Transmitter for qam encoded data

A transmitter for encoding encoded data bits

Full description

Background of the invention

The striped invention relates to sending digital data over channels with limited bandwidth,

Such as wireless communication channels for telephone lines. For example, digital data could include digitally encoded audio.

It is known that quaternary value modulation (QAM) involves encoding data pulses into composite vector signals that can take both real and imaginary parts of one multilevel. For example, both the real section and the imaginary section in 16QAM can take one of four equally spaced values 3, 1, -1 or -3. The possible 4×4 points that appear are called the set, and are shown in Figure 1(a). As shown in the figure, modulation according to this technique involves matching the four-pulse data values, known as B3B2B1B0, to the 16 characteristic values of a composite signal.

A 16QAM transmitter from the previous field is shown in Figure 1(b), generating the above-described set of four pulse data values. The first pair of pulses B1B0 feeds the digital-to-standard data 101 to the D/A converter for the first pair of pulses to determine which of the four values the real section will receive. The second pair of pulses B3B2 is fed to the D/A converter 103 for the second of two pulses in order to determine which of the four values the imaginary section will receive. The D/A converters 101, 103 each feed their output to one of the low-pass filters 105, 107. The function of the low-pass filters 105, 107 is to contain the transmitted spectrum by smoothly transitioning from one value to the next if any of the B3B2B1B0 pulses changes. The low-pass filters 105, 107 are preferably Nyquist filters, which have the property that, at regular sampling times after a pulse shift, the filter output will obtain exactly the value specified by the input B3B2B1B0. Smoothed real values 109 were applied to the cosine integrator 113, while smoothed imaginary values were applied to the sine vector integrator 115. The values of the modulated sine and cosine waves 117, 119 are summed at the summation point 121 to form a modulated composite carrier signal 123 which varies in both phase and value. To preserve the value variations, the former requires the amplification of these

Signal by linear power amplifier 125. Due to the Nyquist property of the low-frequency filters 105, 107, if you sample the vector output signal 127 at the appropriate regular moments, you will see one of the composite values 16 as shown in the grid diagram in

The former domain 16QAM transmitter has the disadvantage that the linear power amplifier 125 is ineffective, and that if it exhibits distortion or non-linearity, then the desired set of points 16 will not be seen in the output vector signal 127. Equally, if the communications channel containing the low-pass filters 105, 107 is not Nyquist conclusive, interference from symbol changes (ISI) will prevent the desired set of points from being seen.

General description of the invention

According to one aspect of the invention, an ingenious transmitter is provided for quadruple value modulation transmission using constant-curve transmitter amplifiers connected with high efficiency better than linear pumps in which a first amplifier of a first level amplifies signals consisting of pulses of highest symbol importance in QAM and a second amplifier of The lowest power level signals are made up of pulses with the lowest significance of the symbol in QAM. In one embodiment, quad-valued modulated signals are generated from the data pulses by using a first modulated quad phase shift shift (QPSK) to encode the first pair of data pulses into one of the four phases of the carrier signal, thus generating the first QPSK signal. The second QPSK modulator encodes the second pair of data pulses in one of the four phases of the carrier signal, thus generating the second QPSK signal. The first QPSK signal is amplified to the first level, and the second QPSK signal is amplified to the second level. The first and second signals are then combined to generate the signal in which the four pulses of data are encoded.

According to a second appearance of the invention. A novel method for active-spectral modulation is disclosed, which is named skewed quadrupole phase shift modulation (OQPSK), in which data pulses are alternately encoded into cosine and sine carrier signal levels. In one embodiment, the data pulses are encoded by encoding the first subset of data pulses in the real section of the component signal at odd moments of a timer, and by encoding the second subset of data pulses in the imaginary section of the component excitation at even moments

For temporary. OQPSK modulation provides the benefit of all signal transformations being restricted to smooth paths and circles of constant radius, allowing better spectral containment when constant curve amplifiers are used.

According to a third aspect of the invention, an inventive skewed quadrilateral value modulation transmitter is disclosed comprising two or more connected constant transmit power amplifiers which amplify skewed MSK or GMSK signals respectively formed from pulses of maximum importance and pulses of minimum importance from the symbol in OQAM.

