampling a transmission symbol when generating a transmission signal. In the following specification, such a digital transmitter will be described using a Quadrature Phase Shift Keying (QPSK) modulation method as an example. ) It can be applied to all cases of over-sampling transmission symbols such as modulation schemes and using modulation schemes that require pulse shaping filtering and modulation operations.

Hereinafter, a digital transmitter and method according to at least one embodiment of the present invention will be described with reference to the accompanying drawings.

1 is a block diagram of a digital transmitter according to at least one embodiment of the present invention.

As shown in FIG. 1, the digital transmitter according to at least one embodiment of the present invention may include a symbol mapping unit 110, a filtering and modulating unit 120, and an amplifying and transmitting unit 130.

The symbol mapping unit 110 receives digital information to be transmitted by the digital transmitter and converts the input information into a symbol signal according to a preset modulation scheme.

The filtering and modulator 120 converts a symbol signal having an infinite bandwidth into a symbol signal having a finite bandwidth to prevent intersymbol interference, and oversamples the changed symbol signal to obtain a waveform signal, and obtains the obtained waveform signal. It performs a 'modulation' to generate a 'modulated signal' by shifting the frequency band from the baseband to the passband.

The amplification and transmission unit 130 amplifies the passband signal for transmission to a long distance and transmits the amplified result through a sensor or an antenna.

A main feature of at least one embodiment of the present invention resides in the operation of the filtering and modulator 120 shown in FIG. 1, and a detailed description of the operation of the filtering and modulator 120 according to at least one embodiment of the present invention 8A to 11, which will be described later, a detailed description of the general case of the filtering and modulator 120 will be described first with reference to FIGS. 2 to 7.

FIG. 2 is an example of a detailed block diagram of the filtering and modulating unit 120 in a general case, and as shown in FIG. 2, the pulse shaping filtering unit 210 and the modulating unit 220 are included.

The pulse shaping filtering unit 210 converts the symbol signal into a waveform signal by performing filtering by interpolating N (where N is a natural number) for each symbol in order to limit the bandwidth of the symbol. In this case, a root-raised cosine (RRC) filter may be used as a filtering agent, but the type of filter is not limited.

The modulator 220 modulates the frequency band of the waveform signal passed through the pulse shaping filter 210 from the baseband to the passband.

FIG. 3 is an example of a detailed block diagram of the pulse shaping filtering unit 210 shown in FIG. 2, and may include a zero insertion unit 310 and a filter 320. 4 is a reference diagram for describing an operation of the pulse shaping filtering unit 210 illustrated in FIG. 3. More specifically, FIG. 4 is a reference diagram for describing an operation of FIG. 3 when the aforementioned N is 4. 3 will be described with reference to FIG. 4.

A symbol signal as shown in FIG. 4A is input to the zero insertion unit 310. The symbol signal shown in (a) of FIG. 4 is a continuous symbol string and shows only three symbols for explanation.

As illustrated in (b) of FIG. 4, the zero insertion unit 310 inserts (N-1) zero signals between symbols.

The filter 320 receives the output of the zero insertion unit 310 and performs filtering to generate a waveform signal as shown in FIG. 4C.

On the other hand, to further explain the 'modulation' function of the filtering and modulator 120 mentioned in Figure 1, 'modulation' multiplies the baseband waveform signal by the IF (Intermediate Frequency) or the modulation signal having a carrier frequency By shifting the frequency band of the waveform signal to the pass band, the multiplied result is converted to an analog signal by DAC (Digital-to-Analog Converting). The output signal of the DAC can be expressed as in Equation (1) below.

(1)

(One)

Here, s (t) means the output signal of the DAC, c (k) means the complex symbol signal in the k-th symbol interval, g (t) means the instantaneous response function of the pulse shaping filter, T Denotes a symbol time, f _c denotes an IF frequency or a carrier frequency (hereinafter, for convenience of description, referred to as IF frequency), Re (x) denotes an operation representing a real part of a complex number x, Im (x) means an operation representing an imaginary part of a complex number.

