Ship detection method.

0001The invention relates to a method for detecting vehicles, in which operating noises emitted by the vehicle are received.

0002The operating noise of a vehicle is mainly caused by its drive machines and is detected as driving noise and evaluated for a bearing. However, other operating units located on the vehicle, such as pumps, winches or cooling units, emit sound energy and also contribute to the operating noise of the vehicle, so that detection and bearing can be possible even when the vehicle is stationary. Up to now, sonar systems have been used in waterborne sound technology for detection, which can be used to detect an invasion in the case of coastal protection or an attack by enemy submarines or torpedoes in the course of naval war. In airborne sound technology, the detection of aircraft, e.g. B. helicopters, or land vehicles, e.g. B. tanks, possible.

0003The detection of enemy vehicles usually takes place when the operating noise of the vehicle stands out clearly from a general background noise, namely its volume differs from it. This is the case if the vehicle emits a loud driving noise due to a high speed level or is so close that its operating noise as a useful noise is clearly different from a noise. For automatic detection of the vehicle, a detection threshold for the received noise is therefore set in a receiving system, which is higher the lower the false alarm rate is selected. Such an arrangement is described for example in DE-OS 20 52 283.

0004Detection becomes problematic when the background noise covers the operating noise because it is naturally louder or because the vehicle is still far away and the useful / interference ratio is therefore relatively small.

0005The invention has for its object to provide a method of the type mentioned, which makes it possible to ensure reliable detection of a vehicle even with a low useful / interference ratio.

0006The object is achieved by the features specified in the characterizing part of claim 1.

0007When using the method according to the invention, a noise is received, for example, via a receiver with all-round characteristics. The receiver monitors a marine area and is intended to detect a watercraft, e.g. B. a ship report. During its voyage, the ship generates an operating noise which, due to the propeller driving it, shows periodic volume fluctuations in the range of a few hertz. These volume fluctuations are comparable to an amplitude modulation of a noise carrier and are characteristic of every propeller-operated watercraft. The method according to the invention is used to check whether the envelope of the received noise fluctuates, since the noise is amplitude-modulated. As soon as a modulation can be determined, it is recognized on the vehicle.

0008The advantage of the method according to the invention is in particular that the detection is carried out with the greatest possible certainty, since it is based on the detection of fluctuations in the envelope of the received noise. These fluctuations occur in propeller-powered vehicles, such as surface ships, submarines, torpedoes, or in propeller planes or helicopters, and can also be detected in the event of poor use / interference ratio. Even loud noises do not call out a false alarm because they lack the characteristic fluctuations in the envelope. The false alarm rate is therefore much lower than with conventional methods, which carry out the detection depending on a set volume threshold. Even if the vehicle is far away and the volume of its operating noise is not significantly different from the background noise, detection is possible because the envelope is also modulated with periodic character, which is caused by rotating propellers.

0009If the operating noise is received by a direction finder with which direction finding is used, the detection and direction finding of a vehicle are present at the same time, as stated in the advantageous development according to claim 2. When using a small base direction finder, which has four electroacoustic transducers with all-round characteristics, the detection is carried out with the aid of the noises received by the transducers; a simultaneous bearing to the vehicle is just as advantageous via runtime evaluation, as is given, for example, in DE-OS 30 17 797 is.

0010The development of the method according to the invention according to claim 3 has the advantage that, even with a low reception bandwidth of the receiving system, a target can be detected which radiates operating noise in a broadband manner, even if the energy is emitted in a frequency range which is lower than the reception bandwidth lies because the detection is based on the detection of an amplitude modulation and the sidebands caused thereby occur at every frequency of the noise carrier. Such a direction finder would advantageously be used as a torpedo sonar.

0011The particular advantage of this method according to claim 3 is further that the direction finder can be operated in a frequency range in which directional characteristics naturally have a small opening angle or good bundling, so that an angular separation of targets can be carried out particularly elegantly, since the low-frequency components of the noise need not be received immediately.

0012It is also possible to make a broadband received noise narrowband by filtering or Fast Fourier transformation and to examine this narrowband noise with regard to its envelope curve, since each narrow subband of the entire noise band also contains the low-frequency information of the drive machines or other operating units in operation Vehicle contains. Another advantage of this signal processing according to claim 3 is that interference reduction is possible by reducing the frequency range to a subband.