Brief explanation of the drawings

You will understand the objectives and features of the invention after reading the detailed description combined with the drawings in which:

Figure 1(a) is a network diagram of a set of values of a composite signal generated by a 16QAM transmitter according to the previous range;

Figure 1(b) is a block diagram of a transmitter for the previous field;

Figure 2 is a phase diagram of a transmitter according to one aspect of the invention;

Figures 3(a)-3(e) show the signals and set of points associated with various crossover points for one embodiment of the inventive transmitter;

Figure 4 is a block diagram of a transmitter employing OQPSK modulation to provide skewed 16QAM according to another aspect of the invention; And

Figures 5(a)-5(c) show the signals and the set of points associated with various crossover points for one embodiment of the novel diffracted 16QAM transmitter.

The various features of the invention relating to figures, in which similar parts are identified by the same indicated symbols, will now be described.

Figure 2 is a block diagram of a transmitter 200 according to one aspect of the invention. One advantage of the transmitter 200 is its ability to generate a 16QAM point array even if nonlinear amplifiers are used.

In the best modification, the first quadrupole phase shift modulator (QPSK) 210 receives two pulse information, B1B0, and modulates them onto a carrier according to common techniques. That is, the group QPSK encodes two pulses on one of the four values of the vector j ± 1 ± by scaling the real section (I or cosine component) between the values +1 and -1 according to the first information pulse and the imaginary section (Q or sine component) between the values J+ and J- According to the second pulse of information. Since all vectors

<img file="SA1015B1_D0001.tif" />

1,414), it can amplify the output

The four that can be generated have the same value (i.e.,

The signal from the first QPSK 201 is identically integrated to the original by a constant curve power amplifier 203. Optionally, the constant-curve amplifier 203 can be, in addition to other types used when applying the invention, a driver-powered amplifier at the output saturation, a type-C amplifier, or a type-B amplifier. The components I and Q that appear at the output of an amplifier with the constant magnitude of curve 203 are shown in Figure 3(a), and the corresponding set of vectors is shown in Figure 3(b).

The remaining two information pulses are encoded in another QPSK set by the second integrator 205. The output signal from the second inverter is then fed to the constant curve amplifier 207 which, identically to the original, amplifies this signal to a power level that is half that of the first constant curve amplifier 203, such that the value of the components I and Q is equal to the quotient

<img file="SA1015B1_D0002.tif" />

In that which was generated by the first constant curve amplifier 203. Components I and Q

to hit

The amplifier appearing at the output of the second constant-curve amplifier 207 is shown in Figure 3(c). Thus, the set points generated at the output of the second amplifier are described by the constant curve 207 B

It is shown in Figure 3(d).

<img file="SA1015B1_D0003.tif" />

The output values of the first and second constant-curve amplifiers 203, 207 are then fed to the inputs of the means of collecting these signals, such as the directional comparator 209 shown in Figure 2. Furthermore, the directional comparator 209 incorporates the first power signal from the second fixed-curve amplifier 207.

By an amount equal to

<img file="SA1015B1_D0004.tif" />

Regarding the voltage grading of the high power signal from the first fixed amplifier

Curve 203. The additional scaled signal is then added to the higher power signal. These gradients can be achieved using the combination of latent and lossless networks such as the directional comparator 209

<img file="SA1015B1_D0005.tif" />

Limited to the values described by K and

The coupling coefficient should be k relative to the power signal

<img file="SA1015B1_D0006.tif" />

To achieve an equal relative grading

<img file="SA1015B1_D0007.tif" />

The lowest is equal

Then the coupling coefficient for the higher power signal is equal to:

<img file="SA1015B1_D0008.tif" />

Additional relative grading includes D

<img file="SA1015B1_D0009.tif" />

Previous staging

<img file="SA1015B1_D0010.tif" />

The lower power signal relative to the higher power signal, with

Signal level relative to the higher power signal, combined to generate a level signal

Relative 1/2 that to the higher power signal. Therefore, the upper power signal with values 1±j± is combined with the lower power signal to be further scaled to 1±j±/2 to generate a set of sixteen points in the real and imaginary parts, each of which takes one of the four values ±1.5, ±0.5, which then constitutes a premium. On it reduced by

Overall grading of

<img file="SA1015B1_D0011.tif" />

The result of the directional comparator 209. Hence the set of points for the largest amount

(± 1.5 ± 1.5 j) The output values are from the directional comparator

<img file="SA1015B1_D0012.tif" />

Ability

Corresponding to a peak level

<img file="SA1015B1_D0013.tif" />

This is equal to the total power of the amplifier. So, have an arrangement

Coupling with 100% efficiency at peak output power level. It is possible to use other distributions for an overall scaling of 1/2 between the signal contained in the pulse of greatest importance and that in the pulse with the greatest importance.

