Balancing amplitude and phase

[0002] 1. Field of the Invention

[0003] The present invention relates generally to digital communication techniques. More specifically, the invention relates to a system and method for balancing the amplitude and phase of a received, quadrature-phase modulated signal.

[0004] 2. Description of the Prior Art

[0005] One of the common methods for modulating digital signals is the use of multilevel systems or M-ary techniques. M-ary modulation techniques are natural extensions of binary modulation techniques and apply to L-level amplitude or phase shift keying. A commonly used quadriphase scheme is called quadrature phage shift keying or QPSK. Like all of the M-ary amplitude or phase schemes, its principle advantage is bandwidth reduction.

[0006] Since pulse rate f_pis:

[0007] f_p=f_slog_LM,  Equation 1

[0008] where f_sis the symbol rate and M is the number of messages; with L representing the number of modulation levels, the larger L is, the smaller the pulse rate and hence, the smaller the bandwidth.

[0009] In telecommunication applications, QPSK modulates two different signals into the same bandwidth creating a two-dimensional signal space. This is accomplished by creating a composite phase modulated signal using two carriers of the same frequency but having a phase difference of 90 degrees as shown in FIG. 1A. By convention, the cosine carrier is called the in-phase component I and the sine carrier is the quadrature component Q. The I component is the real component of the signal and the Q component is the imaginary component of the signal. Each of the I and Q components are bi-phase modulated. A QPSK symbol consists of at least one sample from both the in-phase I and quadrature Q signals. The symbols may represent a quantized version of an analog sample or digital data.

[0010] All phase modulated schemes must overcome the inevitable problem of phase synchronization. For proper operation of QPSK signaling, the I and Q channels should have the same gain throughout processing both received channels, keeping the I and Q channels uncorrelated. Mismatched signal gains or magnitudes between the uncorrelated I and Q channels create errors when processing. Phase differences other than 90 degrees between the signals cause spillover between the channels and similarly result in degraded performance.

[0011] Typical receivers exhibit different overall gains for the separate I and Q channels due to mismatched gains in the mixers, filters, and A/D converters caused by variations in component values due in part to temperature, manufacturing tolerances and other factors. Amplitude and phase imbalance between the I and Q channels result in the distortions shown in FIGS. 1B and 1C, decreasing overall signal-to-noise ratio (SNR).

[0012] Prior art approaches taken to avoid amplitude and phase imbalance rely upon very precise circuitry controlling each gain stage with active temperature compensation. These expensive designs require components that are manufactured with extremely low temperature coefficients and with the mixers for the I and Q channels custom matched during manufacture.

[0013] Accordingly, there exists a need for a system that balances the amplitude and phase of a QPSK signal upon reception increasing signal integrity and thereby reducing bit error rate (BER).

[0014] The present invention balances the amplitude and phase of a received QPSK signal that may have been corrupted during transmission. The output from the system is a signal corrected in both amplitude and phase. The system determines the amplitude of the I and Q channels of a received signal, compares them, and applies a correction to one or both channels correcting amplitude imbalance. For phase imbalance, the system calculates the cross-correlation of the I and Q channels which should average to zero. A correction factor is derived from the cross-correlation product and is applied to both channels, returning the phase cross-correlation to zero.

[0015] Accordingly, it is an object of the invention to provide a system which balances the amplitude of a received QPSK signal.

[0016] It is a further object of the invention to provide a system which balances the phase of a received QPSK signal.

[0017] Other objects and advantages of the system and method will become apparent to those skilled in the art after reading the detailed description of the preferred embodiment.

[0018] FIG. 1A is a plot of a QPSK symbol, balanced in both amplitude and phase.

[0019] FIG. 1B is a plot of a QPSK symbol, amplitude imbalanced.

[0020] FIG. 1C is a plot of a QPSK symbol, phase imbalanced.

[0021] FIG. 2 is a block diagram of an amplitude balancing system in accordance with the present invention.

[0022] FIG. 3 is a block diagram of a phase balancing system in accordance with the present invention.

[0023] FIG. 4 is a vector representation showing phase correction.

[0024] FIG. 5 is a block diagram of a combined amplitude and phase balancing system in accordance with the present invention.

[0025] The preferred embodiment will be described with reference to the drawing figures where like numerals represent like elements throughout.

[0026] An embodiment showing the amplitude balancing system 17 of the present invention is shown in FIG. 2 where two bi-phase modulated signals 19 are input 21I, 21Q. Quantizing is the process of measuring the intensity of a signal in each sample and assigning a digital number to that measured value. Each time the sampling circuit samples the signal, it measures the intensity of the varying analog signal at that discrete moment in time. The input 23I, 23Q data streams represent the discrete samples of data assembled into finite words each having a plurality of bits. The number of bits that define each word determine the total quantization of each sample or symbol. For example, six-bit quantization:

[0027] quantization levels=2ⁿ−1  Equation 2

[0028] with n equal to 6 would yield a resolution of 63 levels. Desired signal resolution determines n.

[0029] Each signal 23I, 23Q component, I and Q, is coupled to an input of an amplifier 25I, 25Q which has an adjustable gain. The output 27I, 27Q of the amplifiers 25I, 25Q are coupled to an absolute value processor 29I, 29Q to obtain the relative magnitudes of each incoming symbol 23I, 23Q. The output 31I, 31Q of the absolute value processors 29I, 29Q are coupled to inputs of respective low pass filters 33I, 33Q.