In order to generate such an analog transmission signal efficiently in a digital signal processing unit (not shown) in at least one embodiment of the present invention as well as in the description of FIGS. 5 to 7 below, the IF frequency fc is represented by a symbol rate. The sampling rate (fs) for the DAC conversion is determined to be an integer multiple of the IF frequency (fc). Hereinafter, for convenience of explanation, it is assumed that the IF frequency is P (where P is a natural number) times the symbol rate and the DAC sample rate is Q (where Q is a natural number) times the IF frequency. In this case, the DAC conversion sample rate is a symbol. (P * Q) times the rate. That is, the digital signal processor (not shown) should generate a signal that is oversampled by (P * Q) times and perform signal processing to generate a transmission signal.

FIG. 5 is a block diagram of an example of a digital transmitter in a general case, and as shown in FIG. 5, the digital transmitter includes a zero insertion unit 510 and a (P * Q) oversampled filter 512. ), A multiplier 514, a zero insertion unit 516, a (P * Q) oversampled filter 518, a multiplier 520, and a subtractor 522.

As shown in FIG. 5, the 0 inserting unit 510 and the 0 inserting unit 516 correspond to the 0 inserting unit 310 (see FIG. 3) of the pulse shaping filtering unit 210 shown in FIG. 2. The (P * Q) oversampled filter 512 per symbol, and the (P * Q) oversampled filter 518 per symbol are filters 320 of the pulse shaping filter 210 shown in FIG. 3, the multiplier 514, the multiplier 520, and the subtractor 522 correspond to the modulator 220 illustrated in FIG. 2.

In terms of the real part, the zero insertion unit 510 inserts (P * Q-1) zeros between symbols, and the (P * Q) oversampled per symbol so that N mentioned above is P * Q. In operation 512, the symbol signal input from the 0 inserting unit 510 is filtered by applying a 'P * Q times oversampled filter' to the symbol signal, thereby oversampling the symbol signal by (P * Q) times. The multiplier 514 multiplies the sampled signal (by (P * Q) times per symbol) with cos (2n [pi] fc / fs). Similarly, on the imaginary side, the zero inserter 516 inserts (P * Q-1) zeros between symbols and the (P * Q) oversampled filter 518 per symbol inserts the zero inserter 516 By filtering by applying a 'P * Q times oversampled filter' to a symbol signal inputted from), the multiplying unit 520 multiplies the symbol signal by (P * Q) times. Multiply the sugar (P * Q) by the sampled signal by sin (2nπfc / fs). The subtractor 522 subtracts the multiplication result of the multiplier 520 from the multiplication result of the multiplier 514 to generate a waveform signal of an IF frequency band (= modulated signal).

According to this FIG. 5, (P * Q) per symbol, if the (P * Q) oversampled filters 512 and 518 have a length of R (where R is a natural number) symbol time, The length of the oversampled filters (512, 518) is the (P * Q * R) tap, only the multiplications needed to obtain one sample (P * Q * R), and (for convenience only the multiplication In order to obtain a sample corresponding to one symbol time, (P * Q * R) * (P * Q) multiplication operations are required. In addition, since the QPSK signal requires pulse shaping filtering on the real axis and the imaginary axis, the required multiplication operation requires twice the calculation amount. As a result, in the case of FIG. 5, the multiplication operation required to obtain a sample corresponding to one symbol time, that is, the multiplication operation required to generate a waveform signal from a symbol signal corresponding to one symbol time, is 2 * ( P * Q * R) * (P * Q). On the other hand, a multiplication operation of 2 * (P * Q) is required for one symbol time, that is, 'modulation' during one symbol period (only considering the multiplication operation).

As described above, the direct implementation method as illustrated in FIG. 5 has a disadvantage of requiring a large amount of computation.