0013With the advantageous development of the method according to the invention according to claim 4, early detection, separation and tracking of several goals is possible. In particular, when the frequencies occurring in the frequency spectrum of the time profile of the envelope of the received noise are displayed depending on the direction. If two vehicles are located at adjacent bearings, further observation can be used to identify a crossing of bearings using the frequencies assigned to each bearing, and a subsequent correct assignment of the bearing and target can be made.

0014With the development of the method according to claim 6, interference is achieved by integration or averaging. If spectral lines occur only briefly in the frequency spectrum of the envelope over time, it can be assumed that these spectral lines do not belong to the operating noise of a vehicle, but were caused by stochastic noise. A vehicle has only been detected when the spectral lines of the frequency spectra determined over a predefinable time interval occur repeatedly at the same frequency and are thereby confirmed.

0015The advantageous development of the method according to claim 7 allows information as to whether the operating noise first detected was emitted by one or more vehicles. If the number or frequency of the spectral lines occurring under a bearing changes, it can be assumed that the detected vehicle has changed its speed and / or has switched off some of its operating units. If the same spectral lines are recognized under neighboring bearings, several vehicles have previously been sighted at the same angle and multi-target separation has now been achieved. If, on the other hand, all spectral lines suddenly disappear under a bearing, this could be a sign that the detected target has switched off its drive units or switched to listening. As a result, its operating noise has suddenly become so quiet that it can no longer be detected by the direction-finding system. If you always find the same frequency spectra with the same bearing, it can be concluded that the bearing is standing. The detected vehicle moves, for example, at a constant radial speed with respect to the direction finder. If the level of the spectral lines increases at the same time, the detected vehicle is on a collision course.

0016An advantageous device for carrying out the method according to the invention is specified in claim 8. The received noise is demodulated with the envelope detector and the lower sideband is delivered to the detection display unit with the low-pass filter, which checks the resulting time profile of the envelopes to determine whether an AC voltage signal is present at the output of the low-pass filter and then provides an indication, since then the time profile the envelope shows the characteristic fluctuations, which are caused by an amplitude modulation of the received noise.

0017With the embodiment of the device according to claim 9, a simple possibility is given to recognize the fluctuations in the time course of the envelope. The frequency analysis circuit determines the frequency of the time profile and, if it is at a distance from the frequency zero, that is to say an AC voltage, causes a display on the detection display unit. In the simplest case, the frequency analysis circuit is a high pass. It is particularly advantageous to use a switching arrangement for calculating the entire frequency spectrum of the time course of the fluctuations, since this increases the reliability of detection, in particular if, according to the embodiment of the device according to claim 10, additional interference-freeing is achieved by subsequent averaging of the frequency spectra over time.

0018In the embodiment of the device according to claim 11, a comparison of the spectral lines is carried out and vehicle detection is only effected if they are confirmed within a predefinable time interval. This type of integration is called stacking. It is particularly elegant when the signal processing is carried out digitally.

0019The advantage of the embodiment of the device according to claim 12 is that the threshold circuit evaluates each detection with a predetermined and thereby known false alarm rate. The normalization of the frequency spectra specified in the advantageous further development according to claim 13 facilitates the setting of a threshold for the spectral lines in the frequency spectrum, which, when exceeded, detects a vehicle.

0020It is particularly effective to link the device according to the invention to a panorama receiving system in which adjacent directional characteristics are formed all around the azimuth and an angle is assigned to each directional channel. Such an embodiment specifies claim 14. Depending on the directional characteristic, a detection is carried out according to the method according to the invention and the bearing to the vehicle is specified together with the detection. The intensity-dependent bearing carried out up to now by comparing group signals of adjacent directional characteristics with the uncertainties known there can thus be omitted, since the bearing is also present with the detection.

0021With the advantageous embodiment according to claim 15, a time bearing plot is made, on which the detection of vehicles and their subsequent bearings are displayed.

0022The advantageous embodiment according to claim 16 also enables target separation in the case of crossing bearings and thus target tracking.

0023The embodiment of the device according to claim 17 is advantageous if optimal sensitivity of the receiver is guaranteed in a certain frequency range. Adapting the pass band's passband to this frequency range also has the advantage of interference suppression.

0024The devices according to the invention can be used in waterborne and airborne sound technology to locate ships, submarines, torpedoes in the water or helicopters in the air, since the propellers driving them cause this characteristic modulation of the radiated noise on which the detection method according to the invention is based.

0025The invention is illustrated on the basis of exemplary embodiments shown in the drawing for a device for detecting and carrying out the method according to the invention for vehicles.