[0030] The low pass filters 33I, 33Q time-average the received component symbols 23I, 23Q, giving additional weight to recent samples and decreasing weight to previous samples. In the present embodiment 17, IIR (infinite impulse response) filters 33I, 33Q with one pole are used, however, other types of filters or different order IIR filters can also be used without deviating from the principle of the invention. The low pass filter outputs 35I, 35Q present averaged estimates of the sample amplitudes output from the absolute value processors 29I, 29Q.

[0031] A summer 37 obtains the difference from the outputs 35I, 35Q of the low pass filters 33I, 33Q producing an error reference signal 39. If the I and Q components of an input signal 23I, 23Q are orthogonal to each other, the error reference signal 39 will have zero magnitude, indicating a balanced symbol. If the error reference signal 39 produces a value other than zero, the symbols are not amplitude balanced.

[0032] A non-zero-value error reference signal 39 becomes an error correction value. The reference signal 39 is coupled to an input of a hard limiter processor 41. The hard limiter 41 outputs a signal 43 smaller in magnitude, either positive or negative, in dependence upon the error reference signal 39. The hard limiter processor 41 clips the error reference signal 39 magnitude thereby making the sign of the error reference signal 39 a correction factor. This is done for simplifying the implementation, the hard limiter is not essential to the invention.

[0033] The output 43 of the hard limiter processor 41 is coupled to a leaky integrator which is an accumulator 45. The accumulator 45 adds the present value input with an accumulated value from previous input values and outputs 47 a sum. Since the accumulator 45 has a finite bit width, over time, the accumulated value will self-limit in magnitude and plateau if errors persist and are great. The accumulated plurality of error reference signals 39 in the internal accumulator of the accumulator 45 will average to zero when the system reaches stasis.

[0034] The output 47 from the accumulator 45 is coupled to a gain input 49I, 49Q on each adjustable gain amplifier 25I, 25Q. The amplifiers 251, 25Q balance the amplitudes of the received I and Q symbols 23I, 23Q, increasing or attenuating their gains in dependence with the accumulator 45 output signal 47. As can be seen, the reference signal 39 is negative feedback to the upstream amplification stages 25I, 25Q. A positive control voltage at the gain input 49I, 49Q indicates a gain increase for that amplifier; a negative control voltage indicates attenuation.

[0035] If the amplitudes of the input signals 23I, 23Q are not balanced, the system will adjust the variable amplifiers 25I, 25Q (attenuating one component while boosting the other) according to the accumulator 45 output signal 47 until the I and Q symbol amplitudes are within a predetermined tolerance. If the symbol gains are equal, but vary between received symbols, the system 17 will not effect correction. A downstream automatic gain control (AGC)(not shown) equalizes the system output 51I, 51Q for further signal processing (not shown).

[0036] An embodiment showing the phase correction system 61 of the present invention is shown in FIG. 3. Two bi-phase modulated signals 19 are input 63I, 63Q into the system 61. The input 63I, 63Q data streams 65I, 65Q for the I and Q symbols are coupled to a first input 67I, 67Q of parallel summers 69I, 69Q. The output 71I, 71Q of each summer 69I, 69Q are the system output 73I, 73Q and feedback for the phase correction system 61. Both feedback lines 71I, 71Q are coupled to a mixer 75 for correlation. The mixer 75 cross-correlated output signal 77 is coupled to an integrator 79. The integrator 79 time-averages the cross-correlation product 77. The integrator output is coupled to a hard limiter processor 83. The hard decision processor 83 limits the magnitude of the integrated cross-correlation product. The hard decision processor 83 output 85 retains sign. The hard limiter processor 83 output 85 is coupled to an accumulator input 87. The hard decision processor 83 reduces implementation complexity, one skilled in this art would recognize that it is not essential.

[0037] As previously discussed, the function of an accumulator is to accumulate, over-time, the present input value with previous inputs. The sum is output as a correction signal.

[0038] The correction signal 89 is coupled to a first input 91I of a variable gain amplifier 93I coupling the Q input 65Q with the I input 63I. The correction signal 89 also is coupled to a first input 91Q of a variable gain amplifier 93Q coupling the I symbol input 65I with the Q input 63Q.

[0039] The correction signal 89 adjusts both amplifiers 93I, 93Q increasing or decreasing their gain. The amplifier outputs 95I, 95Q are coupled to a second input 97I, 97Q of the input adders 69I, 69Q.

[0040] The phase correction is shown as a vector representation in FIG. 4. The adders 69I, 69Q subtract the portion of Q component 63Q from the I component 65I;

[0041] I=x−ry,  Equation 3

[0042] −I=−x−ry,  Equation 4

[0043] where r Δ cross correlation,

[0044] and the portion of I component 63I from the Q component 65Q;

[0045] Q=y−xr,  Equation 5

[0046] Q=−y−xr,  Equation 6

[0047] where r Δ cross correlation,

[0048] in order to remove the cross correlation contribution from each. Once the parts of the signals that result in the cross correlation are removed, the outputs 71I and 71Q of the adders 69I, 69Q become uncorrelated I, Q and orthogonal in signal space.

[0049] An alternative embodiment combining both systems correcting amplitude 17 and phase 61 imbalance is shown in FIG. 5. The system 101 is a simple series connection outputting 103I, 103Q a symbol corrected in both amplitude and phase. Another combined embodiment where the amplitude balancer 17 follows the phase balancer 61 is also possible.

[0050] While specific embodiments of the present invention have been shown and described, many modifications and variations could be made by one skilled in the art without departing from the spirit and scope of the invention. The above description serves to illustrate and not limit the particular form in any way.