Combining signals exhibiting multiple types of diversity

Abstract

Combining signals includes receiving first signals having a first frequency and second signals having a second frequency. A first weight reflecting a signal-to-noise ratio associated with a first signal is determined for each first signal, and a first signal output is generate from the first signals in accordance with the first weights. A second weight reflecting a signal-to-noise ratio associated with a second signal is determined for each second signal, and a second signal output is generate from the second signals in accordance with the second weights. The first signal output and the second signal output are combined to yield a combined signal output.

Description

TECHNICAL FIELD OF THE INVENTION

[0004] This invention relates generally to communications systems, and more specifically to combining signals exhibiting multiple types of diversity.

BACKGROUND OF THE INVENTION

[0005] Signals from multiple receivers may be combined to determine content included in the signals. Known techniques for signal combining, however, may not be able to combine signals that exhibit multiple types of diversity such as temporal, frequency, polar, or spatial diversity. Consequently, known techniques for signal combining are unsatisfactory in certain situations.

SUMMARY OF THE INVENTION

[0006] In accordance with the present invention, disadvantages and problems associated with previous techniques for combining signals may be reduced or eliminated.

[0007] According to one embodiment of the present invention, combining signals includes receiving first signals having a first frequency and second signals having a second frequency. A first weight reflecting a signal-to-noise ratio associated with a first signal is determined for each first signal, and a first signal output is generate from the first signals in accordance with the first weights. A second weight reflecting a signal-to-noise ratio associated with a second signal is determined for each second signal, and a second signal output is generate from the second signals in accordance with the second weights. The first signal output and the second signal output are combined to yield a combined signal output.

[0008] Certain embodiments of the invention may provide one or more technical advantages. A technical advantage of one embodiment may be that signals are combined in accordance with the signal-to-noise ratios associated with the signals, such that maximal ratio combining gain may be substantially realized. Accordingly, signals that exhibit multiple types of diversity such as temporal, frequency, polar, or spatial diversity may be combined.

[0009] Certain embodiments of the invention may include none, some, or all of the above technical advantages. One or more other technical advantages may be readily apparent to one skilled in the art from the figures, descriptions, and claims included herein.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] For a more complete understanding of the present invention and the advantages thereof, reference is now made to the following description taken in conjunction with the accompanying drawings.

[0011] FIG. 1 is a block diagram of one embodiment of a system for combining signals;

[0012] FIG. 2 is a block diagram of one embodiment of an interpolator of the system of FIG. 1;

[0013] FIG. 3 is a block diagram of one embodiment of a system that includes phase aligners and magnitude estimators;

[0014] FIG. 4 is a block diagram of one embodiment of a reference constructor of the system of FIG. 1; and

[0015] FIG. 5 is a block diagram of one embodiment of coordinating modules of the system of FIG. 1.

DETAILED DESCRIPTION OF THE DRAWINGS

[0016] The invention and its advantages are best understood by referring to FIGS. 1 through 5 of the drawings, like numerals being used for like and corresponding parts of the various drawings.

[0017] FIG. 1 is a block diagram of one embodiment of a system 10 for combining signals. System 10 may be used to combine signals in accordance with signal-to-noise ratios associated with the signals, such that maximal ratio combining gain may be substantially realized. Signal-to-noise ratio weighting is insensitive to the specific cause or causes of variation of signal-to-noise ratio. Accordingly, signals that exhibit multiple types of diversity such as temporal, frequency, polar, or spatial diversity may be combined.

[0018] According to the illustrated embodiment, system 10 combines signals. The signals typically carry symbol streams, and different signals may carry substantially identical symbol streams. The signals may arrive at different times, resulting in a symbol offset among the symbols. In order to combine the signals, the signals are correlated to resolve the symbol offset. Signals may comprise, for example, phase shift keying (PSK) signals such as binary phase shift keying (BPSK) signals or staggered binary phase shift keying (SBPSK). Signals may have any suitable radio frequency and may be modulated according to any suitable modulation format.

[0019] System 10 includes antenna systems 20a-b, frequency branches 22a-b, a sum-weight module 26, a reference constructor 72, and coordinating modules 24 coupled as illustrated in FIG. 1. In general, antenna systems 20a-b receive signals and separate the signals with respect to frequency. Each frequency branch 22a receives the signals of a particular frequency and determines a signal magnitude estimate for each signal according to a signal-to-noise ratio associated with the signal. Sum-weight module 26 combines signals received from frequency branches 22a-b according to the weights determined by frequency branches 22a-b. Reference constructor 72 generates a correlation reference that frequency branches 22a-b use to estimate signal magnitude. Coordinating modules 24 correlate signals of the different frequency branches 22a-b.

[0020] According to the illustrated embodiment, an antenna system 20a-b includes one or more antennas 30a-b and a splitter 32a-b. Antennas 30a-b receive signals from one or more transmitters, and splitter 32a-b separates the signals with respect to frequency. Antennas 30a-b may comprise any antenna suitable for receiving signals from one or more transmitters. Antennas 30a-b may be polarized, separated, or both. Splitter 32a-b may comprise any device suitable for separating the signals according to frequency, and may perform other functions such as amplifying or filtering the signals.

[0021] Frequency branches 22a-b process the signals of a specific frequency, which allows system 10 to accommodate signals exhibiting frequency diversity. In the illustrated embodiment, system 10 includes two frequency branches 22a-b. System 10, however, may comprise any suitable number of frequency branches. A frequency branch 22a-b may include signal branches 40a-b, a phase aligner 42a-b, a magnitude estimator 44a-b, a branch combiner 45a-b, and buffers 66a-b and 68 coupled as shown in FIG. 1.

