KR100558237B1 - Transmitter for qam encoded data - Google Patents

Abstract

A first QPSK signal is generated by generating quadrature amplitude modulated signals using a first quadrature phase shift keying (QPSK) modulator to encode a first pair of data bits from the data bits into one of four carrier signal phases. The second QPSK modulator encodes a second pair of data bits into one of four carrier signal phases to generate a second QPSK signal. The first QPSK signal is amplified to a first power level and the second QPSK signal is amplified to a second power level. The first and second amplified signals are combined to generate a signal in which four data bits are encoded. In another embodiment of the invention, instead of the first and second QPSK modulators, an offset quadrature amplitude modulation (OQAM) transmitter is constructed using a new kind of modulation called offset quadrature phase shift keying (OQPSK). The OQPSK modulator encodes data bits by encoding the first sub-group of data bits into the real part of the complex signal at odd moments of the clock and the second sub-group of data bits into the imaginary part of the complex signal at even moments of the clock. . OQPSK modulation offers the advantage that all signal transitions are limited to trajectories around a constant radius circle.

Description

JAM-encoded data transmitter {TRANSMITTER FOR QAM ENCODED DATA}

The present invention relates to the transmission of digital information over a channel with limited bandwidth, such as a radio channel of a telephone line. For example, the digital information includes digitally encoded speech.

It is known that Quadrature Amplitude Modulation (QAM) consists of encoding the data bits into a complex vector signal in which the real part and the imaginary part each take one of a plurality of levels. For example, in 16QAM, the real part and the imaginary part may each take one of four equally spaced values 3, 1, -1, and -3. The 4x4 points that can occur are called constellations and are shown in Fig. 1 (a). The modulation according to this method is to map a 4-bit data value represented by B ₀ B ₁ B ₂ B ₃ to 16 different complex signal values as shown in the figure.

A conventional 16QAM transmitter that generates the above constellation from a 4-bit data value is shown in Fig. 1 (b). The first bit pair, B ₀ B _1, is provided to the first 2-bit digital-to-analog (D / A) converter 101 to determine which of the four values the real part has. A second bit pair, B ₂ B _3, is provided to the second two bit D / A converter 103 to determine which of the four values the imaginary part has. Each D / A converter 101, 103 provides its output to the corresponding of the two low pass filters 105, 107. The function of the low pass filters 105, 107 is to limit the transmitted spectrum by smoothing transitions from one value to another when any of the bits B ₀ B ₁ B ₂ B ₃ change. will be. The low pass filters 105, 107 are preferably Nyquist filters whose characteristics are determined only by the input bits B ₀ B ₁ B ₂ B ₃ at a constant sampling time after the bit change. Will have a value. The smoothed real value 109 is input to the cosine modulator 113 and the smoothed imaginary value 111 is input to the sine wave modulator 115. The modulated sine and cosine waves 117 and 119 are added at the summation point 121 to form a complex modulated carrier signal 123 in which the phase and amplitude vary. In order to preserve the amplitude variation, the prior art had to amplify this signal with a linear power amplifier 125. Because of the Nyquist characteristics of the low pass filters 105, 107, if the output signal vector 127 is sampled at a suitable constant moment, one of the 16 complex values will be observed as shown in the lattice diagram of Fig. 1 (a).

The conventional 16QAM transmitter has the disadvantage that the linear power amplifier 125 is not efficient and the desired 16 constellation points are not observed in the output signal vector 127 if the amplifier has nonlinearity or distortion. Similarly, the desired constellation point will not be observed by intersymbol interference (ISI) unless the communication channel comprising the low pass filters 105, 107 is not strictly a Nyquist channel.

US-A-4 485 357 discloses a method for generating a continuously modulated signal of amplitude and phase, such as an analog SSB signal. The method of this patent is described in the paper by Chireix, "High Power Outphasing Modulation", PROC. IRE, vol. 23, No. 11 (1935). This conventional method combines two isopower amplifiers and drives them with signals of different phase angles by arccosine of the ratio of desired and peak amplitudes. This patent does not disclose a technique for digitally modulating bits.