Fig. 6 is a block diagram of another example of a digital transmitter in a general case, which includes a (P * Q) parallel pulse shaping filter 610, a multiplier 616, a (P * Q) parallel pulse shaping filter 620, and a multiplication. The unit 622 may include a subtraction unit 624. Here, the (P * Q) parallel pulse shaping filter 610 includes (P * Q) R tap filters 612-0, 612-1, ..., 612- (P * Q-1), and It may be composed of a parallel-serial converter 614. Although not shown in detail in FIG. 6, the (P * Q) parallel pulse shaping filter 620 is also the same. Although not shown in FIG. 6, the signal input to the (P * Q) parallel pulse shaping filter 610 is an input signal of the 0 insertion unit 510 of FIG. 5, and (P * Q) the parallel pulse shaping filter. The signal input to 620 is the input signal of the zero insertion unit 516 of FIG. 5.

The multiplier 616 of FIG. 6 is the same as the multiplier 514 of FIG. 5, the multiplier 622 of FIG. 6 is the same as the multiplier 520 of FIG. 5, and the subtractor 624 of FIG. 6. Is the same as the subtraction part 522 of FIG.

6 (P * Q) parallel pulse shaping filter 610 of FIG. 6 corresponds to 'P * Q times oversampled filter 512' of FIG. The operation is different from the oversampled filter 512 '. Specifically, the (P * Q) parallel pulse shaping filter 610 of FIG. 6 performs an R tap filter (P) in sampling (P * Q) times per symbol, that is, generating P * Q sample signals per symbol. 612-X) (where X is an integer of 0 ≦ X ≦ (P * Q−1)) generates an X th sample signal (of P * Q sample signals), and R tap filter 0 to Both R-tap filters (P * Q-1) operate simultaneously, thereby generating (P * Q) sample signals simultaneously.

The parallel-to-serial converter 614 of FIG. 6 sequentially sorts and outputs (P * Q) sample signals generated at the same time. The sample signals output as described above are identical to the sample signals output by the 'P * Q oversampled filter 512' of FIG.

The function of the (P * Q) parallel pulse shaping filter 620 of FIG. 6 is also the same as the function of the (P * Q) parallel pulse shaping filter 610 of FIG.

According to FIG. 6, as shown in FIG. 4B, the output signal sequence of the zero insertion unit has a signal structure in which one symbol signal and (P * Q-1) zeros are repeated, and one filter is used. Since the number of coefficients of the filter that effectively affects the result (ie the input signal is multiplied by a non-zero value) is R apart by (P * Q), one in pulse shaping filtering In order to obtain a sample result of R, multiplication of R times is required, and R * (P * Q) multiplication is required to obtain a sample of one symbol interval, provided that only multiplication operations are considered. As mentioned above, the coefficient of “R tap filter X” in FIG. 6 has an index from 0 to (R-1), X has an index from 0 to (P * Q-1), and (P The coefficient of the * Q * R) tap length pulse shaping filter ((P * Q) times sampled per symbol in FIG. 5) has a relation as shown in Equation 2 below.

(2)

(2)

Here, h (k) is the k-th coefficient value of the filter having a (P * Q * R) tap length, and h _{X, P * Q} (k) is the k-th coefficient value of "R tap filter X".

In addition, twice the amount of computation is required to generate the QPSK signal on the real and imaginary axes, respectively. In addition, as in the case of the direct implementation as illustrated in FIG. 5, the oversampled signal is multiplied by the 'cosine sample sequence' by (P * Q) times, the oversampled signal is multiplied by the 'sine sample sequence', and the electron multiplication is performed. The result of the latter multiplication is subtracted from the result to produce a signal modulated to the IF band. In this case, considering only the multiplication operation, the total multiplication required for generating a transmission signal for one symbol period is' for pulse shaping filtering operation (i.e., for generating P * Q sample signals from one symbol signal). In other words, since 2 * R * P * Q 'for generating a waveform signal using one symbol signal and 2 * P * Q' for generating a modulated signal, the total required multiplication operation is '2 * P'. * Q * (R + 1) '. This method is a method that is commonly used as a method that can reduce the amount of association much compared to the (direct) implementation method disclosed in FIG.

