• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method of single pulse tracking radar, which works with sum and difference signals to determine target direction and comprises a Doppler tracking loop for keeping an intermediate frequency target signal within the passband of a speed gate filter and a digital filter device arranged in each of the sum and difference channels , which at periodic update intervals performs analysis over a number of adjacent frequency slots of target signal components within the passband, confirming the presence or absence of a target signal within a particular frequency slot. A series of comparison procedures are performed, each affecting the sum of a predetermined number of consecutive dignity signal values from the particular frequency compartment accumulated together with any previous such sums. Comparison of this accumulated sum is performed with upper and lower threshold values, which get closer and closer to each other for each comparison procedure. Confirmation of the presence of a target signal during each comparison procedure is indicated when the upper threshold value is exceeded by the accumulated sum, while confirmation of absence of a target signal is indicated when the lower threshold value exceeds the accumulated sum. Another comparison procedure is triggered when the accumulated sum is between the upper and lower thresholds.
    公开号:SE8803569A1
    申请号:SE8803569
    申请日:1988-10-07
    公开日:2012-04-17
    发明作者:D W Joynson；P J Macbean；N Stansfield
    申请人:Alenia Marconi Systems Ltd；
    IPC主号:
    专利说明:

    The Marconi Company Limited sat on a single pulse tracking radar (Derived from p. 8206787-7). The present invention relates to a set of the kind set forth in the preamble of claim 1.
    The radar can be an active system, in which the grind is illuminated Aedelst straining from the robot and reflects straining from the grind, which is called surface echo, is received by the robot during its movement. The reflected strain is treated to obtain directions in bearing and elevation and the grinding speed or distance, so that the robot can follow the changes in the grinding direction and speed.
    The invention includes based on the task of improving the final accuracy of a pulse Doppler radar.
    According to the invention, a method can be drawn that for single pulse tracking radar, which works with sum and difference signals to determine direction and comprises a Doppler tracking loop for lining an intermediate frequency target signal with the passband for a speed gate filter and one in each of sum and difference channels. filter device, which at periodic update intervals performs analysis over a number of adjacent frequency slots of template signal components such as the passband, confirms the presence or absence of a male signal in a particular frequency slot, by performing a series of comparison procedures, each depending on the sum of a predetermined number of consecutive dignity signal values from the special frequency compartment accumulated together with any preceding such sums, and comparison of this accumulated sum with the liver and lower threshold values, which come closer to each other for each comparison procedure, thereby confirming the grain aro of a target signal during the usual comparison procedure is indicated when the upper 2 threshold value is exceeded by the accumulated sum, while confirmation of the absence of a target signal is indicated, nfir the lower threshold value exceeds the accumulated sum, and that an additional sum is exceeded, between the upper and lower threshold values.
    According to an embodiment of the invention, the method is characterized in that a digital filtering is performed for each of the sum and difference signals, which at periodic update intervals & provides analysis of a number of adjacent frequency compartments of potential mile signal components, identification is performed of a target frequency compartment. amount-! and difference signals with respect to the identified target frequency compartment as inputs to a device for forming a complex product of one of these inputs and the complex conjugate to the other, an indication of signal-to-noise derivation being derived from the imaginary component of the complex product and a indication of the dignity level of the sum channel signal in relation to the target frequency range, so that this indication has a Mgt value in the occurrence of non-coherent reflections from multiple targets and a low value in the event of coherent reflections from a single target, as well as a basic indication of its signal The target frequency compartment and the average dignity level across all frequency compartments, this basic indication having a high value in the occurrence of single or several targets in the single frequency compartment and the high density in the presence of broadband noise, and a device which, depending on the signal indications of single or multiple targets. The invention is described in more detail below in exemplary form with reference to the following drawing, in which Figs. 1, 2 and 3 together in the form of a block diagram show the basic elements included in the plant, Fig. 2 showing a Doppler following loop and Figs. Fig. 3 shows an angle tracking loop together with a discrimination coupling for distinguishing between single and several targets, Fig. 4 a table for Asking a combination of target confirmation investigations and Fig. 5 a block diagram for a system for angle tracking and power steering.
    At the part of the tracking system shown in Fig. 1, an antenna 1 transmits radar pulses with a frequency which is controlled by a local oscillator 4 and a control path 6 for Doppler tracking, the oscillator and control signals being combined in a mixer 8 and supplied to the antenna via an amplifier 9 and a circulator 11 pA kAnt sat. The antenna comprises a so-called comparator stage, which receives mAlekot and possibly other received signals, which is received by a square directional antenna system with four antenna elements, and feeds three output channels, namely the bearing and elevation difference channels 13 and 15 and the sum channel 17 also on the edge.
    The antenna is controllable with respect to the robot by means of servo-controlled motors 19 to keep the antenna's line of sight directed at one goal.
    The current direction of the line of sight in relation to the robot is determined by gimbal-mounted interceptors 21.
    The sum and difference signals are each supplied by mixers 23, 25 and 27, each of which is supplied with an input signal from the oscillator 4. The true frequency is the sum of the frequency of the local oscillator, a fundamental intermediate frequency which vibrates from a voltage controlled oscillator 29, and the estimated frequency corresponds to the voltage supplied to the oscillator 29 and constitutes an output signal from the Doppler tracking loop. The output signal from mixers 23, 25 and 4 27 is therefore the intermediate frequency plus the error in estimating the Doppler frequency (actual Doppler frequency = fd). Mixers 23, 25 and 27 are followed by main amplifiers 33, 35 and 37.