US-A-4 737 968 discloses a method of using a phase-locked loop primarily to filter the phase of a principally phase modulated signal. This filter removes amplitude modulation that may have existed originally. This phase-filtered signal is upconverted to the desired transmission frequency (at mixer 84).

EP-A-0 382 697 discloses a control voltage generator in a transmitter arrangement for a digitally modulated signal. This document discloses the use of constant envelope power amplifiers in QAM transmitters and the use of OQPSK for data bit encoding.

<Overview>

According to one aspect of the invention, an apparatus for generating a modulated signal from data bits is provided. In one embodiment, a quadrature amplitude modulated signal encodes a first pair of data bits into one of four carrier signal phases to generate a first quadrature phase shift keying Generated from data bits using a Quadrature Phase Shift Keying (QPSK) modulator. The second QPSK modulator encodes a second pair of data bits into one of four carrier signal phases to produce a second QPSK signal. The first QPSK signal is amplified to a first power level and the second QPSK signal is amplified to a second power level. The first and second amplified signals are combined to produce a signal in which four data bits are encoded.

According to a second aspect of the present invention, a new and spectral-efficient modulation of offset quadrature phase shift keying (OQPSK) modulation is disclosed, in which the data bits are alternately cosined. It is encoded into a wave carrier level and a sinusoidal carrier level. In one embodiment, the first subgroup of data bits is encoded at an odd instant of a clock as the real part of the complex signal and the second subgroup of data bits is even instant of the clock. The data bits are encoded by encoding the imaginary part of the complex signal at. OQPSK modulation has the advantage that spectral containment is improved when a constant envelope power amplifier is used because all signal transitions are limited to smooth trajectories around a circle with a constant radius.

According to a third aspect of the present invention, an offset QPSK formed from first and second bits of an OQAM symbol. A transmitter of the present invention is disclosed for Offset Quadrature Amplitude Modulation comprising two or more power amplifiers that amplify MSK, or GMSK signals. The first QPSK (or MSK or GMSK) signal is amplified at the first power level, and the second QPSK (or MSK or GMSK) signal is amplified at the second power level. The first and second amplified signals are combined to produce a signal in which four data bits are encoded.

The objects and advantages of the present invention will be understood by reading the following figures and the following detailed description.

Is a lattice diagram of a constellation of complex signal values generated according to a prior art 16QAM transmitter.

Figure 1 (b) is a prior art 16QAM transmitter.

2 is a block diagram of a transmitter in accordance with an aspect of the present invention.

3 (a) to 3 (e) illustrate signals and constellation points associated with various nodes of one embodiment of the transmitter of the present invention.

4 is a block diagram of a transmitter providing offset 16QAM using OQPSK modulation in accordance with another aspect of the present invention.

5A-5H illustrate signal and constellation points associated with various nodes of one embodiment of an offset 16QAM transmitter of the present invention.

Various features of the invention will be described with reference to the drawings in which like parts are identified with the same reference numerals.

2 is a block diagram of another transmitter 200 in one aspect of the present invention. An advantage of the transmitter 200 is that even when a nonlinear amplifier is used, the 16QAM constellation point can be restored.

In a preferred embodiment, a first quadrature phase shift keying (QPSK) modulator 201 receives two information bits B ₀ B ₁ and modulates it on a carrier wave in a known manner. That is, the QPSK constellation changes the real part (I or cosine component) between +1 and -1 according to the first information bit and the imaginary part (Q or sine component) according to the second information bit + j and-. By changing between j, two bits are coded into one of four vector values ± 1 ± j. The magnitude of the four vectors that can be generated (i.e. ) Are all the same so that the output signal from the first QPSK modulator 201 can be accurately amplified by the constant envelope power amplifier 203. The constant envelope power amplifier 203 and other amplifiers used in the practice of the present invention may be power amplifiers, class-C amplifiers, or B-class amplifiers operated at output saturation. . The I and Q components appearing at the output of the constant envelope power amplifier 203 are shown in FIG. 3 (a) and the corresponding vector constellations are shown in FIG. 3 (b).