[0022] A signal branch 40a-b processes signals from an antenna system 20a-b, and includes a receiver channel 50a-b and an interpolator 52a-b. Receiver channel 50a-b receives and processes signals to yield signal images. Interpolator 52a-b applies a delay to the signals in order to compensate for symbol offsets, which may result from temporal, or spatial diversity. Accordingly, interpolator 52a-b allows system 10 to accommodate signals exhibiting such diversity. Interpolator 52a-b may comprise, for example, a polyphase interpolator that uses polyphase finite impulse response (FIR) structures to shift the phase of signals. A finite impulse response filter used in conjunction with a delay line may delay the signals with a phase resolution finer than one sample and allow the magnitude of the delay to vary over many symbol periods. Interpolator 52a-b may also provide additional or other functionality such as dynamic delay control, decimation of signals, or linear-phase channel filtering. Interpolator 52a-b is described in more detail with reference to FIG. 2. Other methods for realizing interpolation may be substituted for the polyphase filter interpolator in this invention.

[0023] Phase aligner 42a-b aligns the phases of signals received from interpolators 52a-b of frequency branch 22a-b. Magnitude estimator 44a-b calculates a signal magnitude estimate for each signal that reflects a signal-to-noise ratio associated with the signal. Magnitude estimator 44a-b is described in more detail with reference to FIG. 3. Branch combiner 45a-b combines the signals of frequency branch 22a-b. Buffers 66a-b and 68 store the combined signals, and may comprise first-in-first-out buffers.

[0024] Sum-weight module 26 generates weights from the signal magnitude estimates received from magnitude estimator 44a-b, and sums the signals according to the weights in order to generate a combined signal. The weights reflect the signal-to-noise ratios at signal branches 40a-d. A signal-to-noise-ratio improvement may be realized if the combined signal-to-noise ratio is better than the worst individual signal-to-noise ratio of a signal branch 40a-d. Maximal Ratio Combining performance may be substantially achieved if the signal-to-noise ratio of the combined output is better than the best individual signal's signal-to-noise ratio. Sum-weight module 26 is described in more detail with reference to FIG. 3.

[0025] Reference constructor 72 generates a correlation reference, which is used to estimate signal magnitude, from the signals at one or more frequency branches 22a-b. Frequency branches 22a-b may be selected in accordance with the signal energies at the frequency branches 22a-b. For example, if frequency branch 22a has acquired a signal, but frequency branch 22b has not acquired a signal, the correlation reference may derived using frequency branch 22a. A correlation reference may be derived from multiple frequency branches 22a-b. The contribution of each frequency branch 22a-b may be determined with respect to signal-to-noise-ratios at the frequency branches 22a-b. A correlation reference derived from multiple frequency branches 22a-b may be used at none, some, or all frequency branches 22a-b. Accordingly, any number of frequency branches 22a-b may contribute to and benefit from the correlation reference. Reference constructor 72 is described in more detail with reference to FIG. 4.

[0026] Coordinating modules 24 determine the correlation among the signals of frequency branches 22a-b, and provide instructions to interpolators 52a-d to align the symbols of the signals in order to correlate the signals. The signals are correlated if the symbols of the symbol streams are aligned, and are uncorrelated if the symbols are misaligned. Coordinating modules 24 include a correlation monitor 64, a correlator 62, and a symbol aligner 60.

[0027] Correlation monitor 64 monitors signal correlation and initiates correlation procedures in response to detecting signal decorrelation. Correlation monitor 64 generates correlation values that indicate positive and negative correlations among the signals. If the correlation values fail to satisfy certain threshold values, correlation monitor notifies correlator 62. Correlator 62 generates a symbol alignment tracking error, which is provided to symbol aligner 60. In response, symbol aligner 60 provides instructions to interpolators 52 to align symbols in order to correlate the signals. Coordinating modules 64 are described in more detail with reference to FIG. 5.

[0028] Modifications, additions, or omissions may be made to the system without departing from the scope of the invention. For example, system 10 may be configured to combine signals from any suitable number of antenna systems 20 at any suitable number of frequencies. For example, system 10 may be configured to combine signals from four antenna systems and at four different frequencies.

[0029] As another example, although system 10 may approximate maximal ratio combining, it is within the scope of the invention to include non-optimal weighting that may result from, for example, employing a non-optimal weight calculator. System 10 may include other compromises to optimal performance resulting from, for example, stability enhancement or single signal enhancement. Additionally, functions may be performed using any suitable logic comprising software, hardware, other logic, or any suitable combination of the preceding.

[0030] FIG. 2 is a block diagram illustrating one embodiment of interpolator 52 of system 10 of FIG. 1. Interpolator 52 provides dynamic delay compensation to equalize symbol offsets between different signals, each of which carry substantially identical symbol streams. Typically, if identical symbol streams of signals are offset in time, the time offset is equalized before combining the signals.

[0031] Interpolator 52 includes a delay line 151, tap weight coefficients 150, multipliers 152 and 154, and an adder 156. Delay line 151 stores samples of the signals received from a receiver channel 50a-b. For example, delay line 151 may comprise a 512-word delay line that stores samples {di}, where 0≦i≦511. Delay line 151, however, may have any suitable length. Delay line 151 may be implemented using linear delay techniques or circular buffering techniques.

[0032] Tap weight coefficients 150 are used to adjust the delay applied to the samples. For example, tap weight coefficients 150 comprise tap weight coefficient sets {hi(φ)} for 0≦i≦127, where φ, 0≦φ≦4, represents a phase of a finite impulse response filter. Phase φ and a pointer value τ determine the relative delay applied to the symbols streams. Initial values for phase φ and for pointer value τ may provided by symbol aligner 60, and the values of pointer value τ and phase φ may be adjusted to change the relative delay. Delay resolution may be controlled by selecting a tap weight coefficient set for an increment of the samples.

[0033] In operation, multipliers 152 and 154 multiply tap weight coefficients hi(φ) with samples di to form products di+τhi(φ). Adder 156 adds the products to yield an output 1$y_n\left(\tau ,\phi \right)=\sum _i{d}_{i+\tau }{h}_i\left(\phi \right)$

[0034] that may be used to apply a delay to the samples. A processor may be used to control the operation of the elements of interpolator 52.