7 is a block diagram of another example of a digital transmitter in a general case. The digital transmitter includes a zero insertion unit 712, an A-fold oversampled filter 714, a B-fold interpolator 716, and a multiplier. 718, a zero insertion unit 720, an A-fold oversampled filter 722 per symbol, a B-fold interpolator 724, a multiplier 726, and a subtractor 728. In this specification, each of A and B is a natural number which is one of the factors of (P * Q).

As illustrated in FIG. 7, the zero insertion unit 720 and the zero insertion unit 714 correspond to the zero insertion unit 310 (see FIG. 3) of the pulse shaping filtering unit 210 illustrated in FIG. 2. The A-fold oversampled filter 714 per symbol, the A-fold oversampled filter 722 per symbol, the B-fold interpolator 716, and the B-fold interpolator 724 are shown in FIG. The filter 320 (refer to FIG. 3) of the 210 and the multiplier 718, the multiplier 726, and the subtractor 728 correspond to the modulator 220 illustrated in FIG. 2.

In terms of the real part, the zero inserter 712 inserts (P * Q-1) zeros between symbols, so that N mentioned above is P * Q, and the filter 714 that is A times oversampled per symbol By applying the 'A times oversampled filter per symbol' to the symbol signal input from the zero insertion unit 712, the symbol signal is oversampled by A times and the B times interpolation unit 716 performs the A times oversampling. By multiplying the received signal by B times, the resultant A * B times oversampled signal, which is the result of A * B times oversampling of one symbol signal, is generated, and the multiplication unit 718 generates (per symbol) Multiply the sampled signal by cos (2nπfcT / fs). Similarly, on the imaginary side, the zero insertion unit 720 inserts (P * Q-1) zeros between symbols, and the A times oversampled filter 722 per symbol is input from the zero insertion unit 720. By applying a 'A times oversampled filter per symbol' to the symbol signal, the symbol signal is over sampled by A times and the B times interpolator 724 interpolates the A times and the sampled signal by B times A 'B * oversampled signal' (per symbol) is generated as a result of A * B times oversampling, and the multiplier 726 is oversampled (by P * Q times symbol). Multiply the signal by sin (2nπfcT / fs). The subtractor 728 subtracts the multiplication result of the multiplier 726 from the multiplication result of the multiplier 718 to generate a waveform signal of an IF frequency band (= modulated signal).

In FIG. 7, the B-fold interpolators 716 and 724 are implemented as various interpolators, such as a linear interpolator using two adjacent sample signals, a parabolic interpolator or a cubic interpolator using four sample signals. In the following description, it is assumed to be implemented as a cubic interpolator for convenience of description. The parameters of such interpolators are determined by the time to be interpolated.

Meanwhile, the 'A times oversampled filters 714 and 722' shown in FIG. 7 are implemented according to a direct implementation method such as the 'P * Q oversampled filters shown in FIG. 5'. It may be implemented according to an implementation method such as '(P * Q) parallel pulse shaping filter' illustrated in FIG. 6.

8A and 8B are block diagrams illustrating a filtering and modulation unit 120 according to at least one embodiment of the present invention, wherein the filtering and modulation unit 120 includes an oversample filtering unit 810 and an interpolation and modulation unit. 812 may include.

The oversample filtering unit 810 performs P-sample oversample filtering on the input symbol signal for each symbol. Here, the oversample filtering unit 810 may be implemented according to a direct implementation method such as 'per-symbol (P * Q) oversampled filter' shown in FIG. Q) may be implemented according to an implementation method such as 'parallel pulse shaping filter'.

The interpolation and modulator 812 may interpolate Q-sampled intersampled signals (oversampled P-by-fold by the oversample filter 810) for each symbol in parallel. Hereinafter, unless there is confusion in meaning herein, the oversampled signal refers to an oversampled filtered signal. In this case, Q-time interpolation in parallel means that Q samples are not sequentially generated one by one in Q-time interpolation, that is, Q samples are generated for one oversample signal. This means generating two sample signals at the same time. Of course, the interpolation and modulator 812 may sequentially generate Q sample signals as in the B-time interpolation unit of FIG. 7, instead of performing the Q-time interpolation in parallel.