The two remaining bits B ₂ B ₃ are encoded by the second QPSK modulator 205 into another QPSK constellation. The output signal from the second QPSK modulator 205 is provided to the second constant envelope power amplifier 207 and the amplifier 207 accurately signals the signal to half the power level of the first constant envelope power amplifier 203. Amplify so that the magnitudes of the I and Q components are produced by the first constant envelope power amplifier 203 Double your stomach. The amplified I and Q components appearing at the output of the second constant envelope power amplifier 207 are shown in FIG. 3 (c). The constellation point generated at the output of the second constant envelope power amplifier 207 is It can be described as and shown in Figure 3 (d).

The outputs of the first and second constant envelope power amplifiers 203, 207 are provided to respective inputs of the means for adding these signals, such as the directional coupler 209 of FIG. 2. The directional coupler 209 is configured to convert the low power signal from the second constant envelope power amplifier 207 to the voltage scale of the high power signal from the first constant envelope power amplifier 203. To scale further. A further scaled signal is added to the high power signal. As such, scaling using a passive, lossless combining network, such as a directional coupler 209, And It is limited to the value described by.

Relative scaling To ensure this, the coupling factor for low power signals end Where the coupling factor for the high power signal is to be.

The original low power signal has a higher signal level than the high power signal. The low power signal to the high power signal By further scaling, signals of relative level 1/2 of the high power signal are generated. Thus, with a high power signal of ± 1 ± j The further scaled low-power signals combine to produce 16 constellation points, with the real and imaginary parts taking one of four values: ± 1.5 and ± 0.5, respectively, which are scaled by the directional coupler 209. Is reduced again. Max size constellation point ( 1.5, 1.5j), the output from the directional coupler is ( , j) and the peak power value is to be. This is equal to the sum of the amplifier powers. Thus, the coupling arrangement is 100% efficient at peak power output levels. Power amplifier difference In addition to assigning to relative coupling, other methods of distributing the full scaling half between the most significant bit modulated signal and the least significant bit modulated signal may be used. However, according to the preferred arrangement, maximum efficiency can be obtained at the peak power output point. The 16QAM constellation point shown in FIG. 3 (e) is thus correctly restored using constant envelope (nonlinear) amplifiers 203 and 207.

The smoothing of the transition between constellation points in the transmitter of the present invention shown in FIG. 2 can be achieved by smoothing the I or Q transitions within the first and second QPSK modulators (so-called linear filtering) or phase from one point to another. This is done by smoothing the transition. Linear filtering causes the signal to deviate from a constant magnitude between the constellation points and nonlinear, constant envelope amplifiers 203 and 207 distort this amplitude variation. Nevertheless, when the vectors reach the correct value at the proper sampling time, they reach the constellation point correctly. Distortion can often extend spectral energy into neighboring channels. Nevertheless, the spectral containment of the transmission using the 16QAM transmitter of the present invention is superior to other conventional modulations suitable for use with constant envelope power amplifiers. Distortion of the power amplifier can be reduced by known prior art predistortion techniques described in US Pat. No. 5,191,597 to Ekelund et al., Incorporated herein by reference. Echeland et al. Describe the nonlinearity included in quadrature type radio transmitters for nonlinearity and linear digital modulation of termination amplifiers with predetermined transfer functions H _R and H _ø (for amplitude and phase, respectively). A method for compensating is disclosed, wherein the table reference units (ST, CT) are configured to obtain digital sine and cosine values (I (t, α), Q (t, α)) of orthogonal components determined by a predetermined signal vector α. Save it. According to the transfer function H _R , H _o value method for orthogonally modulated radio signals, r (t, α) is calculated by addressing a memory device that stores a number of H _R and H _o values. The sine and cosine of the addressed H _R and H _ø values are also formed. The calculated value is multiplied by the digital value stored in the table reference units (ST, CT) and the inverse of H _R. As a result, new modified values (i (t, α), q (t, α)) for orthogonal components are obtained, which compensate for nonlinearity in the final amplifier.