[0035] Modifications, additions, or omissions may be made to interpolator 52 without departing from the scope of the invention. For example, interpolator 52 may include more or fewer tap weight coefficients 150. In addition, the phase resolution of interpolator 52 may be varied to realize a specific filter response. For example, the phase resolution may be increased or decreased by increasing or decreasing the number of coefficients or the number of tap weight coefficient sets, such that a lesser or greater phase difference exists between the adjacent tap weight coefficient sets, respectively. Additionally, functions may be performed using any suitable logic comprising software, hardware, other logic, or any suitable combination of the preceding.

[0036] FIG. 3 is a block diagram of one embodiment of a system 210 that includes phase aligners 241a-b and magnitude estimators 243a-b. In general, system 210 correlates signal images xi(t) with a correlation reference y′(t) to yield phasors ri, and generates signal magnitude estimates Mi from the phasors ri. Weights Wi of signal images xi(t) are determined from the signal magnitude estimates Mi. Each weight Wi reflects the signal-to-noise ratio associated with the corresponding signal image xi(t). “Each” as used in this document refers to each member of a set or each member of a subset of the set. The phase of unity magnitude phasors Ri that substantially track the phases of phasors ri are applied to signal images xi(t) to yield phase-aligned signal images Rxi(t) that are substantially phase aligned.

[0037] Weights Wi are applied to phase-aligned signal images Rxi(t) to yield weighted phase-aligned signal images WRxi(t), which have substantially the same phase. System 210 adds the weighted phase-aligned signal images WRxi(t) and normalizes the result to generate a combined signal output y(t), which is modified to generate correlation reference y′(t).

[0038] According to the illustrated embodiment, system 210 receives signals that are sampled as signal images xi(t) defined by Equation (1):

x i (t)=αid(t)+ni(t)  (1)

[0039] where i corresponds to a signal branch 40a, d(t) represents a signal, αi represents a complex branch gain, and ni(t) represents noise, where each noise term ni(t) is mutually uncorrelated and uncorrelated relative to the signal d(t). Signal d(t) may comprise a phase-shift keying (PSK) signal, and noise ni(t) may comprise additive white Gaussian noise (AWGN).

[0040] Each channel branch 235a-b receives signal images xi(t) and generates weighted phase-aligned signal images WRxi(t) from the signal images. Each channel branch 235a-b includes a phase aligner 241a-b, a magnitude estimator 243a-b, a weight calculator 260a-b, and scalar multipliers 262a-b and 266a-b. Phase aligner 241a-b includes multipliers 222a-b and 226a-b and phase-locked loop (PLL) circuits 224a-b. Complex multipliers 222a-b correlate signal images xi(t) with a correlation reference y′(t) to yield phasors ri. An expression for combined signal output y(t), which may substantially represent correlation reference y′(t), may be given by Equation (2): 2$\begin{array}{cc}y\left(t\right)=\frac{1}{\sum _i{W}_i{M}_i}\sum _i\left[W_i\&LeftBracketingBar;{\alpha }_i\&RightBracketingBar;d^{*}\left(t\right)+W_i{R}_i^{*}{n}_i^{*}\left(t\right)\right]& \left(2\right)\end{array}$

[0041] The correlation of signal images xi(t) with the correlation reference y′(t) may remove modulation, since d*(t)d(t)=1.

[0042] Phase-locked loop circuits 224a-b perform tracking and filtering to yield known magnitude phasors Ri having frequencies and phases that match the frequencies of phasors ri or track the phases of phasors ri. A phase-locked loop circuit 224a-b may comprise a complex multiplier 250a-b, a numerically controlled oscillator (NCO) 256a-b, and a loop filter 254a-b. Complex multiplier 250a-b correlates unity magnitude phasor Ri with phasor ri to drive the imaginary component ei of phasor ri to zero and to resolve the real component mi of phasor ri. The real component mi may be used to represent the magnitude of phasor ri by tracking the magnitude of the signal content αid(t) embedded in signal image xi(t).

[0043] According to one embodiment, a numerically controlled oscillator 256a-b comprises a perfect integrator coupled to a complex sinusoid generator such that the output frequency is proportional to the input drive. According to the embodiment, a loop filter 252a-b comprises an imperfect integrator and a feed-forward path. A loop filter 252a-b, along with an NCO 256a-b, determines the type of phase-locked loop realized. Complex multipliers 226a-b apply unity magnitude phasors Ri to signal images xi(t), adjusting the phase alignment of αid(t) embedded in xi(t), to yield adjusted signal images Rxi(t). The signal images are phase aligned to support phase coherent constructive addition to generate combined signal output y(t).

[0044] Magnitude estimators 243a-b include estimation filters 228a-b that generate signal magnitude estimates Mi from real components mi. Signal magnitude estimates Mi provide an estimate of the signal content of signal images xi(t). An estimation filter 228a-b may comprise, for example, a narrow low pass filter to reduce the variance of the estimate. A real component mi may be based on the stationary ergodic processes of signal image xi(t), so the signal magnitude estimate Mi may provide an accurate estimate of the signal content |αid(t)| of signal image xi(t).

[0045] Expressions (3) through (6) provide an explanation of the estimation process for determining signal magnitude estimate Mi. Expressions (3) through (6) assume that coefficient K1 is set to a value close to unity, such that correlation reference y′(t) substantially comprises combined signal output y(t).

