Coherent Phase and Frequency Recovery Method and Circuit

The present invention relates to a method and a circuit for coherently recovering the phase and frequency of a received signal. More particularly, the invention relates to a method and a circuit for rapidly coherently recovering the phase and frequency of bursts of received signals from plural sources, which signals may differ from one another in frequency and phase.

Conventionally, coherent phase and frequency recovery has most often been performed with a PLL (Phase-Locked Loop) circuit, a block diagram of which is shown in FIG. 1. As shown in FIG. 1, the conventional PLL circuit is composed of a mixer 11, loop filter 12, and VCO (Voltage-Controlled Oscillator) 13 connected in a loop configuration. An input signal .theta. is applied to the positive input of the mixer 11, and the output of the mixer 11 is applied through the loop filter 12 to the frequency-control input terminal of the VCO 13. The output of the VCO is connected back to the negative input of the mixer 11. As is well known, when a new input signal .theta. is applied to this circuit, if the frequency and/or phase of the input signal .theta. differ from those of the output of the VCO 13 (and they are within the pull range of the circuit), the resulting output of the loop filter 12 will drive the VCO 13 in the direction to cause the output of the VCO to follow the input signal .theta. in frequency and phase.

The PLL circuit of FIG. 2 can be modeled by the nonlinear phase parameter control system shown in FIG. 2. In this model, the VCO is represented as an integrator having a transfer function K.sub.0 /s, and the mixer 11 by an adder 14 and a block having a transfer function K.sub.d .multidot.sin(s). The loop filter has a transfer function F(s).

To achieve phase lock with an incoming signal of arbitrary phase, the output phase estimate .theta. from the integrator must change to approximate the value of the phase of the input signal .theta.. When a new input signal .theta. is applied, it is represented in the model of FIG. 2 by a step change in input phase. For the VCO to instantaneously follow a step change, it must be driven by an impulse, which requires a very large bandwidth. This results in two fundamental and inherent problems.

First, for applications that employ a crystal-controlled VCO with a relatively narrow pull range, it is generally difficult, if not impossible, to prevent the VCO input from saturating at wide recovery loop bandwidths. In addition, if the input SNR (Signal-to-Noise Ratio) is not sufficiently large, the wider bandwidth will yield a poorer acquisition reliability. That is, the probability of missing acquisition and falsely detecting acquisition are limited by what is termed the "hangup" phenomenon. In general, the lower the SNR, the greater the chance hangup will occur.

Consequently, an alternate circuit arrangement that eliminates the PLL so as to avoid the hangup phenomenon while maintaining a moderate acquisition bandwidth is often needed. For this purpose, a tuned filter is sometimes used. A tuned filter, however, has the disadvantage that if there is a frequency offset in the incoming signal, a phase error in the output estimate will result.

Accordingly, it is an object of the present invention to provide a coherent phase and frequency recovery method and circuit in which the above-discussed disadvantages are eliminated.

More specifically, it is an object of the present invention to provide a coherent phase and frequency recovery method and circuit capable of producing an output signal which is rapidly locked in phase and frequency with incoming signal bursts of varying phase and frequency.

In accordance with the above and other objects of the present invention, a coherent phase and frequency recovery method and circuit are provided in which phase and frequency control signals are processed separately in a dual path feedback loop. A phase correction signal is directly injected as an output phase estimate into the phase path of the feedback loop, bypassing the integrator in the frequency path. Acquisition reliability is enhanced by presetting the initial phase and frequency of the VCO to closely match the incoming signal.

FIG. 1 is a block diagram of a conventional PLL circuit;

FIG. 2 is a diagram showing a nonlinear phase parameter model of the PLL circuit of FIG. 1;

FIG. 3 is a diagram of a model of a coherent phase and frequency recovery circuit of the present invention;

FIG. 4 is a block diagram of a phase and frequency modulator used in the coherent phase and frequency recovery circuit of FIG. 3;

FIG. 5 is a block diagram of an apparatus used for deriving initial phase and frequency estimates;

FIG. 6 is a timing diagram used to explain the operation of the apparatus of FIG. 5;

FIG. 7 is a block diagram of another apparatus for deriving initial phase and frequency estimates;

FIG. 8 is a block diagram showing an example of a phase/frequency estimator used in the apparatus of FIG. 7; and

FIG. 9 is a block diagram of an example of a carrier recovery processor circuit used in the apparatus of FIG. 7.

Preferred embodiments of the invention will now be described with reference to the attached drawings.

FIG. 3 is a diagram showing a transfer function representation of a coherent phase and frequency recovery circuit constructed in accordance with the invention.