<img file="SA1015B1_D0014.tif" />

<img file="SA1015B1_D0015.tif" />

Otherwise, the preferred arrangement gives maximum efficiency at the points of peak output power. The 16QAM array points, shown in Figure 3(e), are thus generated identically to the original despite the use of two fixed-curve (non-linear) amplifiers 203, 207.

In the innovative transmitter shown in Figure 2, smoothing the transitions between group points within the first and second QPSK modulators 201, 205 is achieved either by smoothing the I or Q transitions (called linear filtering) or by smoothing the phase transitions from one point to the other. Linear filtering causes the signal to deviate from a fixed value between set points, and the non-linear fixed-curve amplifiers 203, 207 will distort these changes in value. However, if the vectors reach their correct values at the appropriate sampling timing, we will obtain the group points correctly. Inter-temporal distortion causes the spread of spectral energy to expand to places within adjacent channels. However, the spectral containment of the transmission is using the innovative 16QAM transmitter

Better spectral containment than other pre-domain modulation methods suitable for use with constant curve power amplifiers. Distortion in power amplifiers may also be reduced by use of well-known distortion pretreatment techniques such as those described in U.S. Pat. No. 597,5191, of Oakland et al., incorporated herein by reference. In the patent of Auckland et al., a method of equalizing non-linear variations in a terminal amplifier having a shift function HR, Hø (for value and phase, respectively) is described and included in a quad-type radio transmitter for the purpose of linear digital modulation, in which the analog table units (ST, CT) stores numerical sine and cosine values (I (t,α), Q (t,α))

For the quadrilateral components defined by the given signal vector α. According to the value method of the transformation function HR, Hø for four-way modulated wireless signals, (t,α) r is calculated using memory addressing units that store a number of values of Hø, HR. Also formed the sine and cosine values for the labeled values Hø, HR. The just calculated values are multiplied by the numeric values stored in the units of the lookup table (ST, CT) and by the reciprocal of the HR value. As a result, he obtained new normalized values (t,α) i (t,α),q for the quadrilateral components, which are equivalent to the non-linear variations in the final amplifier.

<img file="SA1015B1_D0016.tif" />

Equal

It is also possible to smooth the transitions between the set of points in the fixed value path. For example, the transformation between the values j+1 and j-1 in the QPSK array generated by the curve constant amplifier 203 may be accomplished by moving clockwise at a 90 degree angle around a circle of equal radius. On the other hand, it may require the use of QPSK, shifting to a point Diagonal offset, neither clockwise nor counterclockwise rotation by 180 degrees around a circle of constant radius provides good spectral fit compared to the shift through the origin, which is a non-constant shift that may not be satisfactorily treated by Hard subwoofer Curve 203, 207.

According to another aspect of the invention, a new modulation method called QAM provides an equalizer in which diagonal shifts in the fundamental component of the signals are prevented by changing the real and imaginary sections at alternate time intervals rather than changing them at the same time. This modulation in its most general form does not have the property that the data symbols correspond to one of a number of points on the grid, but it has the property that half of the underlying data segments are encoded into the real values of a composite signal acquired at, about, odd periods of the data timer and are encoded The other half is in

The imaginary values of a composite signal obtained at, in this example, even periods of the timer. Of course, labeling the alternating periods of the timer as “odd” or “even” periods is arbitrary, and is not intended to be specific. On the other hand, in continuation of the first example; Resolving data encoding or embedding by sampling the real section of the signal at odd-numbered intervals, at timings when parts of the imaginary section are between imaginary values or otherwise averaged, and sampling the imaginary section at even-timing intervals when the real section is averaged. Thus there is no grid of dots, but instead a set of imaginary or horizontal 'stripes' at even times and a set of real or vertical 'stripes' at odd times.

For example, a skewed QPSK only has shifts between j+1 and j-1 (i.e., within 90 degrees) and never has diagonal shifts within 180 degrees, say from j+1 to j-1. Transformations can thus be restricted to paths around circles with a fixed radius. This limitation produces flat-curve signals that can be amplified by Type-C power amplifiers. High efficiency. Therefore, the spectral containment of the resulting diffracted 16QAM signal is improved compared to previous domain modulation and 16QAM methods that use non-linear amplifiers.

