GB2181548A - Pulse detection using correlation - Google Patents

Abstract

A detector for detecting a communication pulse in the presence of noise is disclosed having a transmitter for transmitting a communication pulse 103 of carrier frequency having a main pulse portion 104 and a pre-pulse portion 105, the pre-pulse portion being of substantially opposite phase to the main pulse portion, a receiver for receiving the communication pulse, a correlation signal circuit connected to the receiver for providing a correlation signal, the correlation signal having first (108) and second (109) slopes in response to the communication pulse, the second slope being steeper than the first slope, and a comparator connected to the correlation signal circuit for providing a pulse detection output signal when the correlation signal reaches a predetermined threshold, the threshold being set at at level to detect the correlation signal at a point on the second slope. <IMAGE>

Description

SPECIFICATION Enhanced pulse time-of-arrival detector The present invention relates to enhanced detection of pulses in the presence of noise and spurious signals.

Signal pulses of known carrier frequency are used in a variety of situations for com m unication, object detection, and object location, as well as for other purposes. To the extent that the signal pulses remain relatively undistorted and for significantly higher amplitudes than background noise, pulse detection can be accomplished with relative ease and simplicity. However, there are many situations in which it is necessary to reliably detect communication pulses contaminated with random noise of significant amplitude relative to the amplitude of the communication pulse. In such situations, simple amplitude and/or frequency discrimination may not provide reliable pulse detection, and other more sophisticated detection techniques are required.

One such technique is described in United States Patent 4,038,540 issued to James L. Roberts and assigned to the same assignee as is the present in vention. This patent discloses an apparatusfordet- ecting a communication pulse of known carrierfre quency,which pulse may be distorted by random noise, by correlating the received communication pulse with a reference signal having a repetition rate equal to the carrier frequency. Correlation is accomplished by multiplying the receigved communication pulse with each of quadrature components of the reference signal and sampling the product signal to provide binary sample pulse trains which are supplied to a pair of shift register accumulators.Reversible counters and associated logic determine the differences in the numbers of samples of noncorresponding binary values in each oftheshiftregisters, and provide counts which are added to provide a moving correlation signal which indicates presence of a communication pulse when it exceeds a predetermined threshold.

The detection circuitry of prior art systems such as that shown in United States Patent No. 4 038 540, have a time-of-arrival uncertainty, the extent of which is determined bytheamountthatthecom- munication pulse is distorted by noise.

According to the present invention a quadrature correlation pulse detector for detecting a communication pulse in the presence of noise comprises: transmitting meansfortransmitting a communication pulse of carrier frequency having a main pulse portion and a prepulse portion, said prepulse portion being ofsubstantially opposite phase to said main pulse portion; receiving means for receiving said communication pulse; quadrature reference signal source meansforproducing first and second quadrature reference signals; multiplier means connected to said receiving means and to said quadrature reference signal source means for multiplying said received communication pulse and said first and second quadrature reference signals, said multiplier means providing a multiplier output signal; clock means for supplying clock pulses;; correlation signal means connected to said multiplier means and including accumulator means, said accumulator means responsive to said clock pulses to shiftthrough said multiplier output signal and hav ing fewerstorage locationsthan would be required to store all of a signal as long as said communication pulse, said correlation signal means providing a correlation signal having an increased slope at the point where said prepulse portion begins shifting out of said accumulator means; and comparator means connected to said correlation signal means for providing a pulse detection output signal when said correlation signal reaches a predetermined threshold, said threshold being set at a level to detect said correlation signal at a point on said increased slope.

The present invention reduces the aforesaid timeof arrival uncertainty by providing a detector for detecting a communication pulse in the presence of noise having a transmitterfortransmitting a communication pulse of carrier frequency, the communication pulse having a main pulse portion and a pre-pulse portion, the pre-pulse portion being of substantially opposite phase to the main pulse portion, a receiver for receiving the communication pulse, a correlation signal circuit connected to the receiver for providing a correlation signal, the correlation signal having first and second slopes in responseto the communication pulse, the second slope being steeper than the first slope, and a comparator connected to the correlation signal circuit for providing a pulse detection output signal when the correlation signal reaches a predetermined threshold,the threshold being set at a level to detect the correlation signal at a point on the second slope.