In the coherent phase and frequency recovery circuit of FIG. 3, the input signal .theta. is differenced at 21 with the loop output signal .theta.. The difference signal is multiplied by a gain factor K.sub.d in an amplifier 22, and the resulting signal K.sub.d (.theta.-.theta.) is applied to the inputs of two loop filters 23 and 24 having respective transfer functions F.sub.0 (s) and F.sub.1 (s). The output .theta..sub.0 of the loop filter 23 is summed at 25 with an initial frequency estimate .DELTA.f. The resulting sum signal is integrated by an integrator (VCO) 27 having a transfer function K.sub.0 /s. The integrated signal, which contains the frequency information of the loop output signal .theta., is applied to the signal input of a phase-shift network 29, represented in this transfer function diagram as a summation point. Of course, in an actual circuit implementation, the summation point would be implemented as indicated in the circuit shown in FIG. 4.

A parallel loop is formed by the path composed of the loop filter 24, the summation point 26, and amplifier 28. The output .theta..sub.1 of the loop filter 24 is summed at 26 with an initial estimate .DELTA..theta..sub.1 of the phase of the input signal, the sum signal is multiplied by a factor K.sub.1 in an amplifier 28, and the resulting signal, which contains the phase information for the loop output signal .theta., is coupled to the phase-control input of the phase-shift network 29.

In the above-described circuit, the frequency control signal produced at the summation point 25 is integrated by the integrator 27 to produce a signal having the proper output frequency but with arbitrary phase. On the other hand, the phase control signal applied to the phase control input of the phase shift network sets the phase of the above signal such that the net output phase closely matches that of the incoming signal. If the phase-shift network has a very fast response time, synchronization is nearly instantaneous when the initial phase offset .DELTA..theta. is accurately predicted or measured (as will be described in more detail below).

Thus, with the coherent phase and frequency recovery circuit of the invention, the frequency and phase information for the loop output signal are derived separately and then combined to produce the latter signal. As will be explained in more detail below, this approach achieves rapid locking onto the input signal while substantially eliminating any danger of false locking.

The summation points 25 and 26, the integrator (VCO) 27, and the phase-shift network 29 together form a phase and frequency modulator, which is shown in more detail in the block diagram of FIG. 4. More specifically, the phase-shift network 29 includes a quadrature phase splitter 31 which divides the output of the VCO 27 into in-phase (0.degree.) and quadrature (90.degree.) components, which are coupled to inputs of respective mixers 32 and 33. The phase control signal drives a phase-to-amplitude mapper 35, the i and q outputs of which are applied to the second inputs of the mixers 32 and 33. The phase splitter 31, the mixers 32 and 33, and the signal summer 34 can be implemented with RF hybrids having a bandwidth of the order of several hundred megahertz. The outputs of the mixers 32 and 33 are summed at 34 to produce the loop output signal. The phase-to-amplitude mapper 35 amplitude-weights the quadrature sinusoidal outputs from the VCO 27 with cosine and sine weighting functions, as shown in Table 1 below:

In the circuit of FIG. 3, if the initial estimates .DELTA..theta. and .DELTA.f for phase and frequency are reasonably accurate, the model can be linearized with the following approximation:

Further, with appropriate selection of the parameters of the loop filters 23 and 24 (F.sub.0 (s) and F.sub.1 (s), respectively), the closed-loop response of the circuit of FIG. 3 can be made the same as that of a conventional second-order PLL. To compute the parameters of the loop filters 23 and 24, with the initial offsets .DELTA..theta. and .DELTA.f set to zero, the closed-loop transfer function is: ##EQU1##

Substituting F.sub.0 (s)=1 and F.sub.1 (s)=1/(1+s.tau..sub.1) into equation (2) gives: ##EQU2##

The transfer function of a conventional second order PLL can be written as: ##EQU3##

Setting equation (3) equal to equation (4) yields the design equations in terms of the parameters of the standard second order PLL system: ##EQU4##

Using these parameters, the dual path phase and frequency loop of the present invention is precisely equivalent to the conventional PLL with respect to the steady state. This is advantageous in that the conventional PLL can be demonstrated to be the optimum minimum mean-square error tracking configuration for an unmodulated carrier. In addition, for acquisition of an incoming signal burst, the circuit of the invention exhibits a nearly instantaneous response due to the injection of the initial phase offset into the output estimate.

The initial phase and frequency estimation procedure to produce the initial phase and frequency estimates .theta..sub.1 and .theta..sub.0, respectively, will now be described.