An embodiment of a transmitter 400 employing 16QAM diffracted 4 will now be described.

In this embodiment, the first deflected QPSK modulator 401 receives two B1B0 information pulses and modulates them onto a carrier wave according to the inventive technique described above. That is, one of the pulses, for example B0, is symbolized by the real division (I or cosine element) between the values +1 and -1 at the individual periods of the data timer according to the value of the information pulse B0, and the other pulse (in this example B1) is symbolized by the division The imaginary (Q or sine component) between the values Q; And - at even intervals of the data timer according to the value of the information pulse B1. These real and imaginary bias-inline sections 401 are combined and the resulting signal is fed to a bias-inline amplifier 403. Since all transformations of this signal are restricted to paths around circles of fixed radius, thus producing constant-curve signals, they can be amplified identically to the original by the constant-curve amplifier 403, which may be a high-efficiency Type-C amplifier. The I and Q components appearing at the output of the constant curve amplifier 403 are shown in Figure 5(a). As explained above, this type of modulation does not produce a point set of vectors, as in traditional QPSK modulation. Instead of that,

The real components take values of ±1 at odd timer timings when the imaginary components are averaged, as shown in Figure 5(b), and the imaginary components take values of ±j at even timer times when the real components are averaged, as shown in Figure 5(c). .

The remaining two information pulses B3B2 are encoded in another set of QPSK shifted from the vertical and horizontal strips by means of a second shift 405 inline. The output signal is then fed from the second QPSK modulator. Deflector 405 to the second constant-curve amplifier 407 which, identically to the original, amplifies this signal to a magnitude level that is half that of the amplifier.

<img file="SA1015B1_D0017.tif" />

Multiplied by that

The first is a constant of the curve 403, such that the values for the components I and Q will be equal

The output from the first amplifier is constant curve 403. The amplified I and Q components appearing at the output of the second constant-curve amplifier 407 are shown in Figure E(d). The amplified vertical bars that appear during odd periods of the timer are shown in Figure 5(e), and the amplified horizontal bars that appear at the output of the second constant-curve amplifier 407 during even periods of the timer are shown in Figure 5(f).

The outputs of both the first and second constant-curve amplifiers 403, 407 are then fed to the input of a means for collecting these signals, such as the directional comparator 409 shown in Figure 4. The directional comparator 409 additionally incorporates the lower power signal from the second constant-curve amplifier.

<img file="SA1015B1_D0018.tif" />

407 by quantity

Which is proportional to the voltage gradient of the higher power signal

The first fixed-curve amplifier 403. The technique described above to accomplish this relative scaling of the directional comparator with respect to the rendering of Figure 2. The graded signal is then additionally added to the higher power signal.

Bound with the current ratio

<img file="SA1015B1_D0019.tif" />

<img file="SA1015B1_D0020.tif" />

Includes additional relative grading for

To fit the lower power signal to the higher power signal,

To fit the signal level to a higher power signal, to generate a signal

With a relative level of 2/1 that indicates the highest power. Therefore, the higher power signal and the lower power signal are additionally joined by 1±j±/2 to produce the vertical (real) strips as 5(g) during odd timer timings, and the horizontal (imaginary) strips as 5(h) during even timer timings. The vertical bars may assume any of four values of ±1.5, ±0.5 which are further reduced by

Overall grading

<img file="SA1015B1_D0021.tif" />

Output by directional comparator 409. Similarly, we borrow strips

Horizontal is any of the four values ± 1.5 j ± 0.5 j, which are further reduced by the overall scaling.

<img file="SA1015B1_D0022.tif" />

Output by Directional Comparator 409. By sampling the signal at the odd time of the timer and also at the even time of its corresponding timer, one of sixteen different values may be determined.

In another aspect of the invention, for example the principles summarized above may be extended to include a high-order QAM array having 64 or 256 points by adding a third or fourth constant-curve amplifier, of appropriate magnitude and further scaling using the output from the comparator units such that each comparator contributes a portion of Group output in binary voltage ratios 4/1:2/1:1:...etc. Such changes shall be deemed to be consistent with the scope and spirit of this invention as described by the appended claims.