0026Show it:<ul id="ul0001" list-style="none"><li>1 shows an example of a diagram of the time profile of a received noise,</li><li>2 shows a block diagram of the detection device for evaluating the received noise shown as an example in FIG. 1,</li><li>3 shows a circuit diagram of an integrator arrangement in the device according to FIG. 2,</li><li>4 shows a block diagram of a sonar reception system for directionally selective reception of noise in connection with a detection device according to FIG. 2,</li><li>5 shows a block diagram of a logic circuit for a detection display arrangement in the detection device according to FIG. 4,</li><li>6 and 7 each show a schematic representation of a display for the detection display arrangement in the detection device according to FIG. 4.</li></ul>

00271 shows a possible time course of a received noise, which contains ambient noise and operating noise emitted by a vehicle. Significantly, the time course of the envelope E stands out, which is almost constant within a period t 1 and assumes fluctuations during a period t 2, the frequency and amplitude of which increase after the period t 2. This received noise contains, for example, the operating noise of a submarine, which is during the span span t₁ on creep speed, i.e. behaves quietly and inconspicuously at low speed, then increases its speed and after the period of time t₂ runs at a higher speed level. With the method presented below, this noise is evaluated for the detection of vehicles. The time course of the envelope E of the received noise is determined and examined for fluctuations. Volume fluctuations in the form of amplitude modulation of a noise-like carrier only occur if the noise source not only emits stochastic components, but these components are caused or influenced by drive units and rotating parts, for example propellers. These noise components in the received noise are the significant detection criteria for vehicles. It can be seen in Fig. 1 that even with the noise level hardly changing, the time profile of the envelope E changes and is detected after the time period t 1 on the vehicle, since suddenly the significant fluctuations of the envelope occur.

00282 shows a block diagram of a device for the detection of vehicles which radiate operating noise in a broadband manner. With a hydrophone 10, noises according to FIG. 1 are received. The hydrophone 10 has the same sensitivity all around. The hydrophone 10 is followed by an envelope detector 12 with a low-pass filter 13 via a bandpass filter 11, at the output of which the time course of the envelope E appears. The bandpass filter 11 is only provided if the hydrophone 10 used does not have a constant sensitivity over the entire frequency range, but rather an increased sensitivity in a narrower frequency band. Then the pass band of the band pass 11 is tuned to this frequency band . The low-pass filter 13 is followed by a frequency analysis circuit 14, which contains a filter device or a calculation circuit 141 for determining the Fast Fourier transform. The frequency spectrum of the envelope E is determined in the frequency analysis circuit 14. The computing circuit 141 in the frequency analysis circuit 14 is followed by an integrator arrangement 15, the integration time of which comprises at least one time interval, this time interval being selected and set depending on the useful / interference ratio or a possible maximum speed of the vehicle to be detected.

0029In the frequency analysis circuit 14, the envelope E of the noise received by the hydrophone 10 according to FIG. 1 is examined with regard to its frequency content. The frequency spectrum at the output of the frequency analysis circuit 14 has, for example, only spectral lines during the period t 1, which lie at the frequency f = 0, since the envelope is almost constant here. During the period t₂ a frequency spectrum appears at the output of the frequency analysis circuit 14, which has a spectral line at least at a frequency f₁ and possibly at multiples of the frequency f₁. If several operating units of the vehicle emit noise, further spectral lines will appear in the frequency spectrum. After the period t₂, the frequency spectrum at the output of the frequency analysis circuit 14 has changed again. There appears one or more spectral lines at frequencies that are greater than f 1, for example at frequencies f 2 and f 3 and f₄, as shown in FIG. 2 in the display of a detection display unit 16.

0030The detection display unit 16, which is connected downstream of the frequency analysis circuit 14, contains a threshold circuit 17 and a display unit 18. In the threshold circuit 17, a detection threshold is set depending on the desired false alarm rate. All spectral lines that exceed this detection threshold are displayed on the display unit 18 depending on their frequency. To simplify the determination of whether the detection threshold has been exceeded, the frequency spectrum at the output of the frequency analysis circuit 14 is related to a reference spectrum, which is calculated in a computing circuit 19, in a quotient generator 20. In the arithmetic circuit 19, a reference spectrum is obtained from the frequency spectrum by folding the frequency spectrum at the output of the integrator circuit 15 with a rectangular function. Other methods of creating a reference spectrum are also possible. This signal processing makes the "colored" noise background in the frequency spectrum "white", ie, apart from the outstanding spectral lines that come from the operating noise, the rest of the frequency spectrum has the same, lower level. This signal processing is called normalization. The normalized spectral lines are compared in a threshold stage 21 downstream of the quotient former 20 with a set detection threshold and are displayed on the downstream display unit 18 if they exceed the detection threshold.