It is also possible to smooth the transition between constellation points in a constant amplitude trajectory. For example, in the QPSK constellation generated by the constant envelope amplifier 203, the transition between 1 + j and 1-j has a radius This is accomplished by moving clockwise by 90 degrees around the circle. However, it is necessary to use QPSK to transition to the opposite point on the diameter of the circle. Rotating 180 degrees around a circle with a constant radius in the clockwise or counterclockwise direction has less spectral constraints than transitioning through the origin, Transitions through the origin are non-constant amplitude transitions that cannot be satisfactorily handled by the constant envelope amplifiers 203 and 207.

According to another aspect of the present invention, a new modulation method called offset QAM is provided, in which a diametric transition of a constituent QPSK signal is alternated by alternating instead of simultaneously changing the real part and the imaginary part. prevent. This modulation, when in its most common form, does not have the property that a data symbol corresponds to one of a number of constellation points on the grid, but rather that half of the underlying data bits are at odd intervals of the data clock, e.g. at odd intervals) and is encoded in the real part of the obtained complex signal, and the other half is encoded in the imaginary part of the obtained complex signal at even intervals of the data clock in this example. Of course, alternating clock intervals to determine "odd" and "even" intervals is arbitrary and not limiting. However, if we continue with the first example, by sampling the real part of the signal at odd intervals, where the imaginary part is between imaginary values or is indeterminate, or the imaginary part at even time intervals when the real part is indeterminate By sampling, the data is decoded or demodulated. So there is no grid of constellation points, only a group of imaginary or horizontal stripes at even times, and a group of real or vertical lines at odd times.

The offset QPSK, for example, only transitions between 1 + j and 1-j (i.e. by 90 degrees) and does not transition to the opposite side by 180 degrees like -1-j at 1 + j. Thus, all transitions are confined to the trajectory around a circle of constant radius. This constraint yields a constant envelope signal that can be amplified by a high efficiency C-class power amplifier. The spectral containment of the resulting offset 16QAM signal is improved over prior art 16QAM modulation using nonlinear amplifiers.

An embodiment of a transmitter 400 using offset 16QAM is described with reference to FIG. 4. In this embodiment, the first offset QPSK modulator 401 receives two information bits B ₀ B ₁ and modulates them on the carrier according to the technique of the present invention described above. That is, for example, one of the bits B ₀ is encoded by changing the real part (I or cosine component) between +1 and -1 values in the odd interval of the data clock in accordance with the value of the information bit B ₀ , and the other bits ( In the example B ₁ ) is encoded by changing the imaginary part (Q or sine component) between + j and -j values at even intervals of the data clock in accordance with the information bit B ₁ value. This real and imaginary parts are combined in an offset QPSK modulator 401, and the resulting signal is supplied to a constant envelope amplifier 403. All transitions of this signal can be accurately amplified by a constant envelope amplifier 403, which may be a high efficiency c-class power amplifier, because it is confined to a trajectory around a circle with a constant radius to produce a constant envelope signal. . The I and Q components appearing at the output of the constant envelope amplifier 403 are shown in FIG. 5 (a). As mentioned above, this type of modulation does not produce constellation of point vectors as in conventional QPSK modulation. Instead, real components obtain ± 1 values at odd clock times where the imaginary components are indeterminate as shown in Fig. 5 (b), and imaginary components at even clock times where the real components are indeterminate as shown in Fig. 5 (c). Get the value of j.