    This is followed by an electronic connection for angular tracking, in which each part of the sum signal is supplied with each of the difference signals, this part being for the bearing difference and for the elevation difference. This part is calculated to form a resulting difference channel signal equal to zero and thus simulation of adaptation of the antenna's line of sight to the target, whereby the current angular information outside the line of sight (apart from errors in the lead) is contained in A control signals Ea and E (estimated error between Ea and E). antenna line of sight and target line of sight), which are supplied to the elements 39 and 41 included in the connection for angular tracking via lines 47 and 49.
    The resulting sum and difference signals are applied to spacer gates 53, 55 and 57, which are opened for a predetermined short period of time with a controlled delay period after sanding of each radar pulse. The duration of this demand period is controlled by means of a step 59 for distance tracking. The sum channel is branched for feeding an auxiliary distance gate 51, which deviates from other distance gates in that the gate period is divided into two halves, of which the transmitted signal inverts below one half. The output signal from the sum distance gate 57 and the partially inverted output signal from the distance gate 51 are applied to a phase-sensitive detector, which thus emits an output signal equal to zero, since the received signal is equal to the first and the second half of the auxiliary gate. If the signal is earlier or later, the output signal of the detector will have a resulting negative or positive value, which is used by step 59 to re-center all the distance gates on the signal. The goal is therefore distance-tracked effectively.
    Step 59 also supplies a ratchet signal to the transmitter amplifier 9, while the spacer gates are open, to ensure that no breakdown of a true pulse through the circulator 11 can interfere with paint signals.
    After the distance gates 51, 53, 55 and 57, the sum and difference signals are fed to speed gates 61, 63, 65 and 67. These are, in fact, bandpass filters, which thank the possible range of Doppler frequencies of interest and which are centered on the above-mentioned intermediate frequency. The Doppler tracking line tries to keep the various sum and difference mile frequencies at the center of the speed gates' passband by controlling the oscillator 29. The sum and difference signals are then divided according to the sum of the sum channel signal & by means of amplifiers 71, 73, 75 and 77, which are automatically amplified. by means of a detector 69. là Up to this point, all signals have been analog but are now converted by means of analog-to-digital converters 81, 83, 85 and 87 in for digital frequency analysis. In each of the sum and difference channels there is a filter 91, 93, 95 and 97 for fast Fourier transformation. Each such filter comprises a group of e.g. 32, 64, 128, 256, filter elements, typically e.g. 128, whereby these filter elements each take their adjacent narrow frequency band together thanks to the speed band passband. The frequency bands for such filters are known as the frequency slots below. The characteristics of habitual frequency compartments are pointed and overlap the characteristics of adjacent compartments. The output signals from the frequency slots appear in digital form and represent the amplitude and phase of the signal component of the special frequency slot. For a set of such outputs, information from 128 analog sample values is required. The output frequency of these filters, ie. the refresh rate, is 78 Hz, if the total bandwidth of the filters is 10 kHz, with 128 frequency slots. The refresh rate therefore equals the individual frequency slot width. The output frequency at the filters, ie. the update frequency, can be about 80 Hz. The width of the frequency slots is controllable, as indicated by the control coupling 89, so that it covers the intermediate frequency range at fA, wide frequency slots or a larger number of corresponding narrower frequency slots. Each of the sum, difference and distance channels is thus analyzed in e.g. 128 frequency bands with Doppler intermediate frequency range, so that a target signal frequency can be identified very narrowband.
    Fig. 2 shows the remainder of the Doppler tracking loop, which is completely digital and formed by procedures performed by a data processor. The input signal appears on line 97 'of the sum channel filter 97 in Fig. 1, this single input signal 97' representing all of the 128 sum compartment signals. A detector 101 for detecting the grinding frequency compartment performs an initial estimate of the identity of the grinding compartment, as explained below, which gives an original grinding compartment frequency 00, i.e. offset from the center frequency of the speed gate.
    A matching filter 103, which is supplied with the output signals from the transformation filters, generates, depending on the initial estimate of grinding frequency fo, a derived characteristic, comprising two individual frequency characteristics, which are determined symmetrically above and below fo. One of these frequency compartments is derived from the left compartment and the middle compartment by three adjacent compartments and the other from the middle compartment and the right compartment by the three. A discriminator 105, an amplifier 107 with automatic gain control and a diversion stage 109 form an error signal Ed, which is the frequency error between the estimated target frequency (io in this case) and the current target frequency.
    As will be explained, the error signal d is further processed but is also added directly to the current estimated value f of the grinding frequency (originally io) in a treatment step 111. The result is therefore the current magnitude of the grinding Doppler frequency in relation to the center frequency of the speed gate. . This digital value is subjected to force amplification control at 113 and acts as a digital output signal from the Doppler tracking loop.
    To return to Fig. 1, the output signal from the Doppler tracking loop on line 113 'is converted back to analog form and supplied with an integrator 121, which accumulates the Doppler frequency error and controls the oscillator 29 accordingly. In the event of a constant frequency error, ie. constant grinding acceleration, the loop will be read at the grinding frequency and cause the frequency of the oscillator 29 to be adjusted to follow it. At a constant relative milling speed, the frequency error will disappear, while the input signal to the oscillator 29 will be zero and the output signal from it will be constant at the intermediate frequency.
    An accelerometer 117 detects the acceleration of the robot and contributes 7 with a corresponding factor to the input signal of the integrator via an addition step 119.