Arrangements according to the invention will now be described by way of example and with reference to the accompanying drawings, in which: Figure 1A, iB and iCshow a schematic diagram of a detector according to the present invention; Figure2 shows a pulse such as that which is transmitted by the prior art system as shown in United States Patent No. 4 038 540; Figure 3 shows the signal of Figure 2 after it has been filtered and clipped by the prior art system; Figures4A and4Bshowthe correlation signal and the threshold signal provided bythe priorartsystem as a result ofthe pulse shown in Figure 2; Figures5A and5B show the uncertainty region of the correlation signal produced by the prior art system;; Figures 6and7compare the timing uncertainty at the detection threshold of a correlation signal having the slope which is obtained from the prior art system and for a correlation signal having an increased slope; Figures 8A and8B show how a correlation signal of increased slope at the point of detection can be obtained with the present invention; Figures 9, lOA and 70B show correlation signal diagrams useful in explaining howthetime-or-arrival can be more accurately determined; Figures 1 lA and 1 78 show the correlation signal generated in response to a communication pulse according to another aspect of the present invention;; Figures 12A and 12B show the threshold and detection pulses derived from the correlation signal shown in Figure 11 B; Figures 13A-l3Dshowhowthe prior art detection system respondsto direct and multi-path pulses; and, Figures 14a- show how the present invention responds to direct and multi-path pulses.

In Figure 1A, receiver 11 provides a clipped input signal S(t) to the inputs 13 and 16 of respective digital multipliers 12 and 15. Second inputs 14and 17 of digital multipliers 12 and 15 respectivelyaresupplied by quadrature reference signal source 18. Source 18 produces first and second substantially square wave signal coseoO(t) and sinwo(t) .Asapparentfrom the signal functions, these square wave signals are 90 degrees out of phase or in quadrature with one another. They also have a repetition rate equal to the carrier frequency ofthe signal pulse, and upper and lower values of 1 and 0. These signals may be considered quadrature components of a square wave reference signal.

Digital multipliers 12 and 15 may be respective EX CLUSIVE OR gates which will produce a low output onlywhen both inputs have corresponding amplitudes, i.e. both 1 or both 0.

The outputs oaf digital multipliers 12 and 15 are supplied to sample gates 20 and 23 through respective inputterminals 21 and 24. Sample gates 20 and 23, under control of clock pulses supplied by clock 25 through frequency divider 26, sample the outputs of digital multipliers 12 and 15 at a sample ratewhich is adequate to retain the desired information in the communication pulses, and produce first and second sample trains attheir respective output terminals.

Clock 25 and frequency divider 26 synchronizethe operations of the various portions of the pulsedetec- tor.

The sample trains, which comprise uniformly spa ced short duration pulses having values of either 1 of 0, are supplied to gates 28 and 31 through corresponding inputs 29 and 32. One-zero generator 35, driven by clock pulses from frequency divider 26, provide a second input to each of the gates 28 and 31 through corresponding second inputterminals 30 and 33. Control line 34to each of gates 28 and 31 controls which of the two inputs of each gate is connected to the respective gate output.

By way of example, it may be assumed that gates 28 and 31 each operate such that, in the absence of a signal from line34,thesampletrainatitsfirstinput terminal will be transmitted to its outputterminal.

Conversely, if a signal is present at its control terminal connected toline 34, the signal at its second input (i.e. from one-zero generator 35) will be transmittedto its output terminal. During the actual pulse detection process, no signal is carried on conductors 34, thereby causing the first and second sample trains to be produced at the output terminals of gates 28and31 respectively. The pulsesfrom one-zero generator 35 are connected through gates 28 and 31 during an initialization operation ofthe pulse detector.

The output signals from gates 28 and 31 are supplied to moving window accumulators 36 and 37 respectively. The moving window accumulators, which may comprise shift registers, each having a predetermined number of storage locations, are under the control of the timing signal on conductor 27 and operate to accept the output signals from gates 28 and 31, and shiftthe signal serially to the accumulator outputterminals. Accumulators 36 and 37 thus serve to define a window of fixed duration, typically less than the communication pulse, which duration is determined by the number of storage locationstherein and the rate at which signals are shift ted therethrough.