[0046] Equation (3) defines phasor ri: 3$\begin{array}{cc}{r}_1=x_1\left(t\right)y\left(t\right)=\left[{\alpha }_{1}d\left(t\right)+n_1\left(t\right)\right]\left[\frac{\sum _i\left[W_i\&LeftBracketingBar;{\alpha }_i\&RightBracketingBar;d^{*}\left(t\right)+W_i{R}_i^{*}{n}_i^{*}\left(t\right)\right]}{\sum _i{W}_i{M}_i}\right]& \left(3\right)\end{array}$

[0047] Expanding Equation (3) yields Equation (4): 4$\begin{array}{cc}{r}_1={\alpha }_{1}d\left(t\right)d^{*}\left(t\right)\frac{1}{\sum _i{W}_i{M}_i}\sum _i{W}_i\&LeftBracketingBar;{\alpha }_i\&RightBracketingBar;+{\alpha }_{1}d\left(t\right)\frac{1}{\sum _i{W}_i{M}_i}\sum W_i{R}_i^{*}{n}_i^{*}\left(t\right)+n_1\left(t\right)\frac{1}{\sum _i{W}_i{M}_i}{d}^{*}\left(t\right)\sum _i{W}_i\&LeftBracketingBar;{\alpha }_i\&RightBracketingBar;+n_1\left(t\right)\frac{1}{\sum _i{W}_i{M}_i}\sum _i{W}_i{R}_i{n}_i^{*}\left(t\right)& \left(4\right)\end{array}$

[0048] The uncorrelated terms of Equation (4) average to zero at estimation filter 228, resulting in the following relationship described by Equation (5): 5$\begin{array}{cc}{M}_1\cong E\left[m_1\right]\cong E\left[\&LeftBracketingBar;r_1\&RightBracketingBar;\right]\cong \&LeftBracketingBar;{\alpha }_1\&RightBracketingBar;+\frac{W_1}{\sum _i{W}_i{M}_i}{n}_1\left(t\right)R_i^{*}{n}_i^{*}\left(t\right)& \left(5\right)\end{array}$

[0049] If the signal-to-noise ratio is high, the noise component n1(t) is low, and if the signal-to-noise ratio is low, weight W1 is very low. In either situation, the noise-cross-noise term is sufficiently small to substantially ignore. Therefore, the signal magnitude estimates Mi may be described by Equation (6).

M 1 ≅|α1d(t)|, or generally Mi≅|αid(t)|  (6)

[0050] Sum-weight module 230 determines weights Wi applicable to signal images xi(t) from signal magnitude estimates Mi. Each weight Wi reflects the signal-to-noise ratio associated with the corresponding signal image xi(t). Within sum-weight module 230, weights Wi are applied to phase-aligned signal images to yield weighted phase-aligned signal images. Sum-weight module 230 combines the weighted phase-aligned signal images and normalizes the result to generate the combined signal output y(t), which is modified to yield correlation reference y′(t).

[0051] Weight calculators 260a-b calculate weights Wi from signal magnitude estimates Mi. According to classical theory, maximal-ratio combining requires that signal images xi(t) be weighted in accordance with S/N2 and that {E[|xi(t)|]}=C is maintained by the automatic gain control. Accordingly, E[n12(t)]=C2−M12. As a result, in general the weights Wi may given by Equation (7): 6$\begin{array}{cc}{W}_i=\frac{M_i}{C^2-M_i^2}& \left(7\right)\end{array}$

[0052] Scalar multipliers 262a-b apply weights Wi to adjusted signal images Rxi(t) to generate weighted adjusted signal images WRxi(t). A summing node 264 combines the weighted adjusted signal images WRxi(t) to yield weighted sum SumA(t). The signal component of weighted sum SumA(t) may be defined by Equation (8): 7$\begin{array}{cc}\&LeftBracketingBar;\sum _i{W}_i{R}_i{\alpha }_{i}d\left(t\right)\&RightBracketingBar;=\sum _i{W}_i{M}_i& \left(8\right)\end{array}$

[0053] Multipliers 266a-b apply weights Wi to signal magnitude estimates Mi to yield weighted signal magnitude estimates WMi.

[0054] A normalizer circuit 271 substantially maintains the embedded signal energy of correlation reference y′(t) at unity. Normalizer circuit 271 comprises an adder 268, a multiplicative inverter 270, and a dual scalar multiplier 272. Summing node 268 combines the weighted signal magnitude estimates WMi. Multiplicative inverter 270 inverts the combined weighted signal magnitude estimates WMi to yield a normalization factor Gout, which may be defined by Equation (9): 8$\begin{array}{cc}{G}_{\mathrm{out}}=\frac{1}{\sum _i{W}_i{M}_i}& \left(9\right)\end{array}$

[0055] Multiplier 272 normalizes signal component of weighted sum SumA(t) to unity by applying normalization factor Gout to weighted sum SumA(t). In the final expression, |Rαi|=|αi|, and combined signal output y(t) may be described by Equation (10): 9$\begin{array}{cc}y\left(t\right)=\frac{\sum _i{W}_i{R}_i{x}_i\left(t\right)}{\sum _i{W}_i{M}_i}=\frac{1}{\sum _i{W}_i{M}_i}\sum _i\left[W_i\&LeftBracketingBar;{\alpha }_i\&RightBracketingBar;d^{*}\left(t\right)+W_i{R}_i{n}_i^{*}\left(t\right)\right].& \left(10\right)\end{array}$

[0056] Carrier recovery loop 233 drives phase-locked loop circuits 224a-b to resolve the combined signal output y(t) to baseband, that is, zero frequency error. Carrier recovery loop 233 includes a collective phase detector 231 and a loop filter 232 that derive a phase error measurement, which may be applied to phase-locked loop circuits 224a-b. Carrier recovery loop 233 may have, for example, a second order Type-II topology.

[0057] Modifications, additions, or omissions may be made to system 210 without departing from the scope of the invention. For example, a carrier recovery loop may be employed outside of system 210, such that carrier recovery loop 233 may be omitted. Carrier recovery loop 233 is not required to achieve phase alignment of signal images xi(t), estimation of signal magnitude estimates Mi, determination of weights Wi, or normalization. Additionally, functions may be performed using any suitable logic comprising software, hardware, other logic, or any suitable combination of the preceding.

[0058] FIG. 4 is a block diagram illustrating one embodiment of reference constructor 72 of system 10 of FIG. 1. Reference constructor 72 combines adjusted signal images Rxi(t) in accordance with magnitude estimates Mi to generate a correlation reference y′(t).

[0059] Reference constructor 72 includes branch reference constructors 310a-b, a composite reference constructor 312, and a selector 314 coupled as shown in FIG. 5. Branch reference constructor 310a generates a correlation reference from the signals received at frequency branch 22a, and branch reference constructor 310b generates a correlation reference from frequency branch 22b.