There are two methods available for determining the initial phase and frequency offset. In accordance with the first method, the estimates are predicted in a statistically adaptive manner based on information collected from previous transmission bursts from the various input signal sources. In accordance with the second method, the initial estimates are derived at the beginning of each burst in an open-loop fashion. Also, it is possible to combine these two approaches.

A block diagram of an arrangement for determining the estimates in accordance with the first method is shown in FIG. 5. In this approach, the phase and frequency estimates, .theta..sub.1 and .theta..sub.0, respectively, from the outputs of the dual path loop filters 23 and 24 (FIG. 3) are sampled and converted to digital form by an analog-to-digital converter 51 and the samples applied to a microprocessor 52. The microprocessor 52 formulates predictions of the initial phase and frequency offsets, .DELTA..theta. and .DELTA.f, respectively, for future bursts. The predictions are converted to analog form by a digital-to-analog converter 53 and then added to their respective loop control signals as described above.

A timing diagram for this case is shown in FIG. 6. The first line in FIG. 6 shows a hypothetical series of bursts from three stations S1, S2 and S3. Station S1 is assumed to be transmitting on frequency, Station S2 below frequency, and Station S3 above frequency and drifting low. The resultant frequency control voltage .theta..sub.0 has an instantaneous level proportional to the amount of frequency offset of the incoming bursts, as shown on the second line in FIG. 6. The uncorrected phase variations .theta..sub.0 of the VCO are related to the integral of the control voltages, as depicted in the third line. Straight-line projections of the phase estimate for the next burst occurrences are also shown for the first three bursts. This information is used by the microprocessor 52 in conjunction with the current value of the local oscillator phase to predict the initial phase offset for the next burst from the same stations. Higher-order projections than linear are also possible. The resulting initial frequency and phase estimates, .DELTA.f and .DELTA..theta., respectively, are shown in the third and fourth lines of FIG. 6, where .DELTA..theta. is equal to the difference between the actual phase and the predicted phase. It should be noted that the burst timing information is required by the microprocessor 52 to store past information and to preset the initial estimates. The microprocessor 52 also should restrict the initial phase estimate to .+-.180.degree. by subtracting out modulo 360.degree. corrections. The integration constant should be appropriately selected to prevent overflow for the worst-case frequency offset.

An arrangement for implementing the second method discussed above, that is, the open-loop approach, is illustrated in FIG. 7. The baseband outputs of two mixers 61 and 62 in phase quadrature with a phase shifter 63 are applied to a carrier recovery processing circuit 65 which produces the signals for driving the dual path loop filters 23 and 24. The baseband outputs of the two mixers 61 and 62 are also applied to a phase/frequency interval estimator 64. At the end of a predetermined estimation interval, a switch 66 is closed to thereby close the loop through the loop filters 23 and 24, and the initial phase and frequency estimates are injected into the phase/frequency modulator to thus minimize the lock-up transient.

Although suitable constructions of the phase/frequency estimator 64 and the carrier recovery processing circuit 65 are known and the details thereof do not form a part of the present invention, specific examples will be briefly discussed.

FIG. 8 is a diagram showing a possible construction of the phase-frequency estimator 64. The baseband i and q inputs are each multiplied by a preamble pattern in multipliers 71 and 72 to remove preamble data modulation. The resulting signals are integrated by integrators 73 and 74 to smooth noise fluctuations contained therein and then subjected to an inverse tangent operation in a circuit 75 to yield an estimate of the carrier phase error. The preamble is divided into two estimation intervals such that two consecutive phase estimates .theta..sub.2 and .theta..sub.1 are generated. The frequency estimate .DELTA.f, computed in an arithmetic circuit 76, is proportional to the phase difference .theta..sub.2 -.theta..sub.1 divided by the time interval between .theta..sub.2 and .theta..sub.1. The phase estimate .DELTA..theta. at the end of the preamble is computed in an arithmetic circuit 77 as (3.multidot..theta..sub.2 -.theta..sub.1)/2.

An example of the carrier recovery processing circuit 65 is shown in FIG. 9. The baseband i and q signals are each split into two paths: one for data detection by data detectors 81 and 82, and the other for computing the carrier phase error for processing through the dual path loop filter 23 and 24. The post-detected signals on the outputs of the data detectors 81 and 82 are fed back and cross-multiplied with i and q by multipliers 83 and 84, and differenced by a subtractor circuit 85 to yield the carrier phase error, which is applied to the dual path loop filters 23 and 24.

This completes the description of the preferred embodiments of the invention. Although preferred embodiments have been described, it is believed that numerous modifications and alterations thereto would be apparent to one of ordinary skill in the art without departing from the spirit and scope of the invention.