Accumulators 36 and 37 have associated therewith logic which produce a signal when the samplles entering and leaving the accumulator have different values. Stated otherwise, the logic associated with each shift register produces a signal only when the relative number of 1's and 0's in the shift register is changing. It may further be observed that a change in the relative number of 1's and 0's in each shift register is indicative of a change in the degree of correlation between the received communication pulse and one ofthe quadrature components of the square wave reference signal.

The output signals of accumulators 36 and 37 comprise trains of samplers having values of either 1 of 0.

These samples are supplied to reversible counters 38 and 39 respectively, each ofwhich is controlled by the signal on control line 34 and each of which receives clock pulses from divider 26. Counters 38 and 39 produce counts which are incremented or decremented by data signals received from accumulators 36 and 37. Accordingly, the magnitude ofthe count in counters 38 and 39 are respective ofthe difference in numbers of samples in accumulators 36 and 37 having 1 and 0 values. These counts are indicative of the degree of correlation between there ceived communication pulse and the quadrature components ofthe reference signal.

With reference to the operation of accumulator 36 and 37, it may be observed that noise will correlate with the reference signal in a random manner resulting in sampletrains of l's and 0's in random order.

Accordingly, the differences in the numbers of 1's and 0's in the accumulator may vary slightly with time but, on the average, there will be no difference in the number of l's and 0's in the accumulators due to noise. Consequently, counters 38 and 39 will respond to noise by maintaining an average count of 0.

However, as the outputs of gates 28 and 31 contain components which relate more closely with the reference signal (i.e. as might be expected upon receipt of a communication pulse), a predominate number of 1's or 0's will be present in each accu mu lator. This predominance will be reflected in the counts in the counters.

The output counts from counters 38 and 39 are supplied to adder 40 which adds the counts to pro duce a moving correlation between the input signal and the reference signal. The moving correlation signal is supplied to comparator 41 through inputter- minal 42. Comparator41 alsohasasecondinputter- minal 43 which is supplied with a reference signal indicative of a correlation threshold representing a minimum degree of correlation chosen to characterize presence of a communication pulse. Comparator41 compares the signals at inputterminals 42 and 43 and produces an output signal when the moving correlation signal exceeds the threshold reference signal. The moving correlation signal is also supplied to a delay register44through an inputterminal 45.Delay register 44 has a control terminal 46, a clockterminal, and an output terminal. Register44 receives clock pulse through conductor 27, and is op erable to delay by one clock pulse the signal received through input terminal 45 provided an enabling signal is present at control terminal 46. If the enabiing signal is present at control terminal 46, the delay signal appears at the output terminal.

The undelayed moving correlation signal and the delayed moving correlation signal as produced by delay register 44 are supplied to a second comparator 47 through input terminals 48 and 49 respectively.

Comparator 47 has an output terminal at which a signal is produced indicating whether the undelayed moving correlation signal is largerorsmallerthan the delayed moving correlation, thus detecting peaks in the correlation signal.

Reference numeral 50 identifies inhibit logic having an input terminal 51 connected to receive the outputsignal of comparator 41, an inputterminal 52 connected to receive the output signal of comparator 47, an input terminal 53 connected to receive a pulse window signal, and an output terminal which is connected to supply a signal to control terminal 46 of delay register 44. Inhibit logic 50 functions to provide an output signal which will disable register 44 only during presence of a pulse window signal at input terminal 53 and only when the signals supplied by comparators 41 and 47 respectively indicatethatthe moving correlation signal exceeds the correlation threshold signal and the undelayed moving correlation signal is smallerthan the delayed moving correlation signal.Accordingly, register 44 is caused to detect and hold the maximum correlation signal which occurs during a pulse detection window.

Detection logic 54 operates such that after the maximum correlation during a pulse detection window has occurred and a moving correlation signal has fallen belowthethreshold reference, in effect indicating that the input signal pulse is passing out of the pulse detection window, a pulse detection signal is produced at output 57. This pulse detection signal is supplied over conductor 58 to any suitable utilization apparatus. It is also supplied to a pulse timer60 through an input terminal 62. Timer 60 also has an input terminal 61 connected to receive timing signals from frequency divider 26.

Timer 60 operates so as to produce a pulse window signal on conductor63. The pulse window signal has a duration equal to the duration ofthe pulse detection window, and commences a predetermined length of time after the last detected pulse as indicated by a pulse detection signal at input terminal 62.

Conductor 63 supplies the pulse window signal to in- hibit logic 50 and to logic circuitry 64 designated as clear correlator logic.