[0060] According to one embodiment, branch reference constructor 310a includes adders 320a-b and 322, an amplitude normalizer 324, and multipliers 326a-b. Adders 320a-b add adjusted signal images Rxi(t) generated from signals at receiver channels 50a-b. Adder 322 adds magnitude estimates Mi from receiver channels 50a-b. Amplitude normalizer 324 normalizes the sum of the magnitude estimates Mi. Multipliers 326a-b multiply the normalized sum of the magnitude estimates Mi with the sum of the adjusted signal images Rxi(t) from receiver channels 50a-b, respectively, to yield correlation references for receiver channels 50a-b, respectively. Branch reference constructor 310b may be substantially similar to branch reference constructor 310a.

[0061] Composite reference constructor 312 generates a composite correlation reference from signals received at frequency branches 22a-b. Composite reference constructor 312 includes adders 330 and 322a-b, an amplitude normalizer 334, and multipliers 336a-b. Adder 330 adds the sum of the signal magnitude estimates Mi from receiver channels 50a-b with the sum of the signal magnitude estimates Mi from receiver channels 50c-d to yield a combined signal magnitude estimate.

[0062] Adders 322a-b add the sum of the adjusted signal images Rxi(t) from receiver channels 50a-b with the sum of adjusted signal images Rxi(t) from receiver channels 50c-d to yield combined adjusted signal images. Amplitude normalizer 334 normalizes the amplitude of the combined signal magnitude estimate. Multipliers 336a-b multiply the combined adjusted signal image with the combined signal magnitude estimate.

[0063] Selector 314 selects the result of branch reference constructor 310a, branch reference constructor 310b, or composite reference constructor 312 in order to generate correlation reference y′(t). The selection may be made in accordance with the signal energies at the frequency branches 22a-b. For example, if frequency branch 22a has acquired a signal, but frequency branch 22b has not acquired a signal, the correlation reference may derived using the signals from frequency branch 22a. If frequency branches 22a-b have acquired signals, the correlation reference may derived using the signals from frequency branches 22a-b. The contribution of each frequency branch 22a-b may be determined with respect to signal-to-noise-ratios at the frequency branches 22a-b.

[0064] Modifications, additions, or omissions may be made to reference constructor 72 without departing from the scope of the invention. For example, reference constructor 72 may include more or fewer branch reference constructors 310. Additionally, functions may be performed using any suitable logic comprising software, hardware, other logic, or any suitable combination of the preceding.

[0065] FIG. 5 is a block diagram of one embodiment of coordinating modules 24 of system 10 of FIG. 1. Coordinating modules 24 monitor the correlation among signals from frequency branches 22a-b, and provide instructions to interpolators 52a-b to delay the signals in order to correlate the signals.

[0066] In general, coordinating modules 24 perform an initial correlating procedure to initially correlate the signals, and then perform a dynamic correlating procedure to perform subsequent adjustment. For purposes of illustration only, it is assumed that the samples from the signals of frequency branch 22a are used as reference samples with which the samples from the signals from the signals of frequency branch 22b are to be correlated.

[0067] Coordinating modules 24 include symbol aligner 60, correlator 62, and correlation monitor 64. Symbol aligner 60 includes a delay sweep control 410, an integrator 412, a multiplexer (MUX) 414, and a loop filter 416. In general, delay sweep control 410 sets a range of delays that is applied to the signals during the initial correlation procedure, and loop filter 416 dynamically adjusts the delay during the subsequent dynamic correlation procedure. Multiplexer 414 switches between delay sweep control 410 and loop filter 416 depending on the procedure.

[0068] According to one embodiment, delay sweep control 410 sets a range of delays during an initial correlating procedure by providing initial values for interpolators 52a. The delay range may have any suitable range such as plus or minus four symbols. The initial values may comprise values for pointer value τ and phase φ. Delay sweep control 410 sends a square wave through multiplexer 414 to integrator 412. In response, integrator 414 produces a triangular ramp that results in the desired delay range. The sweep is continued until the signals are correlated.

[0069] This invention is not dependent on any specific triangular shape of the sweep function, and is not dependent on any method of generating a triangular sweep function. Other embodiments of this invention may employ, for example, a sinusoidal sweep function generated by any suitable mathematical or heuristic method that causes the delay of the interpolator delay line to vary over the desired range of time offset.

[0070] Loop filter 416 dynamically adjusts the delay during a subsequent dynamic correlation procedure. The dynamic correlation procedure is initiated if correlation monitor 64 detects signal decorrelation. Correlator 62 generates correlation information comprising a symbol alignment tracking error, which is provided to symbol aligner 60. Loop filter 416 receives the correlation information and sends instructions to interpolators 52b in response to the information. The instructions may comprise a pointer value, which may be calculated according to correlation values that indicate the degree of signal correlation. Signals may be determined to be correlated if the correlation values satisfy a particular correlation threshold. A pointer value may comprise a value corresponding to the centroid of the correlation distributions satisfying the threshold, or may comprise a value corresponding to the midpoint between an initial correlation value satisfying a threshold and a final correlation value satisfying a threshold. Other criteria may be used to determine the delay associated with the best correlation.

[0071] In operation, according to the illustrated embodiment, frequency branch 22a acquires a signal. Frequency branch 22a provides samples from the signal to correlator 62 to be used as reference samples with which samples of frequency branch 22b are to be correlated. The pointer value τ and the phase φ for interpolators 52a of frequency branch 22a are set to initial values to designate a delay range.

[0072] Frequency branch 22b acquires a signal and provides samples from the signal to correlator 62. Interpolators 52b of frequency branch 22b sweep over the delay range. Correlator 62 monitors symbol alignment to generate correlation information. If the symbols align, multiplexer 414 switches control of interpolators 52b to loop filter 416. Loop filter 416 dynamically controls interpolators 52b in response to correlation information received from correlator 62.