The interpolation and modulator 812 modulates the frequency band of the waveform signal which is the result of Q times interpolation.

In order for the interpolation and modulator 812 to perform Q-time interpolation in parallel as described above, the interpolation and modulator 812 is interpolated to the interpolator 814 and the modulator 816 as shown in FIG. 8B. Can be done.

The interpolator 814 interpolates the symbols P-fold oversampled Q times in parallel for each symbol. The modulator 816 modulates the frequency band of the result interpolated by the interpolator 814.

8B will be described in more detail with reference to FIG. 9.

9 is a detailed block diagram of a digital transmitter according to a first embodiment of the present invention, wherein a zero insertion unit 912, an oversample filtering unit 914, an interpolation unit 916, a multiplier 922, and zero insertion unit are provided. The unit 924, the oversample filtering unit 926, the interpolator 928, the multiplier 930, and the subtractor 932 may be included.

As shown in FIG. 9, the zero insertion unit 912 and the zero insertion unit 924 are the same as the zero insertion unit 310 (see FIG. 3) of the pulse shaping filtering unit 210 shown in FIG. The unit 922, the multiplier 930, and the subtractor 932 are the same as the modulator 220 shown in FIG. 2, and the oversample filter 914 and the oversample filter 926 are illustrated in FIG. 8A. The intersampler 916 and the interpolator 918 are the same as the intersampler 814 illustrated in FIG. 8B.

The oversample filtering unit 914 oversamples the P times P symbol for every symbol of the input symbol signal to generate a P times oversampled signal (ie, P sample signals).

The interpolator 916 may be composed of Q uncoupled interpolators 918-0 to 918-(Q-1) and a parallel-to-serial converter 920.

Each of the Q uncoupled interpolators, i.e., the q th uncoupled interpolator (where q is an integer of 0 ≦ q ≦ (Q−1)) 918-q (corresponds to the q th uncoupled interpolator) A q th sample signal is generated for one sample signal (P times oversampled signal). Each uncombined interpolator operates simultaneously, i.e., in parallel, whereby Q sample signals are simultaneously generated for one sample signal of its P-fold oversampled signal. The interpolator 916 performs Q times interpolation for each sample signal constituting the P times oversampled signal. The parallel-serial converter 920 sequentially sorts and outputs the generated Q sample signals. The Q sample signals output by the parallel-to-serial converter 920 become part of the result of P * Q oversampling one symbol signal.

On the other hand, when the interpolators 916 and 928 use the parallel interpolator structure in this manner, the number of coefficients of each of interpolators 0 to Q-1 is determined according to the interpolation method used. As mentioned above, for convenience of explanation, the interpolation method will be described using a cubic interpolator. The cubic interpolator uses one past sample signal, one current sample signal, and two future sample signals around the point of oversampling. Since the time interpolated by the transmitter is predetermined, the coefficients of the interpolator can be calculated and stored in advance, and the coefficients from each interpolator 0 to interpolator Q-1 do not change. If you look at the signal sequence of cosine and sine that are multiplied by the output of the interpolator, it is a signal sequence with Q periods.

(3)

(3)

(4)

(4)

Here, f _c / f _s = 1 / Q because the DAC sample rate is designed to be Q times the IF frequency. That is, each of the Q parallelized interpolator outputs is always multiplied by the same cosine or sine value to be multiplied for modulation. Therefore, if the interpolator operation is performed by multiplying the cosine or sin value multiplied for modulation in advance with the coefficient of the interpolator, the function of modulation can be performed simultaneously with interpolation. Using this characteristic, the present invention proposes a combined interpolator as shown in FIG. 10.

10 is a detailed block diagram of a digital transmitter according to a second embodiment of the present invention, and includes a zero insertion unit 1010, an oversample filtering unit 1012, an interpolation and modulation unit 1014, and a zero insertion unit 1022. , And a sample filtering unit 1024, an interpolation and modulation unit 1028, and an adder 1032 may be included. Here, the interpolation and modulator 1014 may include Q combined interpolators 1016-0 to 1016- (Q-1) and a parallel-serial converter 1020, and the interpolation and modulator 1026. ) May include Q coupling interpolators 1028-0 to 1028- (Q-1) and a parallel-serial converter 1030.