Logic circuitry 64 is operable to supply signals to gates 28 and 31 and counters 38 and 39. It operates such that, following detection of an input signal pulse, a signal is supplied to gates 28 and 31 ofsuf- ficient duration to cause accumulators 36 and 37 to be filled with alternate 1 and 0 samples through operation of one-zero generator 35. The signal from logic circuitry 64 is also supplied to the reset terminals of counters 38 and 39, whereby the counts therein are set toO. This condition, in addition to the existence of identical numbers of 1 's and 0's in accumulators 36 and 37, initializes the pulse detection so that it is immediately ready to sense correlation between the input signal and the reference signal.

The pulse as described in United States patent 4,038,540, takes the form as shown in Figure 2. The regular sine wave as shown in Figure 2 represents the communication pulse which was transmitted and which it is desired to receive, and the irregularsignal on each side ofthe regular sine wave represents noise. The filtering and clipping circuit convertsthe signal in Figure 2 into a square wave as shown in Figure 3. The correlation signal which is supplied as an outputfrom adder40 is shown in Figure 4A and the output of the detector is shown in Figure 4B. The detection threshold as shown in Figure 4A is established by the reference input 43 to comparator41.

As described above, because noise occurs in a random manner resulting in random l's and 0's entering accumulators 36 and 37, on average there will be no net difference in the signals accumulated by the ac cumulators. Consequently, the counters will respond to noise by maintaining an average countsubstanti allyequal to 0. However, asthe communication pulse as shown by the regular sine wave in Figure 2 is received, there will be significant correlation between the received communication pulse and the quadrature reference signals so that a predominant number of l's and 0's will be present in each accumulator causing the correlation signal as shown in Figure 4Ato increase.Asthe communication pulse fillsthe accumulator and begins to be shifted out,the correlation signal will decrease.

Fig ure 5A shows the communication pulse as received by the detector and Figure 5B shows an expanded view of the leading edge of the correlation signal produced bythe detector in response to the communication pulse. Figures 5Aand SBtaken together show the timing relationship between the received communication pulse 101 and the correlation signal 102 with respectto the timing window.

Noise and creates a region of uncertainty, shown by the shaded area, such that the actual correlation signal will fall somewhere within the region.

A characteristic of finite time window correlation processes is that the leading edge ofthe correlation peak becomes progressively more stable, i.e. less subject to uncertainty, as a signal with the proper characteristics moves into the window. In Figure 5B, the shaded region,which represents uncertainty, is narrowest at the top of the leading edge.

Function 102 as shown in Figure 5B represents a correlation signal based upon a communication pulsewhich is received cleanly, i.e.without noise.

Detection must be made before the window is totally filled with signals since a detection level set too closelytothe maximum achievable level of the pulse will introduce a higher probability of missing the pulse altogether. The uncertainty associated with the time of detection is determined bythewidth ofthe shaded uncertainty region at the point of detection.

Figure 6 is an enlarged view of the correlation signal and itsshaded area of uncertaintyatthepoint ofuncertaintyatthe point of detection. It can be seen from Figure 6 that, for a given amplitude uncertainty, there is an associated timing uncertainty as determined by the slope of the leading edge, i.e. the slope ofthe correlation signal.

Figure 7 shows that with increased slope and the same amplitude uncertainty,the timing uncertainty is reduced. There are two ways of increasing the slope ofthe correlation signal without impacting re ceiver complexity--by shortening the integration window or by adjusting the signal characteristicto produce a steeper slope atthe detection threshold point ofthe correlation signal.

Thefirstway,which involves shortening the pulse duration and the integration time of the associated receiver, is the usual method of improving timing resolution. Unfortunately, this way has the negative ef fectofwidening the receivers noise bandwidth. In the sonar beacon signal detection application, in which the present invention may be used, for example, widening the receiver's noise bandwidth re ducesthe maximum detection range for a given beacon source level. Hence, this particularwayis not a desirable alternative for long-range/deep water applications.

The second way is normally accomplished by increasing the complexity of the signal characteristic.

In doing so, the bandwidth is increased without shortening the communication pulse. This result is normally referred to as increasing the time bandwidth product. Unfortunately, this method usually also requires an undesirable and impractical increase in the complexity of the receiver process, in that the receiver can no longer look for a simple com munication pulse. A matched filter receiverfora complex signal requires a full replica correlation process which may be on the order of twice as complex.