[0073] Modifications, additions, or omissions may be made to the system without departing from the scope of the invention. Additionally, functions may be performed using any suitable logic comprising software, hardware, other logic, or any suitable combination of the preceding.

[0074] Certain embodiments of the invention may provide one or more technical advantages. A technical advantage of one embodiment may be that signals are combined in accordance with the signal-to-noise ratios associated with the signals, such that maximal ratio combining gain may be substantially realized. Accordingly, signals that exhibit multiple types of diversity such as temporal, frequency, polar, or spatial diversity may be combined.

[0075] Although the present invention has been described in several embodiments, various changes and modifications may be suggested to one skilled in the art. It is intended that the present invention encompass such changes and modifications as fall within the scope of the appended claims.

Claims

1. A method for combining a plurality of signals, comprising: receiving a plurality of signals comprising a plurality of first signals having a first frequency and a plurality of second signals having a second frequency; determining a first weight for each first signal of the first signals, the first weight reflecting a signal-to-noise ratio associated with the each first signal; generating a first signal output from the first signals in accordance with the first weights; determining a second weight for each second signal of the second signals, the second weight reflecting a signal-to-noise ratio associated with the each second signal; generating a second signal output from the second signals in accordance with the second weights; and combining the first signal output and the second signal output to yield a combined signal output.

2. The method of claim 1, further comprising: determining a symbol offset among at least two signals, each signal of the signals comprising a plurality of symbols, the symbol offset describing an offset between the symbols of the at least two signals; and correlating the at least two signals in accordance with the symbol offset.

3. The method of claim 1, wherein determining the first weight for each first signal of the first signals, the first weight reflecting the signal-to-noise ratio associated with the first signal comprises: establishing a signal magnitude estimate in accordance with a correlation reference; and determining a weight for the signal in accordance with the signal magnitude estimate.

4. The method of claim 1, wherein receiving the signals comprises: receiving the signals at a first antenna and a second antenna, the first antenna and the second antenna separated by a distance; determining an offset among at least two signals; and correlating the at least two signals in accordance with the offset.

5. The method of claim 1, wherein receiving the signals comprises: receiving the signals at a first antenna and a second antenna, the first antenna having a first polarization, the second antenna having a second polarization, the first polarization orthogonal to the second polarization; determining an offset among at least two signals; and correlating the at least two signals in accordance with the offset.

6. A system for combining a plurality of signals, comprising: a first frequency branch operable to receive a plurality of signals comprising a plurality of first signals having a first frequency; a second frequency branch operable to receive the signals comprising a plurality of second signals having a second frequency; and a sum module coupled to the first frequency branch and the second frequency branch and operable to: determine a first weight for each first signal of the first signals, the first weight reflecting a signal-to-noise ratio associated with the each first signal; generate a first signal output from the first signals in accordance with the first weights; determine a second weight for each second signal of the second signals, the second weight reflecting a signal-to-noise ratio associated with the each second signal; generate a second signal output from the second signals in accordance with the second weights; and combine the first signal output and the second signal output to yield a combined signal output.

7. The system of claim 6, wherein the first frequency branch is further operable to: determine a symbol offset among at least two signals, each signal of the signals comprising a plurality of symbols, the symbol offset describing an offset between the symbols of the at least two signals; and correlate the at least two signals in accordance with the symbol offset.

8. The system of claim 6, wherein the sum module is operable to determine the first weight for each first signal of the first signals, the first weight reflecting the signal-to-noise ratio associated with the first signal by: establishing a signal magnitude estimate in accordance with a correlation reference; and determining a weight for the signal in accordance with the signal magnitude estimate.

9. The system of claim 6, further comprising: a first antenna coupled to the first frequency branch and the second frequency branch and operable to receive the signals; and a second antenna coupled to the first frequency branch and the second frequency branch and operable to receive the signals, the first antenna and the second antenna separated by a distance, the first frequency branch operable to: determine an offset among at least two signals; and correlate the at least two signals in accordance with the offset.

10. The system of claim 6, further comprising: a first antenna coupled to the first frequency branch and the second frequency branch and operable to receive the signals; and a second antenna coupled to the first frequency branch and the second frequency branch and operable to receive the signals, the first antenna having a first polarization, the second antenna having a second polarization, the first polarization orthogonal to the second polarization, the first frequency branch operable to: determine an offset among at least two signals; and correlate the at least two signals in accordance with the offset.

11. Logic for combining a plurality of signals, the logic embodied in a medium and operable to: receive a plurality of signals comprising a plurality of first signals having a first frequency and a plurality of second signals having a second frequency; determine a first weight for each first signal of the first signals, the first weight reflecting a signal-to-noise ratio associated with the each first signal; generate a first signal output from the first signals in accordance with the first weights; determine a second weight for each second signal of the second signals, the second weight reflecting a signal-to-noise ratio associated with the each second signal; generate a second signal output from the second signals in accordance with the second weights; and combine the first signal output and the second signal output to yield a combined signal output.

12. The logic of claim 11, further operable to: determine a symbol offset among at least two signals, each signal of the signals comprising a plurality of symbols, the symbol offset describing an offset between the symbols of the at least two signals; and correlate the at least two signals in accordance with the symbol offset.

13. The logic of claim 11, operable to determine the first weight for each first signal of the first signals, the first weight reflecting the signal-to-noise ratio associated with the first signal by: establishing a signal magnitude estimate in accordance with a correlation reference; and determining a weight for the signal in accordance with the signal magnitude estimate.

14. The logic of claim 11, operable to receive the signals by: receiving the signals at a first antenna and a second antenna, the first antenna and the second antenna separated by a distance; determining an offset among at least two signals; and correlating the at least two signals in accordance with the offset.

15. The logic of claim 11, operable to receive the signals by: receiving the signals at a first antenna and a second antenna, the first antenna having a first polarization, the second antenna having a second polarization, the first polarization orthogonal to the second polarization; determining an offset among at least two signals; and correlating the at least two signals in accordance with the offset.