    In order to return to Fig. 2, the above-mentioned further processing of the discriminator output signal was also used to form a signal 7 on the basis of a final addition 123, which outputs an additional estimated value of the frequency error of the target frequency in relation to the center of the speed gate. This signal was used at a target power step 125, which is explained in more detail below.
    Confirmation of the target frequency compartment is used in several ways: The procedure for electronic angular tracking described with reference to Fig. 1 is activated, the robot's waveguide system is activated and in connection with the main arrangement an indication of signal-to-noise ratio use and detection of several targets is explained. .
    In a closer study of Fig. 2 and in particular the device 101 for detecting target frequency compartments, it is found that the transformation filters in each sum and difference channel form output signals which give the complex amplitude in each of the frequency compartments of the filters, e.g. 16, 32, 64, 128 or 256 frequency compartments according to the compartment width installed by means of 89. The filter compartment width is equal to the total bandwidth of the filter device, divided by the number of filter compartments, while the output data frequency is almost equal to the compartment width.
    Target detection is performed with respect to all dignity output signals, ie. the square of the sum of the real and imaginary components, from the transformation filter in the sum channel with the exception of, ie. the first and last NL 'dar NL = 9, 5, 3, 2, 2 far 256, 128, 64, 32 and 16 frequency slots. The dignity output signal for each compartment is divided by the mean value of all frequency compartments with the exception of the first and last NL, after which this value is compared with a threshold value TD, which is determined in such a way that a certain number of junctions becomes possible if the input signal is purely thermal noise. (fake lamb). The applied technique is to compare each compartment dignity, which is treated by means of the mean dignity as above, together with the maximum calculated value of the preceding compartment and with the threshold value TD. The maximum frequency compartment at the end of the procedure, which also exceeds TD ', is interpreted as a target alarm for the relevant frequency compartment. If no frequency compartment exceeds the threshold value, no alarm occurs and the detection procedure is performed with respect to the next set of filter data. This detection method can be modified in practice to enable multiple alarms.
    If the template alarm is detected, its compartment number and center frequency cow are output, so that the Doppler precipitation loop and the confirmation procedure can work with correct frequencies. An additional calculation is performed during the detection process. The content of the alarm frequency compartment is reduced by the above-mentioned threshold value TD. This is because the probability distribution of the maximum frequency compartment, which exceeds a certain threshold value, is almost a constant (TD) plus a Rayleigh distribution, assuming that TD is> logeN. By subtracting the threshold value, the original alarm can be treated in the same way as subsequent dignity outputs from the union, which simplifies the confirmation technique.
    The calculations performed during the above procedure for each compartment dignity pi to establish the frequency compartment with maximum dignity are: (1) N-NL Pi ri = pi j = NL + 1 where r1 is the scaled dignity of a frequency compartment with no N the total number of frequency compartments, NL is the number, is the number of a frequency compartment among those to which the average value formation is taken into account. If ri> TD and also> rcrimc, set rcmax = r.and stored in the sAsorn the number of the alarm frequency compartment. The quantity rcmax is the maximum value of r, which has occurred so far. The stored value of 1, which remains when the view is taken to special frequency compartments, is the d & m & alarm compartment. This initial installation or estimation of the milling frequency (k0 in Fig. 2) was then used to close the Doppler tracking loop and trigger the power up procedure. While the 9 mAl frequency is within the detected frequency slot (i), it will of course not generally be exactly the center frequency in front of the frequency slot.
    In order to achieve the probability of target capture together with the legal probability of false confirmation, it is necessary to sum up incoherently over a large number of filter update periods. To enable this, follow-up is required, since either the target must be kept in a certain frequency compartment or the alarm frequency compartment may be lost in some other way.
    For this reason, capture in tv4 was used in steps. The first step, which has already been described, detects the situation for a probable target, while the second, which relates to confirmation, states that this detected signal is indeed a case with a high degree of probability.
    The mile power step 125 shown in Fig. 2 operates in the following manner. The alarm frequency compartment E is indicated by the Doppler tracking algorithm in fi4. 2 except immediately after the first detection, if the detection sub - program includes this information.
    As with the detection algorithm, the equivalent of equation 1 above is used to form an output signal for scaled alarm. If the alarm compartment is outside the range NL + 1 to, the scaled compartment value is set to zero. Thereafter, this scaled technical value in data processing is added to the same quantity in the subsequent update. This addition continues until a predetermined number of dignity values from the target compartment have been summed. At this time, the sum is compared with TWO threshold values, namely an upper and a lower threshold value.
    The sum of the frequency compartments is greater than the upper threshold value, a mark for target presence is set, while target absence is confirmed, if the sum is smaller than the lower threshold value, ie. the presumed target is rejected. If the sum is in between, the selection remains in an unconscious state. In this case, another set of dignity values, derived from updates to transform filters, is added to the first set and the accumulated sum is compared with two new and more closely related threshold values, of which the 8th Aterigen indicates the confirmed NDR and the lower one confirms the AW.
    After a predetermined number of such comparison processes, which occur at the beginning of the program, the first and lower thresholds are brought together, so that a final decision is forced, even if it may not be final, as shown below. This successive comparative study is called the indication of accumulated sum.
    At this point, a current mean value is formed by the last N statements, of which habit updates transformation filters.