Figure 8A illustrates how a tone pulse can be modified according to the present invention to improve the timing uncertainty at the detection threshold of the correlation signal to improvethetime-of-arrival detection performance of a communication pulse receiver. The communication pulse is provided with a main pulse portion 104which maytake the form as shown in Figure 2 and a pre-pulse portion 105which is of a phase substantially opposite to the phase of the main pulse portion 104. Such a signal may be transmitted by transmitter 10 shown in Figure 1.

Transmitter 10 can be any diphase transmitter arran get to provide the propertiming length ratios between pre-pulse portion 105, main pulse portion 104, and the timing window established by accumulators 36 and 37. Figure 8A shows an example of the time duration ratio between main pulse portion 104 and prepulse portion 105, and Figures 8A and 88 taken togethershowthe relationship between the length of communication pulse 103 and the size ofthe integration window as established by the clockfrequency and the number of storage locations in accumulators 36and37.

Figure 8B is a trace showing that as the pre-pulse portion of communication pulse 103 enters the cor- relator's integration window, i.e. enters the accumulators, the expected output level ofthe correlation signal provided by adder 40 increases along segment 106. This increase continues until the phase reversal edge enters the window. Atthis point, the trace decreases along segment 107 because the main pulse portion 104 starts cancelling pre-pulse portion 105, which pre-pulse portion is already in the integration window.The magnitude of this pre-peak occurring at the end segment 108 is limited to approximately 25 percent offull scale by limiting the time duration of the pre-pulse to 25 percent ofthe integration window width. The down trend of segment 107, which has a slope equal in magnitude but opposite in sign to the slope of segment 106, continues until it reaches 0, the point at which equal amounts of pre-pulse and main pulse are in the window.

Beyond this point, the correlation output signal 115 increases along segment 108, which has the same slope as segment 106, until the window is totally filled with signal (all ofthe pre-pulse and some ofthe main pulse). At this point, the noise-produced output level uncertainty is minimum because the window is filled with signal as it can get.

After segment 1 OS the uncertainty stays minimized because the window stays full of signal. However, the absolute level continues to change along segment 109, but at twice the previous rate. This change in rate occurs because as more of the main pulse of consistent phase is added to the window, a corresponding amountof pre-pulse of cancelling phase is lost, thereby doubling the effect of rate of change ofthe integration window. As the level passes through the detection threshold, atime-of-arrival measurement is madewhich gainsthe benefits ofthe following two effects: (1) the integration window is totally filled with signal so that noise uncertainty is minimized; and (2) the slope ofthe leading edge is doubled so that the noise uncertainty is halved.

When all of the prepulse has been shifted out of the accumulators, the correlation signal follows segment 110 until no more communication pulse enters the accumulators at which time the correlation signal begins dropping along segment 111. The slope of segments 106, 107, and 108 may be referred to as the first slope and the slope of segment 109 may be referred to as the second slope.

In another aspect of the invention, an unbiased measurement of the pulse time-of-arrival is derived by determining the midpoint between the leading and trailing edge detection points, as indicated in Figure 9. Under noiseless conditions, the receiver output for a simple communication pulse input, such as that shown in Figure 2, will have the idealized waveform 121 shown in Figure 9. As the signal4o- noise ratio decreases, the effect of the expected (average) output will beto reduce the peak level and the leading and trailing edge slopes as indicated by waveform 122. The uncertainty band will be centered aboutthis expected wave form 122. The expected threshold crossing of the leading slope edge occurs laterthan in the noiseless case and is signal-to-noise ratio dependent.This bias develops on the trailing edge as well and, on the average, is equal in magnitude but opposite in polarity to the leading edge bias such that the amount of uncertainty can be reduced by determining the midpoint between the positive and negative going intersections ofthe correlation signal with the detection threshold. Another benefit from determining pulse arrival from both the leading and trailing edge measurements is that the combined measurements, given that they contain independent noisejittercomponents, result in a combined measurement noise jitter which is approximately 30 percent lowerthan the noise on either individual measurement.

Figure 1 Bshowsthe circuitryfordetermining the midpoint ofthe received communication pulse. As the signal DETECT goes low indicating the beginning of a received communication pulse, flip-flop 201 is cleared which drives its inverted Q output high and flip-flop 202 is set which drives its Q output high.