As shown in FIG. 10, the 0 inserting portion 1010 and the 0 inserting portion 1022 are the same as the 0 inserting portion 310 (see FIG. 3) of the pulse shaping filtering portion 210 shown in FIG. The sample filter 914 and the oversample filter 926 are the same as the oversample filter 1012 and the oversample filter 1024 illustrated in FIG. 9.

The interpolation and modulator 1014 is implemented as an 'interpolator' having the combined interpolator coefficients as the interpolator coefficients, where the combined interpolator coefficients are Q times the P-fold oversampled signal. Interpolator coefficients that complete the function of interpolating and modulating the P-fold oversampled signal by Q times.

Such 'interpolator' may be implemented using an interpolator called a plurality of combined interpolators, as shown in FIG. 10, or may be implemented as a single interpolator, as shown in FIG. 10.

If the 'interpolator' is implemented as a single interpolator, the 'interpolator' interpolates the combined interpolator coefficients for generating different Q sample signals for one sample signal of its P-fold oversampled signal. It is a single interpolator with coefficients.

On the other hand, when the interpolator is implemented using a plurality of combined interpolators, the interpolator, that is, the interpolation and modulator 1014, has Q combined interpolators 1016-0 to 1016- (Q-1). )) And the parallel-to-serial converter 1020.

Each of the Q combined interpolators, i.e., the q th combined interpolator (where q is an integer of 0 ≦ q ≦ (Q-1)) 1016-q is the q th (corresponding to the q th combined interpolator) A modulated sample signal is generated for one sample signal (P times oversampled signal). Each joint interpolator operates simultaneously, i.e., in parallel, thereby simultaneously generating Q '(modulated) sample signals' for one sample signal of its P-fold oversampled signal. The interpolation and modulator 1014 performs the Q-time interpolation for each sample signal constituting the P-fold oversampled signal and performs the modulation at the same time. The parallel-to-serial converter 1020 sequentially sorts and outputs the generated Q '(modulated) sample signals'. The Q (modulated) sample signals output by the parallel-to-serial converter 1020 are the result of modulation after P * Q times one symbol signal.

The combined interpolator coefficients described above may be preset first interpolator coefficients and second interpolator coefficients determined according to periods of a modulation signal. Specifically, the combined interpolator coefficient of the combined interpolator 1016-q corresponding to q is the 'first interpolator coefficient (ie, the interpolator of FIG. 9) of the combined interpolator 1016-q corresponding to q. It is the result of multiplying cos (2qπ / Q) by the interpolator coefficient) of 918-q).

The interpolation and modulator 1026 is implemented as an 'interpolator' having the combined interpolator coefficient as the interpolator coefficient, where the combined interpolator coefficient is Q times the P-fold oversampled signal. Interpolator coefficients that complete the function of interpolating and modulating the P-fold oversampled signal by Q times.

Such 'interpolator' may be implemented using an interpolator called a plurality of combined interpolators, as shown in FIG. 10, or may be implemented as a single interpolator, as shown in FIG. 10.

If the 'interpolator' is implemented as a single interpolator, the 'interpolator' interpolates the combined interpolator coefficients for generating different Q sample signals for one sample signal of its P-fold oversampled signal. It is a single interpolator with coefficients.

On the other hand, when the interpolator is implemented using a plurality of combined interpolators, the interpolator, that is, the interpolation and modulator 1026, has Q combined interpolators 1028-0 to 1028- (Q-1). ) And the parallel-to-serial converter 1030.

Each of the Q combined interpolators, i.e., the q th combined interpolator 1028-q, converts the q th modulated sample signal (corresponding to the q th combined interpolator) into one of its (P times oversampled) signals. Generate for the signal. Each joint interpolator operates simultaneously, i.e., in parallel, thereby simultaneously generating Q '(modulated) sample signals' for one sample signal of its P-fold oversampled signal. The interpolation and modulator 1026 performs the Q-time interpolation for each sample signal constituting the P-times oversampled signal and performs the modulation at the same time. The parallel-to-serial converter 1030 sequentially sorts and outputs the generated Q '(modulated) sample signals'. The Q (modulated) sample signals output by the parallel-to-serial converter 1030 are the result of modulation after P * Q times one symbol signal.