    Since the current value value is included in the current sum, the earliest input value is discarded. The criterion for confirmation is that the mean value in question should be above a threshold value, which is normally the threshold value, which was caused to converge during the last comparison procedure, and that in addition the individual comparison procedures should not be rejected. This latter investigation is carried out to ensure a reasonably rapid reaction, if for some reason the target would disappear locally. The result Az is an indication of the current average.
    The indication of accumulated sum and the indication of current mean value combine according to Fig. 4 to obtain a final judgment. It is found that the indication of accumulated sum acts is preferred, if the indication of the current mean is negative, while the indication of the accumulated sum is in fact increased rued a level of indication certainty, on the indication of the current mean is positive.
    In addition, there is a certain renewed start-up room. If the total confirmation leaves the state of nerve confirmation, the current mean value is dipped. If the examination of accumulated sum rejects or confirms, the accumulation of subsets of sums also ceases and each subsequent subset is examined separately.
    When template power (125) is completed, nerve power 126 is used to activate the angular tracking loop and the robot guidance loop.
    The applied principle of Doppler tracking enables this tracking both in the digital data processor 11 shown in Fig. 2 and by means of the oscillator 29 in Fig. 1. The key function which enables tracking by the data processor itself is the procedure which takes place in the loop 103, 105, 107 , 109 with adjustment filter and discriminator. To achieve frequency discrimination, the dignity signals from two adjacent frequency bays can be subtracted.
    If the target signal is symmetrical over the transition between the two frequency bays, the dignity will usually be the same and the difference will be zero. If the target signal is outside the midpoint towards one or the other direction, the result will be positive or negative accordingly. If two such frequency compartments, which are centered on a special estimation value of mile frequency, can be simulated, therefore, frequency shift from this medium can be detected and the target signal frequency can be monitored continuously.
    In order to continuously set the Second layer for the discriminator frequency, a technique for adjusting the filters is applied to the transformation filter 97, which requires simulation of each of the two above-mentioned simulated compartments. The following applies to the simulation of each of the two. The complex output signal from two existing adjacent frequency bays is determined by the possibility of using these two quantities alone to build a new frequency bin, which, like its peak, will have each selected frequency value between the midpoints of the two original bins. To derive a simple algorithm, a rectangular gap was provided and phase factors of w / N were neglected between adjacent frequency bays. The new characteristic of the frequency compartment can be roughly written as: a sin w x13 sin w (1-x) f (x) = (1-x) (2) ddr x is the brAll part of a frequency compartment from the middle of the first compartment, ddr the reaction is required, a is the amplitude contribution from the first compartment and a is the amplitude contribution from the second compartment. If a, a are related by a parameter 6, 2 - - 6+ wcotw6 (1-6) 7 Ircot76 (6 (1-6) (1-26)) - 1 + 26 - 262 1 - a can show that 12 where the point f (x = 8) is a maximum for the filter according to equation 2. In this way, an adjustment filter can be built up by taking the frequency f, evaluating the frequency compartments closest to it and then calculating the parameters, A, required to obtain tvA filter compartments, each of which is separated by half a compartment spacing from the estimated target frequency f. The same technology is applied to the filters regardless of the type of door used. The equation for the matching filters becomes dA: A.
    Nta = integer part of + 0,5B) ts + 1,5) x C1 = 0 ((x) (ie the above function calculation for a predetermined rate x-value) Ca = C1 b (Nta-1) - (1-C1) b (Nta) Cb = (1-C1) b (Nta + 1) - C1 b (Nta) where Nta is the number of the frequency compartment, which is entered by A the required frequency, f is the estimated Doppler frequency according to the digital tracking loop, B is the total bandwidth of the transformation filters, ts is the updated time (= compartment number B), N is half the number of frequency compartments in the transformation filter and b (N) is the Nth filter compartment, thus the Ca and Cb outputs from the adaptation filter with simulated adjacent frequency compartments are centered on & frequency E and are Ater complex.
    Then the discriminator 105 shown in Fig. 2 must be constructed, which can be done by selecting the frequency error d = ICa12-1Cbl 'But this filter technique results in scaling at the original of this magnitude, which depends on the filter layer, so this must be corrected. A square correction term is used of the form -2 Sd = a1 + a2 x + ae x dar x = min (x, 1-x) and a1, a2, a3 assume different values for the different shutter functions used. As a result, the discriminator can also be scaled automatically for applicable door function. The discriminator output signal must be automatically gain-controlled, which is achieved through a simple first-order feedback system, which works with a square detector 107. The detector works with a new frequency compartment of the shape Cagc = Ca + Cb 13 which gives a filter, which is centered on the & mAl frequency.
    In the case of the discriminator, this seems to provide optimal performance. The automatic gain control is thus agcn = agcn-1k + (1-k) ICagcl2Sc dar Sc is another scale correction factor for the different gains, which are found at different stalls on the adaptation filter and different gaps. This dr is determined by - Sc = 1 + x bl + x2 b2 where b1 and b2 are constants, which depend on the type of gap, while, k is equal to exp (-ts / tagc), where tagc is too strong-. time constant. The resulting output signal cfT (from ..division step 109) is td = d Sc Sd) (agcn t) Hz The digital tracking loop is completely closed in the data processor (see Fig. 2) and forms the estimated value of target frequency required for the validation procedure . The special use. the construction is a type with two loops. A block diagram of the device is shown in Fig. 2.