The two remaining information bits B ₂ B ₃ are encoded by a second offset QPSK modulator 405 into another set of offset QPSKs of vertical and horizontal lines. The output signal from the second offset QPSK modulator 405 is provided to a second constant envelope amplifier 407 which amplifies this signal exactly half the power level of the first constant envelope amplifier 403. do. As a result, the magnitudes of the I and Q components are determined by the first constant envelope amplifier 403. Becomes The amplified I and Q components appearing at the output of the second constant envelope amplifier 407 are shown in FIG. 5 (d). The amplified vertical lines appearing at odd clock intervals are further shown in 5 (e), and the amplified horizontal lines appearing at the output of the second constant envelope amplifier 407 at even clock intervals are shown in further 5 (f).

The outputs of the first and second constant envelope amplifiers 403, 407 are provided to respective inputs of the means for adding these signals, for example the directional coupler 409 of FIG. 4. Directional coupler 409 provides a low power signal from second constant envelope amplifier 407 for voltage scaling of a high power signal from first constant envelope amplifier 403. To scale further. Techniques for such relative scaling in the directional coupler have been described in connection with the embodiment of FIG. 2. The more scaled signal is added to the high power signal. The original low power signal has a higher signal level than the high power signal. The low power signal to the high power signal By further scaling, signals of relative level 1/2 of the high power signal are generated. Thus, with high power signals The further scaled low power signal combines to produce the vertical (real) lines of FIG. 5 (g) during the odd clock time and the horizontal (imaginary) lines of FIG. 5 (h) during the even clock time. The vertical lines can take one of four values: ± 1.5, ± 0.5, which is fully scaled by the directional coupler 409 Is reduced again. Similarly, the horizontal lines can take one of four values: ± 1.5j, ± 0.5j, which is scaled by the directional coupler 409 Is reduced again. One of the 16 different values can be determined by sampling the signal at odd clock times and corresponding even clock times.

In another aspect of the present invention, each of the binary voltage ratios 1: 1/2: 1 by adding scaled addtion using the third and fourth constant envelope power amplifiers with appropriate power scaling and also with an output coupler By contributing to the output constellation with / 4 ..., etc., the above-described principle can be extended to high order QAM constellations having, for example, 64 to 256 points. Such modifications are considered to be within the scope of the invention as set forth in the appended claims.

It is well known in the art that offset QPSK, where the phase angle shifts by only 90 degrees, is more easily spectrally contained when the vector trajectory is limited to follow a constant envelope circle by smoothing the angular change. The rate of change of the phase angle is defined as the instantaneous deviation from the nominal center frequency of the signal frequency. If the phase angle change from one data bit period to the next period is a constant rate, i.e., if the phase angle rotates by ± 90 degrees during one bit period, the corresponding frequency change is 1/4 cycle per bit period or B / 4 Hz (B is the bit rate). This constant rate phase change results in one form of constant envelope OQPSK called Minimum Shift Keying (MSK).

In MSK, the angular change occurs at a constant smooth rate, but its derivative, the instantaneous frequency, changes rapidly as the direction of the phase change changes from clockwise to counterclockwise. Thus, the angle is a continuous function and its derivative (frequency) is also a continuous function (with a steep step), but the derivative of the frequency has infinite discontinuities (dirac functions) at the steep step point.

The rate at which the spectrum falls out of the desired signal band is 6N dB per octave, where N is the minimum order derivative with discontinuities. Thus, in the case of MSK, the second derivative of the phase angle has a discontinuity, so the spectrum falls out of band at 12 dB per octave.