When flip-flop 202 switches, the clear signal is removed from flip-flop 203 ailowing itto be clocked to cause counter 204 to begin counting at half rate.

Atthesametime,theTHLDsignal goes high which signal is delayed by flip-flop 205 for synchronization purposes and then operates through AND gate 206, NAND gate 207, andflip-flop 208to generate atrail ing edgesignal on line 209. Ifthetreilingedgeofthe communication pulse arrives before counter 204 reaches a count of 16, the detector is reset on the assumption that a spurious pulse was received. thus, a holding period is built into the detectorto insure that the detectorwill not respond to spurious signals.

However, if the counter reaches a count of 16 before the trailing edge signal on line 209 isgenerated,the counter continues to count at the half rate until the trailing edge signal appears on line 209.

When this signal appears afterthe count of 16,flipflop 201 is setto 0 on its inverted outputwhich disablesthe divide-by-two operation offlip4lop 203 allowing the counter to count at full rate. Counter 204 willcontinuetocountatfull rate until it reaches a full countatwhich pointflip-flop 210 is switchedto gene- rate the signal RTU STROBE.

At the beginning of each cycle of operation, the counter is preset to a count of 2584 and then counts up from there. Counter 204 and the clock frequency are arranged so that the counter takes 1 ms to reach a full count of 4095 from the preset count of 3584 coun ting atthefull rate and takes 2 msto reach afull count of 4095 (all ones, in binary) counting at the half rate.

Bythis arrangement and by counting at half rate until the detection ofthe trailing edge after which the countercountsatfull rateuntilthefullcountis reached, the signal RTU STROBE will always be generated 1 ms after the midpoint of the receivged communication pulse. The midpoint can, therefore, be accurately determined bythe detector.

As an example, if the signal THLD goes high attime 0 and stays high for 1 ms indicating a received communication pulse of 1 ms duration, counter 204 will count at half rate for as long as the signal THLD re mains high, i.e. 1 ms. Atthis point the counter is only half full since it has counted at half rate for only 1 ms.

When the signal THLD goes low (at the end of 1 ms), the counter will count at its full rate until the counter becomes full, a process which will take 1/2 ms since it is now counting at full rate and since it was only half full. The midpoint of the received communication pulse occurs 1 ms before the point at which the center becomes full (1/2 ms during which the counter counts at itsfull rate and 1/2ofthetimethatthecoun- ter counts at its half rate).

Figure 1 C shows the modifications of the circuit 208 shown in U.S.patent 4,038,540 to generatethe DETECTsignaI and from wheretheTHLDsignal can be derived, which signals are used to detect the midpoint of the received communication pulse.

This time-of-arrival determination can be further enhanced by using a communication pulse with both a pre-pulse portion and a post-pulse portion so that the sown trend segment (111 as shown in Figure8B) has a slope equal in magnitudetotheslopeofthe increased slope segment (109 as shown in Figure 8B) but opposite in sign and by taking a time average of where the correlation signal crosses the detection threshold both in the positive going sense and in the negative going sense.

By providing both pre and post pulses to the main pulse shown in Figure 8A, this time-of-arrival technique can be made more accurate.

Accordingly, Figure 1 OA shows the standard tone pulse as shown in Figure 2. Figure lOB shows the correlation signal which is produced by the receiver shown in Figure 1 in response to the communication pulse shown in Figure 10A. Figure 1 lAshowsacom- munication pulse 130 having a pre-pulse portion 131 and a post-pulse portion 132. Figure 11 B shows the correlation signal which results from the use of communication pulse 130. By comparing Figures 11 B and 8B, it can be seen that the first half of the correlation signal is the same duetothe pre-pulse portion 131.

The other half of the correlation signal shown in Figure 11 B, i.e. the portion of the curve following the flat portion, is the same as the first half shown in Figure 11 B except opposite in sign because of postpulse portion 132. As can be seen, the slope of the cu rve just fol lowing the flat portion is equal in magnitude but opposite in sign to the slope of the curvejust before the fiat portion.

In response to the correlation signal, the comparator produces the signal shown in Figure 1 which is then averaged to produced the signal of Figure 1 2B.