    It is found that the loop emits two outputs, one of which is for the validation procedure (Z) and one for the feedback (f) to the combination filter / discriminator combination. The transition functions for these bAda are different and are determined by 2 zG1 (1 + G2) - G1 Y = z2 + z (G1 (1 + G2) -2) + (1-G) 1 Yz2G1 + G1 (G2-1) z Y. z2 + z (G1 (1 + G2) -2) + (1-G) 1 where f is the target frequency relative to the center of the low-frequency sum channel. These transition functions are obtained by taking the output signal ED from the discriminator and then multiplying by the gain Gl.
    The algorithms for forming these two closed loop transition functions are as follows. If the discriminator's output signal is obtained 14 f (n) = f1 (n-1) + G1 d 1 f 2 (n) = f2 (n-1) + G2f1 (n) A f (n) = f1 (n) + f2 (n) F (n) = fi (n) + f2 (n-1) These algorithms AskAdliggores in Fig. 2 pA following sat. The output signal from the discriminator chain 105, 107, 109 is subjected to a gain G1 at 127, which results in an input signal Greci to an addition stage 129, a storage and celebration device 131 supplying the output signal from stage 129. back to its entry at the next update. If f1 (n) is the output signal from the addition step 129 at the nth update, f1 (n) must therefore be equal to Grid f1 (n-1), so that the loop is clamed for digital integration. The output signal f1 (n) from the loop is supplied with a further storage and retrieval device 133, the output signal f1 (n-1) of which is supplied to an addition stage 135. A second input signal to this addition stage is derived from the output signal f1 (n) from the loop above and subjected to a gain G2 at 137, so that G2.f1 (n) is obtained, and a further integrating loop 139, 141 to obtain f2 (n) equal to f2 (n-1) + G2.f1 (n). Another storage and displacement device 143 then forms f2 (n-1), i.e. the second input to step 135.
    The output signal from the addition step 135 dr sAlunda 2 (n-1) = f1 (n-1) + f2 (n-1), i.e. the estimated target frequency, based on previous updating of the transformation filter information. This output signal is applied to the matching filter 103 for calculating the next error 6.d and is also applied to the addition step 111 together with the current error 6 according to the description above to form the output signal from the Doppler tracking loop.
    The signal f2 (n-1) is applied to an additional addition stage 123, but in this case together with the current input signal to the step 135, i.e. nice). The summed output signal from step 123 is thus f1 (n) + f2 (n-1), which is denoted by f and as the estimated target frequency is applied to the confirmation step 125 pA as described above.
    According to Fig. 3, the outputs from the filters 95 for input difference on lines 95 'and the sum signals of the filters are input in the same way on lines 97'. These sum and difference signals, each from e.g. 128 frequency compartments, are each supplied with matching filters 145 and 147 pA in the same manner as the signals described for the Doppler loop in Fig. 2. In each case, the estimated ual frequency taken from the Doppler loop is used to select the frequency compartment at which it is located, or assumed frequency compartment is derived from this frequency compartment and an adjacent frequency compartment, the simulated frequency compartment with its center or peak frequency being adapted to the estimated monthly frequency. Changes in the mai% frequency, which are determined by the Doppler tracking loop, cause A to become the value of f Others, whereby the simulated frequency range is filled List at the same and clamed slides up and down the intermediate frequency range together with the Doppler error frequency and Achieves a high degree of disk .
    The output signals from the adjustment filters are digital and are therefore supplied to the digital equivalent of a phase-sensitive detector.
    The complex conjugate to the difference signal in complex form is therefore formed by a step 149. The sum and the complex conjugate to the difference are then multiplied by each other in a step 151, which emits real and imaginary output signals. The real part is divided by the output signal from a detector 157 with automatic gain control to achieve the amplitude ratio for each of the two channels, i.e.
    (De-keS) Soch (Da -kaS) S S2S2 Dar S2, ie. the effective value of the sum signal, constitutes an output signal from the detector 157. Only the delivery difference channel is shown, but the elevation difference channel is treated in the same way. This is applied to a Kalman filter in Fig. 5 to obtain an output signal which is added to an aberration correction factor and applied to the gain element in the angular tracking loop to close this loop. The Kalman filter shown in Fig. 5 emits an estimated value Ea of the error in the line of sight of the antenna, i.e. the estimated value of the target angle outside the line of sight. This estimated error value is combined in an addition step 156 with an indication of line of sight direction relative to the robot. This latter 16 is achieved in principle by means of gimbal suspension transducers 21, which are also shown in Fig. 1 and are corrected for radar aberration at 150, which in turn depends on the line of sight direction relative to the radome, radome temperature at 163 and sand frequency at 161. This correction device therefore forms baring and elevation angles for the effective line of sight instead of for the physical line of sight.
    The resulting signal from the addition stage 156 can be used in Fig. 1 to control the element 41 in the angular tracking loop and close this loop.
    When the angle tracking loop is in equilibrium, the vardete is an exact center of the line of sight error.
    In view of the imaginary output signal from stage 151, this is divided by the signal for automatic gain control in the division stage 153 to form a form of indication of signal-to-noise ratio. This output signal from the division stage 153 is an indication of the dignity of the signal components, which are either non-coherent between the sum and difference signals or phase shifted 900 is embedded. The non-coherent dignity component is large when thermal noise or disturbances in either the sum or difference channels are large, and penalties for zero at high signal-to-noise ratios. The non-coherent dignity output signal is also large in the case of elongated targets or targets that fly in formation and are not resolved by the searcher.