By further filtering the frequency waveform so that abrupt steps are replaced by smooth changes in frequency, the discontinuities are shifted higher from the second derivative of the phase, resulting in a faster spectrum drop. However, whenever constant envelope modulation is used, there is a limit to the spectral limitation because the actual component of the transmitted signal is proportional to the sine and cosine of the phase, which is a nonlinear function. It is experimentally known that under these conditions, the spectral limit can be best obtained by Gaussian-shaped low pass filtering. This variation of MSK is known as Gaussian-filtered Minimum Shift Keying (GMSK) and is used in European digital cellular systems known as Global System for Mobile Communications (GSM). GMSK is actually a series of modulations described by parameter BT, where BT is the product of the -3dB band of the Gaussian filter and the information bit duration [bit period] T. The smaller the value of BT, the better the spectrum tighter spectral containment will be before the nominal constellation point is reached before starting on a new trajectory for the next bit period. This phenomenon is known as a "partial response" and makes it difficult to decode the signal efficiently. The designer of a particular system must find a compromise between spectral limitation and partial response phenomena.

Other frequency-waveform filters, such as Nyquist filters, can be used, which ensures that the signal (at least the transmitted signal, if not the received signal) exactly passes the constellation point, ie no partial response effect occurs.

Each time the transition between constellation points is smoothed by filtering, the shape (trajectory) of the transition depends not only on the start and end points (current data bits) but also on the data bits before and after. The number of data bits in the sequence that affect the shape of the trajectory is equal to the length of the impulse response of the filter. If the filter length is a finite number of data bit periods L, such as when using a finite impulse response (FIR) filter, a two-to-two L multiplier that is a finite number corresponding to all possible patterns of L binary bits The other trajectory shape of -the-power L) is generated. This waveform is precomputed and stored in the lookup table as a series of waveform samples and can be called according to an address formed using the L series of data bits. Recalled samples may be digital-analog converted to generate analog I, Q modulated waveforms. The D / A converter can also be deleted by storing a pre-calculated bit sequence of the Delta-Sigma modulation representation of the waveform as described in US Pat. No. 5,530,722, which is incorporated herein by reference.

It has been explained so far that a group of modulations, including at least OQPSK, MSK and GMSK, are closely related to each other, represent data by essentially the same constellation point, and differ only in the shape of the transition between constellation points. . Any of these modulations can be used in the present invention to encode a pair of data bits, and signals encoded with different data bit pairs are scaled and added to form an offset QAM signal of the present invention.

The invention has been described with reference to specific embodiments. However, it will be apparent to one skilled in the art that the present invention may be embodied in other forms than the preferred embodiments described above. Preferred embodiments are illustrative only and should never be considered limiting.

Claims (48)