The combined interpolator coefficient described above may be a first interpolator coefficient and a second interpolator coefficient determined according to a period of a modulation signal. Specifically, the combined interpolator coefficient of the combined interpolator 1028-q corresponding to q is equal to '-sin (2qπ / Q)' in the 'first interpolator coefficient' of the combined interpolator 1028-q corresponding to q. Is the result of multiplying

The adder 1032 outputs the modulated signal by adding the sample signals output by the parallel-serial converter 1020 and the sample signals output by the parallel-serial converter 1030.

10 will be described in more detail.

As shown in Equations (3) and (4), the interpolator and the modulator shown in FIG. 9 use the 'modulated signal sequence becomes a signal sequence whose period is a Q sample'. As described above, the integrated interpolation and modulation unit may be implemented. In FIG. 10, the coefficients of the combined interpolator R0 to the combined interpolator R (Q-1) are cos (2qπ / Q), n = 0, .. , The combined interpolator having a value multiplied by (Q-1) as a fixed coefficient, and the coefficients of the combined interpolator I0 to the combined interpolator I (Q-1) are the interpolators 0 to interpolator Q- of FIG. It is a combined interpolator with a fixed coefficient that multiplies 1) by -sin (2qπ / Q), n = 0, ..., (Q-1).

According to FIG. 10, a multiplication operation required for signal generation of one symbol interval may be calculated as follows. First, in the signal processing block that receives the Re [c (k)] symbol, the multiplication of R * P in the low sample pulse shaping filter operation having the parallel pulse shaping filter structure (i.e., in the oversample filtering section 1012) is performed. Since it requires an operation and requires 4 multiplications per joint interpolator (1016-q) (since it is a cubic interpolator), 4 * P * Q multiplication operations are needed for 'interpolation' for one symbol interval. (R * P) + (4 * P * Q) multiplication operations are needed to generate a modulated signal (only the real side is considered). Since QPSK modulation requires computation on both real and imaginary axes, the total multiplication required for 'pulse shaping filtering' (= P * Q times oversampled signal generation) and 'modulation' is 2 * (( R * P) + (4 * P * Q)), compared with the case proposed in FIGS. 5 to 7, the required number of operations is significantly reduced.

11 is a flowchart illustrating a digital transmission method according to at least one embodiment of the present invention. A description with reference to FIG. 8 is as follows.

The oversample filtering unit 810 performs P times oversample filtering for each symbol of the input symbol signal (operation 1110).

After operation 1110, the interpolation and modulation unit 812 interpolates the (P times) and the sample filtered signal in parallel for each symbol by Q times, and modulates the frequency band of the interpolation result (step 1120). In step 1120, step (b) may be performed by an interpolator having a combined interpolator coefficient as the interpolator coefficient. In this case, the interpolator may be performed by interpolating the P times oversampled signal by Q times. Complete the interpolation and modulation functions.

The program for executing the digital transmission method according to at least one embodiment of the present invention mentioned above in a computer may be stored in a computer-readable recording medium.

Here, the computer-readable recording medium may be a magnetic storage medium (for example, a ROM, a floppy disk, a hard disk, etc.), and an optical reading medium (for example, a CD-ROM, a DVD). : Digital Versatile Disc)).

So far, the present invention has been described with reference to the preferred embodiments. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a block diagram of a digital transmitter according to at least one embodiment of the present invention.

2 is an example of a detailed block diagram of the filtering and modulation unit 120 in a general case.

3 is an example of a detailed block diagram of the pulse shaping filtering unit 210 illustrated in FIG. 2.

4 is a reference diagram for describing an operation of the pulse shaping filtering unit 210 illustrated in FIG. 3.