    The signal-to-noise signal from stage 153 is applied to a division stage 159, the second input of which is derived from the signal-to-noise signal from stage 125 for squeegee confirmation in Figs. 2 and 3. The two signal-to-noise indications deviate in that the basic indication is derived from the confirmation stage. the dignity of the target frequency compartment and the average dignity Over the transformation filter band. It therefore does not distinguish between a coherent single mAl VAT mA frequency frequency box and non-coherent multiple mAl frequency box. The indication is derived only from the target frequency compartment and only distinguishes between coherent or single targets on one side and multiple targets and noise on the other side, so that a low value is obtained for the coherent target and a high value for the non-coherent targets. Therefore, if this indication is divided by the basic indication from the confirmation step, the result will be a low value for a single coherent target frequency box and a high value for the multiple target frequency box. The result of the division thus gives an indication of single targets or multiple targets.
    The described plant is adaptable to different robot-target states, which can arise in the event of an attack and especially in the final stage, since the state can change quickly.
    The parameters in the Doppler loop are adaptable to take into account varying target conditions, so that optimal performance is constantly maintained. Reference is made to Fig. 2.
    Immediately after target detection, the gains G1, G2 and G3 are set to high, so that the loop reacts quickly. The pore strengths are then reduced until, after a predetermined period of time, they assume constant LOW values to reduce the noise in the loop. The gains are increased while the robotic flight continues, in accordance with the signal strength measured in the amplifiers by automatic gain control, or by any of the above-described signal-to-noise ratios described in accordance with the selected filter frequency compartment width.
    During the launch phase of the robot, the Doppler tracking loop is exposed to significant transients and the frequency range width of the transformation filters is selected to a large value. The reinforcements G1, G2 and G3 are also installed at high values to achieve rapid reaction.
    If a large-distance target is not encountered, the frequency compartment width is set to a very small value and G1, G2 and G3 are installed to the low value after loop activation. This brings the capture threshold to its optimum value.
    If the target is at a short distance, install the frequency compartment width to a large value and install G1, G2 and G3 to the high value. This achieves the optimal capture speed and reduces tracking errors to a minimum. A wide frequency compartment width provides faster update speed and thus optimal performance very close to the stop, where fast response is essential. TARGET DISTANCE is determined by means of the distance tracking loop shown in Fig. 1.
    As soon as the loop has started to work, the signal f is proportional to the velocity of the target frequency, i.e. mAlets acceleration. If f is large, the frequency compartment width is installed to a high value to enable high tracking mobility. The reluctance in Doppler discrimination with regard to other cases with similar speeds is not significant, since a maneuvering m & 1 cannot be kept close to other cases. When in is small, small compartment width was used to avoid maximum Doppler discrimination with regard to men flying in formation.
    When multiple targets taken next to each other are detected, e.g. by the output signal from the division stage 159 in Fig. 3 or the acceleration of the mn is small, the frequency compartment width is installed to a low value to achieve maximum Doppler discrimination.
    The entire angle tracking system can be divided into three parts by four parts. Firstly, the receiver in the angular tracking loop, secondly the angular tracking filter and thirdly, the power steering and stabilization system.
    A block diagram of the complete system for angle tracking and power steering is shown in FIG. The angle tracking filter estimates the angular error with respect to the line of sight of the antenna in each of the bearing and elevation channels, this error being used to add a controlled portion of the sum channel corresponding to these angles, in contrast to the stored difference / sum slope at the origin of the difference channels. as described above with reference to Figs. 1 and 3. These difference signals are then transmitted through similar intermediate frequency chains as for the sum channel, sampled and converted from analog to digital form (81-87). The Fourier are then transformed (91-97) to obtain the spectrum of signals on the birth and elevation channels. The sum and difference channels are digitally divided by signals for automatic gain control and elevation and baring signals are formed by phased detectors at 155 by selecting the real part of these 19 signals. These signals are proportional to the angular error between the actual error of the target line of sight and the best estimate of the line of sight error of the antenna. These calculations are all performed at the same refresh rate as the Doppler and distance shark loops, ie. with the refresh rate of the filter transformed output signal.
    After detection at 151 mA, a shaft transformation must be performed to transform from receiver shafts to line of sight shafts.
    This is done digitally. These signals are then transmitted through an Kalman fabric ring tracking filter 171, which forms the best estimated value of line of sight velocity using a variable color gain algorithm. The sight line speed signals are used in different ways. They first provide a signal to the autopilot as a guidance command, for which purpose they are transformed into robotic axes, but they are also used as an input to the integrator included in the angular error loop, which feeds the estimated angle back to the element in this loop, and are finally used as an input signal. step 173 is given a line of sight to drive the antenna mechanism.
    The latter is fed to a gyro output stage 175 for fidelity reference to determine the total motion of the target, which is fed along line of sight axes.
    These angular velocities are still calculated with the loop's low refresh rate. These velocities are then integrated using a digital integrator 177 and convert the output signal from line of sight axes to motor axis angles using digital axis transformations 179. The angular cup signals are then used to control a layer controlling loop 181 for the antenna. The angular request signal at 183 is subtracted from the potentiometer capture elements 21 and the fault is fed to a control unit 185, which supplies request signals to the antenna mechanism, servo and motors 187. The layer of the antenna reflector plate is slave controlled to the integrated gyro signal and the target line error. This control unit operates at a high refresh rate, so that body movements on the receiver's output signal are eliminated. Extra line 189 for line of sightFeedback was used to the element 39, 41 in the angular tracking loop. This is added to the second feedback path and the fact that the most estimated value of line of sight error can be used to be subtracted from the current line of sight error of the angular tracking loop. The combined feedback path must be angularly transformed at 191 back to the receiver shafts and in addition correction of radon aberration at 159 is added depending on the cardan shaft angles, as described with reference to Fig. 3.