Thus, timing noise jitter can be halved because the slopes of both the leading and trailing edges are doubled, timing noise jitter can be reduced because both leading and trailing edge detection time meas urementsto estimatethe mid-pulse point are utilized, and signal-to-noise dependent timing bias error is cancelled by taking advantage of the mirror imagesymmetryofthe leading and trailing edgedetection bias error components.

Furthermore, an added benefit derived from the positive control ofthetrailing edge ofthedetected pulse is that of resistance to multipath elongation which can be produced from multipath reflections of the primary communication pulse. Figure 13A shows the primary communication pulse and Figure 13B shows a reflected multipath pulse. These pulses at the receiver appear, as is shown in Figure 13C, as the sum of the direct pulse and a multi-path pulse. Figure 13B shows the resulting correlation pulse which can result in two detection pulses for one communication pulse ratherthan one detection pulse for one communication pulse.

Figure 14Ashows a communication pulse with both a pre-pulse portion and a post-pulse portion.

Figure 14B shows a multipath pulse which is a re flectionofthe primary direct path pulse. Figure 1 4C shows the resulting pulses that appear at the receiver. Figure 1 4D shows the resulting correlation signal from the received pulse shown in Figure 14C.

As can be seen, the signal shown in Figure 14B will result in only one detection pulse.

Claims (22)

1. A quadrature correlation phase detectorfor detecting a communication pulse in the presence of noise comprising: transmitting means for transmitting a communication pulse of carrierfrequency having a main pulse portion and a prepulse portion, said prepulse portion being of substantially opposite phase to said main pulse portion; receiving means for receiving said communication pulse; quadrature reference signal source means for pro ducing first and second quadrature reference signals; multiplier means connected to said receiving means and to said quadrature reference signal source means for multiplying said received com munication pulse and said first and second quad rature reference signals, said multiplier means prov iding a multiplieroutputsignal; clock means for supplying clock pulses;; correlation signal means connected to said multi plier means and including accumulator means, said accumulator means responsiveto said clock pulses to shift through said multiplier output signal and hav ing fewer storage locations than would be required to store all of a signal as long as said communication pulse, said correlation signal means providing a cor relation signal having an increased slope at the point where said prepulse potion begins shifting out of said accumulator means; and, comparator means connecting to said correlation signal means for providing a pulse detection output signal when said correlation signal reaches a pred etermined threshold, said threshold being set at a level to detect said correlation signal at a point on said increased slope.

2. Thedetectorofclaim 1 whereinsaidtransmitting means transmit a communication pulse of carrierfrequency having a main pulse portion, a pre pulse portion, and a post-pulse portion, said prepulse portion and said post-pulse portion being of substantially opposite phase to said main pulse portion; said correlation signal means providing a correlation signal having increased slopes in a positive sense and in a negative sense as a result of said communication pulse, said comparator means comprising means for detecting the intersection of both the positive and negative slopes of said correlation signal with said predetermined threshold.

3. The detector of claim 2 wherein said comparator means comprises means for determining the midpoint between said intersections for determining atimeofarrivalforsaid pulse.

4. The detector of claim 3 wherein said midpoint determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting atfull rate until said counterisfull.

5. The detector of claim 4wherein said multiplier means comprises first and second multipliers, said firstmultiplierfor multiplying said first reference signal to said received communication pulse and said second multiplierfor multiplying said second reference signal, which is in phase quadrature with said first reference signal, to said communication pulse, and said accumulator means comprises a first accumulatorfor receiving a first output from said first multiplier and a second accumulatorfor receiving a second output from said second multiplier.

6. The detector of claim 5 wherein said first and second accumulators comprise respective first and second reversible counters, said first reversible counter having inputs connected to said clock means and to said first accumulator and said second reversible counter having inputs connected to said clock means and to said second accumulator, whereby said reversible counters provide count signals indicative of the degree of correlation between said first and second reference signals and said received communication pulse.

7. The detector of claim 6 wherein said correlation means comprises an adder connected to add the counts produced by said first and second reversible counters.