5 is a block diagram of an example of a digital transmitter in a general case.

6 is a block diagram of another example of a digital transmitter in a general case.

7 is a block diagram of another example of a digital transmitter in a general case.

8A and 8B are block diagrams illustrating the filtering and modulator 120 according to at least one embodiment of the present invention.

9 is a detailed block diagram of a digital transmitter according to a first embodiment of the present invention.

10 is a detailed block diagram of a digital transmitter according to a second embodiment of the present invention.

11 is a flowchart illustrating a digital transmission method according to at least one embodiment of the present invention.

Claims (15)

An oversample filter for oversampling P (where P is a natural number) for each input symbol signal; An interpolation unit for interpolating the oversampled signal in parallel for each symbol, wherein Q is a natural number; And A modulator for modulating the frequency band of the interpolated result; The interpolator may include Q uncombined interpolators for simultaneously generating Q different sample signals for individual symbols included in the oversampled signal; And And a parallel-to-serial converter for sequentially arranging the Q sample signals generated from the Q uncoupled interpolators. delete delete An oversample filter for oversampling P (where P is a natural number) for each input symbol signal; And An interpolation and modulation unit capable of interpolating the oversampled signal in parallel for each symbol, wherein Q is a natural number, and modulating a frequency band of the interpolation result; The interpolation and modulator are implemented as interpolators having combined interpolator coefficients as interpolator coefficients, and the interpolator completes performing interpolation and modulation functions by interpolating the oversampled signal by Q times. Digital transmitter. 5. The apparatus of claim 4, wherein the interpolator is And a single interpolator having said combined interpolator coefficients as interpolator coefficients for generating different Q sample signals for one sample signal of said oversampled signal. 5. The apparatus of claim 4, wherein the interpolator is Q joint interpolators each of which operates simultaneously and generates a corresponding q (where q is an integer of 0≤q≤ (Q-1)) sample signal for one of the oversampled signals. field; And And a parallel-to-serial converter for sequentially outputting sample signals generated from the combined interpolators to output a result to be obtained by performing the interpolation and modulation functions. 5. The method of claim 4, And the combined interpolator coefficients are second interpolator coefficients determined according to the periods of the first interpolator coefficient and the modulation signal. The method according to claim 6, For the real axis, the combined interpolator coefficient of the combined interpolator corresponding to q is the result of multiplying cos (2qπ / Q) by the first interpolator coefficient of the combined interpolator corresponding to q, In the imaginary axis, the combined interpolator coefficient of the combined interpolator corresponding to q is a result of multiplying -sin (2qπ / Q) by the first interpolator coefficient of the combined interpolator corresponding to q. Digital transmitter. (a) oversampling P (where P is a natural number) per symbol for the input symbol signal; (b) interpolating Q (where Q is a natural number) multiplex of the oversampled signal in parallel for each symbol; And (c) modulating the frequency band of the interpolated result; The step (b) may include simultaneously generating different Q sample signals with respect to one sample signal of the oversampled signal; And And sequentially arranging the generated Q sample signals. delete delete (a) oversampling P (where P is a natural number) per symbol for the input symbol signal; And (b) performing interpolation of the oversampled signal in parallel for each symbol in Q (where Q is a natural number) times and modulating a frequency band of the interpolation result; The step (b) is performed by an interpolator having a combined interpolator coefficient as the interpolator coefficient, and the step (b) completes performing the interpolation and modulation function by interpolating the oversampled signal by Q times. A digital transmission method characterized by the above. 13. The apparatus of claim 12, wherein the interpolator And a single interpolator having said combined interpolator coefficients as interpolator coefficients for generating different Q sample signals for one sample signal of said oversampled signal. The method of claim 12, wherein step (b) Simultaneously generating different Q sample signals for one sample signal of the oversampled signal; And And outputting a result to be obtained by performing the interpolation and modulation functions by sequentially outputting the generated Q sample signals. 13. The method of claim 12, The combined interpolator coefficients are first interpolator coefficients and the second interpolator coefficients determined according to the period of the modulation signal.

Images