    The parameters of the Kalman filter are different, depending on the condition during robot flight, and are generally designed to provide large bandwidth, when the estimated time up to the stroke is small. Time until the stop is estimated by dividing the distance between the robot and the target (from the distance tracking loop) by the decreasing speed (from the Doppler tracking loop). The filter bandwidth is also dependent on the signal-to-noise ratio, the bandwidth being small when this ratio is small.
    权利要求:
    Claims (4)
    [1]
    1. set at single pulse tracking radar, which works with sum and difference signals to determine target direction and comprises a Doppler tracking loop for holding an intermediate frequency target signal with the passband for a speed gate filter and one digitally arranged in each of the sum and difference channels filter device, which at periodic update intervals performs analysis over a number of adjacent frequency compartments of target signal components including the passband, confirms the presence or absence of a target signal at a special frequency compartment, characterized by a series of comparison procedures being performed, each depending on the sum a predetermined number of consecutive dignity signal values from the special frequency range, accumulated together with any preceding such sums, and comparison of this accumulated sum with the upper and lower threshold values, which come closer to each other for each comparison procedure, thereby confirming of a mAlsig The usual threshold procedure is indicated when the upper threshold value is exceeded by the accumulated sum, while confirmation of the absence of a target signal is indicated when the lower threshold value exceeds the accumulated sum, and that another comparison procedure is triggered when the accumulated sum is the lower threshold.
    [2]
    Set according to claim 1, characterized in that the upper and lower threshold values are caused to coincide after the series of a predetermined number of comparison procedures, so that an indication of the presence or absence of a target signal is forced.
    [3]
    Set according to claim 1, characterized in that a current mean value is established, which comprises a certain number of dignity signal values, wherein the earliest incorporated signal value is rejected as current value, so that a current average value is obtained, which is compared with a predetermined threshold value for Providing an indication of the presence or absence of target signal, wherein the indication of accumulated sum and the indication of current mean value contribute to a conclusion, at which the indication of accumulated sum acted, if the indication of current mean indicates absence of mil signal, and the possible results of the indication of accumulated sum: target signal absent, indefinite or present, are treated as indefinite present resp. present, if the indication of the current mean value indicates the presence of a target signal.
    [4]
    4. set according to claim 1, characterized in that a digital filtering is performed for each of the sum and difference signals, which at periodic update intervals & provides analysis of a number of adjacent frequency compartments of potential maize signal components, identification is performed of a target frequency compartment, supply is performed of sum and difference signals with respect to the identified target frequency compartment as inputs to a device for forming a complex product of one of these inputs and the complex conjugate to the other, a signal-to-noise indication being derived from the imaginary component of the complex the product and an indication of the dignity level of the sum channel signal in the mile frequency compartment, so that this indication has been high in the presence of non-coherent reflections from multiple miles and l> in the presence of coherent reflections from a single target, as well as a basic indication of signal noise -forh011ande, which is derived from the level of dignity & mo m mole frequency compartment and average dignity level Above all frequency compartments, this basic indication has increased value in the event of a single or more mAl VAT m & m frequency frequency compartment and added value in the event of broadband noise, and a device which, depending on the indications of signal noise, of single or multiple ma '. / DISTANCE ACCELERATION 7: 5) INST. DEPARTURE STADIUM POCKET WIDTH - "--- MULTIPELMPL BAP. DISTANCE BAR. STUDENT. / 49 126 ACT. ANGLE FOLLOWING + ROBOT GUIDANCE. / 33 Zg7 / -74 11 ••• ■ = 1111111 ■■ ••, L / 73V43 / 47- / t_ / / 59/53 BAR DIFFERENCE cG. IMAG. REAL SIGNAL NOISE FOR KALMANFILTER ETC.I. FIG. / 5 / TOTAL / 1 r / 57 xlz 7-0 g 4 - ,, UNDER INCREASING SUM AMOUNT ROVING AVERAGE VOICED VETEE CONFIRMED VETHE CONFIRMED CONFIRMED VOICE DECLINE VETEY CONFIRMED GRAVE EXPENDITURE {<. —Ro OUTSIGN FROM ROBOT CONTROL HEAD BODY / 87 (4.). ■ ••• ■ • AID - OMV. STEERING ANT. LAGE / 8 / Ix FILTER 9 / g7 OTJ AT2 / 831/77/79 D / AOMV. A / D- OMV. (2 / T4
    类似技术:
    公开号 | 公开日 | 专利标题
    SE8803569A1|2012-04-17|Use single pulse tracking radar to confirm the presence or absence of a target signal
    US6229475B1|2001-05-08|Pulse doppler radar system with improved cluster target resolution capability