16. A system for combining a plurality of signals, comprising: means for receiving a plurality of signals comprising a plurality of first signals having a first frequency and a plurality of second signals having a second frequency; means for determining a first weight for each first signal of the first signals, the first weight reflecting a signal-to-noise ratio associated with the each first signal; means for generating a first signal output from the first signals in accordance with the first weights; means for determining a second weight for each second signal of the second signals, the second weight reflecting a signal-to-noise ratio associated with the each second signal; means for generating a second signal output from the second signals in accordance with the second weights; and means for combining the first signal output and the second signal output to yield a combined signal output.

17. A method for combining a plurality of signals, comprising: receiving a plurality of signals comprising a plurality of first signals having a first frequency and a plurality of second signals having a second frequency by receiving the signals at a first antenna and a second antenna, the first antenna and the second antenna separated by a distance, the first antenna having a first polarization, the second antenna having a second polarization, the first polarization orthogonal to the second polarization, by determining an offset among at least two signals, and by correlating the at least two signals in accordance with the offset; determining a symbol offset among at least two signals, each signal of the signals comprising a plurality of symbols' the symbol offset describing an offset between the symbols of the at least two signals, and correlating the at least two signals in accordance with the symbol offset; determining a first weight for each first signal of the first signals, the first weight reflecting a signal-to-noise ratio associated with the each first signal by establishing a signal magnitude estimate in accordance with a correlation reference, and determining a weight for the signal in accordance with the signal magnitude estimate; generating a first signal output from the first signals in accordance with the first weights; determining a second weight for each second signal of the second signals, the second weight reflecting a signal-to-noise ratio associated with the each second signal; generating a second signal output from the second signals in accordance with the second weights; and combining the first signal output and the second signal output to yield a combined signal output.

18. A method for correlating a plurality of signals, comprising: receiving a plurality of signals at a filter, the filter having a plurality of phases, each phase associated with a weight correlation set; receiving correlation information associated with the signals, the correlation information describing a correlation among the signals; determining a phase associated with the signals in response to the correlation information; selecting a weight correlation set in accordance with the phase; and applying the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals.

19. The method of claim 18, wherein applying the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals comprises: determining a pointer value associated with the signals in response to the correlation information; and applying the selected weight correlation set to the signals in accordance with the pointer value.

20. The method of claim 18, further comprising: receiving an initial phase and an initial pointer value corresponding to a range of delays; and applying the range of delays to the signals in accordance with the initial phase and the initial pointer value.

21. The method of claim 18, further comprising: determining an initial phase and an initial pointer value corresponding to a range of delays, the range of delays applied to the signals; monitoring the correlation among the signals through the range of delays; determining that the signals are correlated; and initiating a dynamic adjustment procedure in response to the determination.

22. The method of claim 18, further comprising: monitoring the correlation among the signals to generate a plurality of correlation values; determining that the signals are not correlated; and calculating a next pointer value in accordance with the correlation values.

23. A system for correlating a plurality of signals, comprising: a delay line operable to store a plurality of signals at a filter, the filter having a plurality of phases, each phase associated with a weight correlation set; and a processor operable to: receive correlation information associated with the signals, the correlation information describing a correlation among the signals; determine a phase associated with the signals in response to the correlation information; select a weight correlation set in accordance with the phase; and apply the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals.

24. The system of claim 23, wherein the processor is operable to apply the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals by: determining a pointer value associated with the signals in response to the correlation information; and applying the selected weight correlation set to the signals in accordance with the pointer value.

25. The system of claim 23, wherein the processor is further operable to: receive an initial phase and an initial pointer value corresponding to a range of delays; and apply the range of delays to the signals in accordance with the initial phase and the initial pointer value.

26. The system of claim 23, further comprising a coordinating module operable to: determine an initial phase and an initial pointer value corresponding to a range of delays, the range of delays applied to the signals; monitor the correlation among the signals through the range of delays; determine that the signals are correlated; and initiate a dynamic adjustment procedure in response to the determination.

27. The system of claim 23, further comprising a coordinating module operable to: monitor the correlation among the signals to generate a plurality of correlation values; determine that the signals are not correlated; and calculate a next pointer value in accordance with the correlation values.

28. Logic for correlating a plurality of signals, the logic embodied in a medium and operable to: receive a plurality of signals at a filter, the filter having a plurality of phases, each phase associated with a weight correlation set; receive correlation information associated with the signals, the correlation information describing a correlation among the signals; determine a phase associated with the signals in response to the correlation information; select a weight correlation set in accordance with the phase; and apply the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals.

29. The logic of claim 28, operable to apply the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals by: determining a pointer value associated with the signals in response to the correlation information; and applying the selected weight correlation set to the signals in accordance with the pointer value.

30. The logic of claim 28, further operable to: receive an initial phase and an initial pointer value corresponding to a range of delays; and apply the range of delays to the signals in accordance with the initial phase and the initial pointer value.

31. The logic of claim 28, further operable to: determine an initial phase and an initial pointer value corresponding to a range of delays, the range of delays applied to the signals; monitor the correlation among the signals through the range of delays; determine that the signals are correlated; and initiate a dynamic adjustment procedure in response to the determination.

32. The logic of claim 28, further operable to: monitor the correlation among the signals to generate a plurality of correlation values; determine that the signals are not correlated; and calculate a next pointer value in accordance with the correlation values.

33. A system for correlating a plurality of signals, comprising: means for receiving a plurality of signals at a filter, the filter having a plurality of phases, each phase associated with a weight correlation set; means for receiving correlation information associated with the signals, the correlation information describing a correlation among the signals; means for determining a phase associated with the signals in response to the correlation information; means for selecting a weight correlation set in accordance with the phase; and means for applying the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals.