An apparatus for generating a modulated signal from data bits, the apparatus comprising: First means for generating a first signal, Second means for generating a second signal, A first power amplifier for amplifying the first signal and supplying a first amplified signal at its output; A second power amplifier for amplifying the second signal and supplying a second amplified signal at its output; Combining means for combining the first and second amplified signals Including, The first means is means for encoding the first pair of data bits into the first signal, The second means is means for encoding the second pair of data bits into the second signal, The first power amplifier amplifies the first signal to a first power level 203, The second power amplifier amplifies the second signal to a second power level 207, The first and second power levels 203, 207 allow the combining means 209 to generate a modulated signal with four data bits encoded. Device characterized in that. The method of claim 1, The first encoding means encodes the first pair of data bits into one of four carrier signal phases to generate a first Quadrature Phase Shift Keying (QPSK) signal. QPSK means 201, And said second encoding means is second QPSK means (205) for encoding said second pair of data bits into one of four carrier signal phases to generate a second QPSK signal. 3. The apparatus of claim 2, further comprising means for smoothing transitions from one phase encoded value of the first and second QPSK signals to another phase encoded value. Device characterized in that. 4. The device of claim 3, wherein said smoothing means comprises at least one low pass filter. 4. The apparatus of claim 3, wherein the smoothing means comprises means for using digitized waveforms previously calculated and stored in a lookup table. 6. The output of the first and second power amplifiers 203, 207 is characterized by the fact that the lookup table holds a precomputed waveform that is precompensated for distortion of the first and second power amplifiers. And a reduced distortion. 2. The apparatus of claim 1, wherein each of the first and second power amplifiers is a class-C amplifier. 2. The apparatus of claim 1, wherein each of the first and second power amplifiers is a class-B amplifier. An apparatus for encoding groups of data bits into a complex signal for transmission, the apparatus comprising: Means (401, 405) for encoding the first subgroups (B0, B2) of data bits into the real part of the complex signal at odd moments of the clock; Means (401, 405) for encoding the second subgroup (B1, B3) of data bits into an imaginary part of a complex signal at an even time of the clock; First and second power amplifiers 403 and 407 for amplifying the real and imaginary parts of the complex signal; Combining means 409 for combining the amplified real and imaginary parts to form a complex signal for transmission Apparatus comprising a. The method of claim 9, The means for encoding the first subgroup comprises means for selecting one of a plurality of predetermined signal values according to bit polarities of the first subgroup of data bits. Including, Means for encoding the second subgroup comprises means for selecting one of a plurality of predetermined signal values according to the bit polarity of the second subgroup of data bits. 11. The apparatus of claim 10, wherein the predetermined signal values are equispaced with each other. The method of claim 1, The first encoding means is means 401 for encoding the first pair of data bits into a first Offset Quadrature Phase Shift Keying signal (OQPSK signal), And said second encoding means is means (405) for encoding said second pair of data bits into a second OQPSK signal. 13. The method of claim 12, wherein the transition from one coded signal value of the first and second OQPSK signal to another coded signal value is smoothed, thereby obtaining spectral containment of the complex vector modulated signal. And further smoothing means. 14. The apparatus of claim 13, wherein the smoothing means comprises at least one low pass filter for smoothing a transition from one coded signal value of the first and second OQPSK signals to another coded signal value. Device. 14. The apparatus of claim 13, wherein the smoothing means uses means for using a previously calculated and stored discrete transition waveform to smooth the transition from one coded signal value of the first and second OQPSK signals to another coded signal value. Apparatus comprising a. 16. The stored discrete transition waveform is precompensated for distortion in the first and second power amplifiers so that the amplified signals supplied by the first and second power amplifiers 403,407 are substantially And a reduced distortion. 16. The apparatus of claim 15, wherein the transition of the stored discrete transition waveforms follows a constant amplitude trajectory. 18. The apparatus of claim 17, wherein the constant amplitude trajectory is formed using Gaussian Minimum Shift Keying modulation. 14. The apparatus of claim 13, wherein the smoothing means smoothes the transition from one coded value to another while maintaining a constant amplitude signal trajectory. 20. The apparatus of claim 19, wherein the constant amplitude signal trajectory forms a Gaussian minimum transition keying (GMSK) signal. 13. The apparatus of claim 12, wherein the first and second power amplifiers (403, 407) are constant envelope power amplifiers. 13. The apparatus of claim 12, wherein the first and second power amplifiers (403, 407) are operated at output saturation. 13. The apparatus of claim 12, wherein the first and second power amplifiers (403, 407) are class-C amplifiers. 