    DK149370B|1986-05-20|APPARATUS FOR DISTINCTING TRUE TARGET COUNTER SIGNALS FROM ISAER DISCRETED SIDE LOAD INTERFERENCES IN AN IMPULS DOPPER RADAR RECEIVER
    EP2062067B1|2011-05-11|Method of and device for tracking an object
    US3896411A|1975-07-22|Reverberation condition adaptive sonar receiving system and method
    CN102947725B|2015-01-21|Radar device
    AU2010202039C1|2012-04-05|Direction-finding method and installation for detection and tracking of successive bearing angles
    US5038145A|1991-08-06|Constant false alarm rate with low probability of false alarms
    US4419659A|1983-12-06|Intrusion detection system using leaky transmission lines
    US4016486A|1977-04-05|Land mine detector with pulse slope, width and amplitude determination channels
    GB2462871A|2010-03-03|Improving the angular resolution of monopulse radars
    US4063238A|1977-12-13|Range tracking apparatus in a doppler radar
    Howland1995|Passive tracking of airborne targets using only Doppler and DOA information
    SE507996C2|1998-08-10|Procedure for determining the relative velocity between two objects in motion
    KR960704238A|1996-08-31|Radar apparatus
    Flanagan et al.2005|Robust distributed detection using low power acoustic sensors
    JPH05188145A|1993-07-30|Method and apparatus for extracting phase error caused by atmosphere in backward scattering signal of interference-type image radar system mounted on carrier out of raw radar data
    US5093665A|1992-03-03|Point target filter
    JPH04147079A|1992-05-20|Method and apparatus for processing radar signal
    KR101454297B1|2014-10-27|System and method for intercepting missile using high range resolution fmicw
    JPH0555941A|1993-03-05|Method and device for reducing noise for detecting target
    US20050046610A1|2005-03-03|Advanced asynchronous pulse detector
    Meng et al.2002|Radar detection improvement by integration of multi-object tracking
    US3061830A|1962-10-30|Signal-to-noise detector
    Damarla2003|Tracking a convoy of multiple targets using acoustic sensor data
    同族专利:
    公开号 | 公开日
    GB2388985A|2003-11-26|
    GB2388985B|2004-04-07|
    FR2852401A1|2004-09-17|
    GB8136055D0|2003-04-09|
    US7053815B1|2006-05-30|
    IT8367632D0|1983-06-09|
    GB8232892D0|2003-04-09|
    SE8206787D0|1982-11-29|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5047781A|1968-06-05|1991-09-10|Hughes Aircraft Company|Radar sensing generator in a monopulse radar system|
    BE754279A|1969-08-12|1970-12-31|Hollandse Signaalapparaten Bv|RADARAPPARAAT MONOPULS|
    US4442431A|1971-07-12|1984-04-10|Hughes Aircraft Company|Airborne missile guidance system|
    US3952304A|1973-11-23|1976-04-20|Hughes Aircraft Company|Tracking system utilizing Kalman filter concepts|
    US5233351A|1975-08-26|1993-08-03|Raytheon Company|Local oscillator arrangement for a monopulse receiver in a semiactive missile guidance system|
    US4179696A|1977-05-24|1979-12-18|Westinghouse Electric Corp.|Kalman estimator tracking system|
    US6771205B1|1977-07-28|2004-08-03|Raytheon Company|Shipboard point defense system and elements therefor|
    US4156875A|1978-03-13|1979-05-29|Raytheon Company|Range gate generator with velocity aiding|
    US4271412A|1979-10-15|1981-06-02|Raytheon Company|Range tracker utilizing spectral analysis|US7741991B1|1987-06-26|2010-06-22|Mbda Uk Limited|Radar tracking system|
    US7417584B1|1989-11-08|2008-08-26|Lockheed Martin Corporation|Monopulse radar estimation of target altitude at low angles of elevation|
    US20060208945A1|2005-02-28|2006-09-21|James Kolanek|Space / time / polarization adaptive antenna for ESM / ELINT receivers|
    US7504982B2|2005-12-06|2009-03-17|Raytheon Company|Anti-Missile system and method|
    FR2904428B1|2006-07-27|2010-04-09|Thales Sa|METHOD FOR IMPROVING THE QUALITY OF BASIC BAND TRANSPOSITION OF SIGNAL RECEIVED BY HIGH-RESOLUTION RADAR COMPRISING AN ANALOGUE RECEIVING CHAIN|
    US7633432B2|2006-12-13|2009-12-15|The Boeing Company|Method and apparatus for precision antenna boresight error estimates|
    US8321491B2|2007-06-26|2012-11-27|The United States Of America As Represented By The Secretary Of The Army|System and method for detecting a weak signal in a noisy environment|
    JP5111031B2|2007-09-14|2012-12-26|キヤノン株式会社|Displacement detection method and motor control device|
    JP2009198363A|2008-02-22|2009-09-03|Omron Corp|Radiolocater and method|
    US7864101B2|2008-07-28|2011-01-04|Raytheon Company|Radar tracking system|
    US8194766B2|2009-05-22|2012-06-05|The Aerospace Corporation|Constant false alarm rate robust adaptive detection using the fast fourier transform|
    US8831155B2|2011-12-29|2014-09-09|Qualcomm Incorporated|Radar detection method and system using low-resolution FFTS|
    US10181643B2|2015-03-05|2019-01-15|The Boeing Company|Approach to improve pointing accuracy of antenna systems with offset reflector and feed configuration|
    CN106707269A|2015-11-13|2017-05-24|南京理工大学|Radar object speed tracking method based on cross-product automatic frequency control|
    CN108563144B|2018-03-30|2021-06-29|西安电子科技大学|Missile-borne radar signal processing semi-physical simulation test system|
    法律状态:
    2013-04-16| NAV| Patent application has lapsed|
    优先权:
    申请号 | 申请日 | 专利标题
    GBGB8136055.4A|GB8136055D0|1981-11-30|1981-11-30|Radar tracking system|
    [返回顶部]