34. A system for correlating a plurality of signals, comprising: an interpolator operable to: receive a plurality of signals at a filter, the filter having a plurality of phases, each phase associated with a weight correlation set; receive an initial phase and an initial pointer value corresponding to a range of delays, and applying the range of delays to the signals in accordance with the initial phase and the initial pointer value; receive correlation information associated with the signals, the correlation information describing a correlation among the signals; determine a phase associated with the signals in response to the correlation information; select a weight correlation set in accordance with the phase; and apply the selected weight correlation set to the signals to delay the signals in order to substantially correlate the signals by determining a pointer value associated with the signals in response to the correlation information, and by applying the selected weight correlation set to the signals in accordance with the pointer value; and a coordinating module coupled to the interpolator and operable to: determine the initial phase and the initial pointer value corresponding to the range of delays, monitor the correlation among the signals through the range of delays, determine that the signals are correlated, and initiate a dynamic adjustment procedure in response to the determination; and monitor the correlation among the signals to generate a plurality of correlation values, determine that the signals are not correlated, and calculate a next pointer value in accordance with the correlation values.

35. A method for generating a correlation reference, comprising: receiving a plurality of first signals having a first frequency; receiving a plurality of second signals having a second frequency; calculating a first correlation reference from the first signals; calculating a second correlation reference from the second signals; calculating a composite correlation reference from the first signals and the second signals; and generating a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference.

36. The method of claim 35, wherein calculating the first correlation reference from the first signals comprises: estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the signal magnitude estimates.

37. The method of claim 35, wherein calculating the first correlation reference from the first signals comprises: adjusting a phase alignment of the first signals to yield a plurality of adjusted signals; estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the adjusted signals and the signal magnitude estimates.

38. The method of claim 35, wherein calculating the first correlation reference from the first signals comprises: determining a signal energy associated with each first signal of the first signals; determining a contribution of each first signal in accordance with the determined signal energies; and calculating the first correlation reference in accordance with the determined contributions.

39. The method of claim 35, wherein generating the resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference comprises: determining a first signal energy associated with the first correlation reference; determining a second signal energy associated with the second correlation reference; and selecting the resulting correlation reference from the first correlation reference, the second correlation reference, and the composite correlation reference in accordance with the first signal energy and the second signal energy.

40. A system for generating a correlation reference, comprising: a first branch reference constructor operable to calculate a first correlation reference from a plurality of first signals having a first frequency; a second branch reference constructor operable to calculate a second correlation reference from a plurality of second signals having a second frequency; and a composite reference constructor operable to calculate a composite correlation reference from the first signals and the second signals; and a selector coupled to the first branch reference constructor, the second branch reference constructor, and the composite reference constructor and operable to generate a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference.

41. The system of claim 40, wherein the first branch reference constructor is operable to calculate the first correlation reference from the first signals by: estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the signal magnitude estimates.

42. The system of claim 40, wherein the first branch reference constructor is operable to calculate the first correlation reference from the first signals by: adjusting a phase alignment of the first signals to yield a plurality of adjusted signals; estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the adjusted signals and the signal magnitude estimates.

43. The system of claim 40, wherein the first branch reference constructor is operable to calculate the first correlation reference from the first signals by: determining a signal energy associated with each first signal of the first signals; determining a contribution of each first signal in accordance with the determined signal energies; and calculating the first correlation reference in accordance with the determined contributions.

44. The system of claim 40, wherein the selector is operable to generate a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference by: determining a first signal energy associated with the first correlation reference; determining a second signal energy associated with the second correlation reference; and selecting the resulting correlation reference from the first correlation reference, the second correlation reference, and the composite correlation reference in accordance with the first signal energy and the second signal energy.

45. Logic for generating a correlation reference, the logic embodied in a medium and operable to: receive a plurality of first signals having a first frequency; receive a plurality of second signals having a second frequency; calculate a first correlation reference from the first signals; calculate a second correlation reference from the second signals; calculate a composite correlation reference from the first signals and the second signals; and generate a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference.

46. The logic of claim 45, operable to calculate the first correlation reference from the first signals by: estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the signal magnitude estimates.

47. The logic of claim 45, operable to calculate the first correlation reference from the first signals by: adjusting a phase alignment of the first signals to yield a plurality of adjusted signals; estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal; and calculating the first correlation reference according to the adjusted signals and the signal magnitude estimates.

48. The logic of claim 45, operable to calculate the first correlation reference from the first signals by: determining a signal energy associated with each first signal of the first signals; determining a contribution of each first signal in accordance with the determined signal energies; and calculating the first correlation reference in accordance with the determined contributions.

49. The logic of claim 45, operable to generate the resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference by: determining a first signal energy associated with the first correlation reference; determining a second signal energy associated with the second correlation reference; and selecting the resulting correlation reference from the first correlation reference, the second correlation reference, and the composite correlation reference in accordance with the first signal energy and the second signal energy.

50. A system for generating a correlation reference, comprising: means for receiving a plurality of first signals having a first frequency; means for receiving a plurality of second signals having a second frequency; means for calculating a first correlation reference from the first signals; means for calculating a second correlation reference from the second signals; means for calculating a composite correlation reference from the first signals and the second signals; and means for generating a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference.

51. A method for generating a correlation reference, comprising: receiving a plurality of first signals having a first frequency, and receiving a plurality of second signals having a second frequency; calculating a first correlation reference from the first signals by: adjusting a phase alignment of the first signals to yield a plurality of adjusted signals, estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal, and calculating the first correlation reference according to the adjusted signals and the signal magnitude estimates; estimating a plurality of signal magnitude estimates according to the first signals, each signal magnitude estimate associated with a first signal, and calculating the first correlation reference according to the signal magnitude estimates; and determining a signal energy associated with each first signal of the first signals, determining a contribution of each first signal in accordance with the determined signal energies, and calculating the first correlation reference in accordance with the determined contributions; calculating a second correlation reference from the second signals; calculating a composite correlation reference from the first signals and the second signals; and generating a resulting correlation reference in accordance with a correlation reference selected from the set consisting of the first correlation reference, the second correlation reference, and the composite correlation reference by determining a first signal energy associated with the first correlation reference, by determining a second signal energy associated with the second correlation reference, and by selecting the resulting correlation reference from the first correlation reference, the second correlation reference, and the composite correlation reference in accordance with the first signal energy and the second signal energy.