It is well known in this field that a diffractor has shifts of no more than 90 degrees in phase angle, and this is because it is easily fit spectrally when the vector path is constrained to follow a constant circle of the curve by smooth changes in angle. The rate of change of the phase angle is by definition the instantaneous deviation of the signal frequency from its nominal center frequency. If the phase angle changes from one data pulse period to the next available place at a constant rate, that is, by rotating the phase angle by ±90 degrees over the range of one pulse period, the equivalent frequency change is equal to a quarter cycle per pulse period, or B/4 Hz, where B It is a pulse rate. Phase changes at a constant rate cause the formation of a constant curve called the minimum sudden displacement (MSK).

Angle changes in MSK occur at a constant smooth rate, but their derived instantaneous frequency changes in an abrupt, highly variable pattern in which the direction of the phase shift changes from clockwise to counterclockwise. Therefore, the angle is a continuous function, and its derivative (frequency) is also a continuous function (but it has very variable abrupt steps). On the other hand, the frequency derivative has an infinite number of discontinuous points (Dirac functions) at the points whose step is abrupt.

The rate at which the spectrum falls outside the required signal bandwidth is 6N dBs per octet, where is the degree of the lowest derived degree containing discontinuous points; Therefore because the second derivative of the phase angle in this case of MSK contains discontinuous points, the spectrum lies out-of-bandwidth by 12 dB per octet.

By further filtering the shape of the frequency wave to such an extent that abrupt steps of extreme change are replaced by smooth changes of frequency, discontinuous points may have migrated from the second derivative of phase to higher degrees, thus causing the spectrum to bounce rapidly. On the other hand, there are limits to spectral containment whenever curve-invariant modulation is used, because the actual components of the transmitted signal are proportional to the sine and cosine of phase, which are non-linear functions. It was found from the experimental results that, under these conditions, filtering the low frequencies formed by a Gaussian of the frequency waveform gives the best spectral fit that can be achieved; This variation of MSK is known as Gaussian Minimum Shift Snap Filtering (GMSK) and is used in the European digital wireless communications network system known as the Global Mobile Communications System (GSM). In fact GMSK is a uniform set of modulations described by the specific value BT, resulting from the 3dB bandwidth of the Gaussian-amplitude filter of the information pulse, T. Smaller values of BT provide tighter spectral containment at the expense that the signal will not fully reach the nominal set of points before starting a new path for the next pulse period, a phenomenon known as 'partial response', which makes it more difficult to effectively decode the signal. The trade-off between spectral containment and partial response phenomenon is left to the designer of any particular system.

Other frequency-waveform filters may be used, such as Nyquist filters, which ensure that the signal (at least as transmitted if not as received) passes completely through the set of nominal points, that is, it will not appear as a partial response.

Any time a transformation between a set of points is smoothed by filtering, the shape of the transformation (path) depends not only on the start and end points (the current pulse of information) but also on previous and future information pulses. The number of successive pulses of information on which the shape of the path depends is equal to the length of the impulsive response of the filter. If this is a finite number of data pulse periods L, as when a finite impulse response (FIR) filter is used, then a 2-power finite number of different path shapes L are generated, corresponding to all possible types of binary pulses L. They may be calculated in advance. These vector shapes are stored in a lookup table as a sequence of vector shape samples, and are retrieved based on a modulated address using successive data pulses L. The retrieved samples may be digital-to-standard (D/A) converted to generate standard I and Q embedded vector shapes. Also may

D/A converters are deleted by storing representative vector shapes of a previously calculated delta-sigma pulse modulation, as described in US Patent Application No. 5,530,722, which is incorporated herein by reference.

It has been explained above how a certain set of at least global embeddings on MSK, OQPSK and GMSK are related to each other and represent data essentially by the same set of points, differing only in the forms of the transitions between the set of points. Any of these modulation processes may be used in the present invention to encode a pair of data pulses, whereupon the signals encoded by the different pairs of data pulses are modulated and combined to form an innovative-shifted QAM signal.

The invention is described with reference to a particular embodiment. On the other hand, it will be readily apparent to those skilled in the art that it is possible that the invention may be embodied in forms other than that preferred embodiment described above. This may be done without departing from the spirit of the invention. The preferred embodiment is only illustrative and should not be considered limiting in any way. The scope of the invention shall be indicated by the attached elements of protection, in addition to the previous description, and all changes and equivalents that fall within the scope of the intended elements of protection are included in that source.