13. The apparatus of claim 12, wherein the first and second power amplifiers (403, 407) are class-B amplifiers. A method of generating a modulated signal from data bits, the method comprising: Generating a first signal; Generating a second signal, Amplifying the first signal to generate a first amplified signal; Amplifying the second signal to generate a second amplified signal; Combining the first and second amplified signals Including, The first signal is generated using an encoding technique for encoding the first pair of data bits to generate the first signal, The second signal is generated using an encoding technique for encoding the second pair of data bits to generate the second signal, The first amplified signal is generated by amplifying the first signal to a first power level 203, The second amplified signal is generated by amplifying the second signal to a second power level 207, The first and second power levels 203 and 207 allow the combining means to produce a modulated signal with four data bits encoded. Characterized in that the method. The method of claim 25, The first signal is generated by encoding the first pair of data bits into one of four carrier signal phases using quadrature phase shift keying (QPSK) 201 to generate a first QPSK signal, And the second signal is generated by encoding the second pair of data bits into one of four carrier signal phases using a QPSK (205) to produce a second QPSK signal. 27. The method of claim 26, further comprising smoothing transitions from one phase coded value of the first and second QPSK signals to another phase coded value. 28. The method of claim 27, wherein the smoothing comprises smoothing the transition from one phase coded value to another phase coded value of the first and second QPSK signals using one or more low pass filters. Characterized in that the method. 28. The method of claim 27, wherein the smoothing step uses a discrete waveform precomputed and stored in a lookup table to smooth the transition from one phase coded value of the first and second QPSK signals to another phase coded value. And comprising a step. 30. The method of claim 29, wherein the lookup table holds a precomputed waveform that is precompensated to reduce distortion resulting from the generation of the first and second amplified signals. The method of claim 26, Generating (203) the first amplified signal comprises amplifying the first QPSK signal to the first power level using a first C-class amplifier, Generating (207) the second amplified signal comprises amplifying the second QPSK signal to the second power level using a second C-class amplifier. The method of claim 26, Generating (203) the first amplified signal comprises amplifying the first QPSK signal to the first power level using a first class B amplifier; Generating (207) the second amplified signal comprises amplifying the second QPSK signal to the second power level using a second Class B amplifier. A method for encoding groups of data bits into a complex signal for transmission, the method comprising: Encoding the first subgroup of data bits into a real part of the complex signal at an odd instant of the clock, Encoding a second subgroup of data bits into an imaginary part of a complex signal at an even moment of the clock, Amplifying the real and imaginary parts of the complex signal; Combining the amplified real and imaginary parts to form a complex signal for transmission Method comprising a. The method of claim 33, wherein Encoding the first subgroup comprises selecting one of a plurality of preset signal values according to the bit polarity of the first subgroup of data bits; Encoding the second subgroup comprises selecting one of a plurality of predetermined signal values according to the bit polarity of the second subgroup of data bits. 35. The method of claim 34, wherein the predetermined signal values are equidistant from each other. The method of claim 25, The first signal is generated by encoding (401) the first pair of data bits into a first OQPSK signal, The second signal is generated by encoding (405) the second pair of data bits into a second OQPSK signal. 37. The method of claim 36, further comprising smoothing a transition from one coded signal value of the first and second OQPSK signal to another coded signal value to thereby obtain spectral constraints of the complex vector modulated signal. Characterized in that the method. 38. The method of claim 37, wherein the step of smoothing comprises smoothing a transition from one coded signal value of the first and second OQPSK signals to another coded signal value using a low pass filter. How to. 38. The method of claim 37, wherein the smoothing step comprises smoothing a transition from one coded signal value of the first and second OQPSK signals to another coded signal value using a precomputed and stored discrete transition waveform. Characterized in that. 40. The method of claim 39, wherein the stored discrete transition waveform is previously compensated for distortion in the first and second power amplifiers 403, 407, and is supplied by the first and second power amplifiers 403, 407. Wherein the amplified signal comprises a significantly reduced distortion. 40. The method of claim 39, wherein the transition of the stored discrete transition waveforms follows a constant amplitude trajectory. 42. The method of claim 41, wherein the constant amplitude trajectory is formed using Gaussian minimum transition keying modulation. 38. The method of claim 37, wherein the smoothing step of smoothing the transition from one coded value to another coded value comprises maintaining a constant amplitude signal trajectory. 44. The method of claim 43, wherein the constant amplitude signal trajectory forms a Gaussian minimum transition keying (GMSK) signal. 37. The method of claim 36, wherein the first and second power amplifiers (403, 407) are constant envelope power amplifiers. 37. The method of claim 36, wherein the first and second power amplifiers (403, 407) are operated at output saturation. 37. The method of claim 36, wherein the first and second power amplifiers (403, 407) are C-class amplifiers. 37. The method of claim 36, wherein the first and second power amplifiers (403, 407) are B-class amplifiers.