8. A detector for detecting a communication pulse in the presence of noise comprising: transmitting means fortransmitting a communication pulse of carrier frequency having a main pulse portion and a prepulse portion, said prepulse portion being of substantially opposite phaseto said main pulse portion; receiving means for receiving said communication pulse; clock means for supplying clock pulses;; accumulator means connected to said receiving means andto said clock meansto shiftthroughan accumulator signal, said accumulator signal based upon said received communication pulse and a reference signal, said accumulator means having fewer storage locations than would be required to store all of a signal as long as said communication pulse; correlation signal means responsive to said accumulator signal for providing a correlation signal having an increased slope at the point where said prepulse portion being shifting out of said accumulator means, said correlation signal dependent upon an amountofcorrelation between said received communication pulse and said reference signal; and, comparator means connected to said correlation signal means for providing a pulse detection output signal when said correlation signal reaches a pred etermined threshold, said threshold being set at a level to detect said correlation signal at a point on said increased slope.

9. The detector of claim 8wherein said transmitting means transmits a communication pulse of carrierfrequency having a main pulse portion, a prepulse portion, and a post-pulse portion, said prepulse portion and said post-pulse portion being of substantially opposite phase to said main pulse portion, said correlation signal means providing a correlation signal having increased slope in a positive sense and in a negative sense as a result of said communication pulse, said comparator means comprising means for detecting the intersection of both the positive and negative slopes of said correlation signal with said predetermined threshold.

10. The detector of claim 9 wherein said com- parator means comprises means for determining the midpoint between said intersections for determining atime of arrival forsaid pulse.

11. Thedetectorofclaim l0wherein said mid- point determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting at full rate until said counter is full.

12. Thedetectorofclaim 11 wherein said correla- tion means comprises reversible counter means, said reversible counter means having inputs connected to said clock means and to receive said communication pulse, whereby said reversible counter means provides a count signal indicative of said amount of correlation between said reference signal and said received communication pulse.

13. A detector for detecting a communication pulse in the presence of noise comprising: transmitting means for transmitting a communication pulse of carrierfrequency having a main pulse portion and a prepulse portion, said prepulse portion being of substantially opposite phase to said main pulse portion; receiving means for receiving said communication pulse; correlation signal means connected to said receiving means for providing a correlation signal dependent upon an amount of correlation between said received communication pulse and a reference signal, said correlation signal having first and second slopes in response to said communication pulse, said second slope being steeper than said first slope; and, comparator means connected to said correlation signal means for providing a pulse detection output signal when said correlation signal reaches a predetermined threshold, said threshold being set at a level to detect said correlation signal at a point on said second slope.

14. The detector of claim 13 wherein said trans- mitting means transmits a communication pulse of carrier frequency having a main pulse portion, a prepulse portion, and a post-pulse portion, said prepulse portion and said post-pulse portion being of substantially opposite phase to said main pulse portion, said correlation signal means providing a correlation signal having increased slopes in a positive sense and in a negative sense as a result of said communication pulse, said comparator means comprising means for detecting the intersection of both the positive and negative slopes of said correlation signal with said predetermined threshold.

15. Thedetectorofclaim 14whereinsaidcomparator means comprises means for determining the midpoint between said intersections for determining a time of arrival for said pulse.

16. Thedetectorofclaim 15whereinsaidmidpoint determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting atfull rate until said counter isfull.

17. A detectorfor determining the midpoint of a communication pulse comprising; receiving means for receiving said communication pulse; pulse determining means for generating a leading edge signal and a trailing edge signal defining said communication pulse; and, midpoint determining means for determining the midpoint of said communication pulse, said midpoint determining means including counter means having a predetermined count capacity, said counter means responsive to said leading edge for counting clock pulses from a clock at half rate and forthen counting clock pulses at full rate in response to said trailing edge until said count capacity is attained, said midpoint occurring an amount of time before said count capacity is reached dependent upon said full and half rates.

18. The detector of claim 17 wherein said midpoint determining means comprises means for resetting said detector if said trailing edge occurs before a predetermined count so that a received signal can be ignored if it is not of sufficient duration.

19. The detector of claim 1 wherein said midpoint determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting atfull rate until said counterisfull.

20. The detector of claim 8 wherein said midpoint determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting atfull rate until said counter is full.

21. The detector of claim 13 wherein said midpoint determining means comprises counter means for counting clock pulses at half rate while said correlation signal is above said threshold and, when said correlation signal falls below said threshold, for counting atfull rate until said counter is full.

22. A quadrature correlation phase detectorfor detecting a communication pulse in the presence of noise, substantially as described herein with reference to Figures 1Ato 1 C of the accompanying drawings.