00313 shows a block diagram of an integrator arrangement 15 contained in the frequency analysis circuit 14, which is constructed as a stack memory. X_m(f) are the spectral lines as a function of the frequency at the output of the arithmetic circuit 141. In a multiplier 151, the individual spectral lines are divided by a stack number N, which can be set depending on the useful / interference ratio of the received noise. The multiplier 151 is followed by an adder 152, the second input of which is connected to a feedback branch from the output of the adder 152. The feedback branch consists of a delay stage 153 and a multiplier 154. In the delay stage 153, a storage time is set which corresponds to the time after which a new frequency spectrum appears at the output of the computing circuit. The frequency spectrum at the output of the delay stage 153 calculated in the previous computing cycle is multiplied 154 by the difference (1-<img file="EP0213541A2_D0001.tif" />) multiplied and added to the currently calculated spectral lines, which are divided by the factor N. This calculation process results in an integration of the spectral lines per frequency, the integration time is set by the number of stacks N.

0032The integrated or smoothed frequency spectrum at the output of the integrator arrangement 15 is denoted by S._m(f) named, where m is the current data set and (ml) the data set calculated in the last calculation cycle and m runs from 2 to N. The stacking then takes place according to the recursion formula:<maths id="math0001" num=""><img file="EP0213541A2_D0002.tif" /></maths>

00334 shows the basic structure of a panorama sonar system with a cylinder base 200 and a directional generator 201. Directional characteristic signals are formed in the directional generator 201 and further processed in a processing channel 202 implementing the method according to the invention. Each processing channel 202 consists of a band pass 11, an envelope detector 12, a low pass 13 and a frequency analysis circuit 14, as shown in FIG. 2. The processing channels 202 are each followed by a threshold circuit 17, which belongs to a detection display device 216. In the detection display device 216, the detection results for the individual directional characteristic signals are displayed on two display units 220 and 221 as a function of time. The threshold circuits 17 are constructed exactly as described in connection with FIG. 2.

0034A time bearing plot for an arbitrarily assumed example is shown on the display unit 220, in which bearings ϑ are shown over the time T. The display unit 220 is connected to an angle generator 222 and a clock generator 223 for providing signals for the horizontal deflection ϑ and vertical deflection T of the electron beam for the display screen. Its bright control input for the bright control of the electron beam is connected to the threshold circuits 217. At time T₁ three points are displayed under the bearings ϑ₁, ϑ₂ and ϑ₃, since bright control was carried out at the bearings ungen₁, ϑ₂ and ϑ₃ and three vehicles were detected via the threshold circuits 17. At time T₂ you get three points under the bearings ϑ₁, ϑ₂ and ϑ₄. The vehicle detected at time T₁ under the bearing ϑ₃ has meanwhile moved on with a tangential speed component with respect to the panorama sonar system, while the vehicles under the bearings ϑ₁ and ϑ₂ either rest or drive on a constant radial course. At the time T₃ only two points light up on the display unit 220, namely under the bearing ϑ₁ and ϑ₂. The vehicle that was last detected under bearing ϑ₄ has either continued on its course and is now in the same direction with respect to the panorama system as the vehicle that was detected under bearing ϑ₂, or it has switched off its operating units or it has switched to crawl speed and is no longer detectable by the panorama sonar system.

0035At time T₄ it can be seen that the vehicle detected under the bearing ϑ₃ has continued on its course, since a vehicle is now detected under the bearing ϑ₅ and the bearing ϑ₂ is retained. But there is also the possibility that the vehicle previously detected under the bearing ϑ₂ has changed course and is now recognizable under the bearing ϑ₅ and the other vehicle can be found under the bearing ϑ₂. In the area of ​​bearing ϑ₁ it can be seen at time T₄ that under the neighboring angle ϑ₇ a vehicle is detected which either previously had the same bearing together with the vehicle bearing under ϑ₁ or was first understood. This cannot be seen from the display unit 220. The display at time T₅ shows four vehicles under the bearings ϑ₁, ϑ₇, ϑ₆ and ϑ₂.

0036The open questions about the detection situation are solved with a logic circuit in the detection display unit 216, which is not shown in Fig. 4, but by way of example in Fig. 5 in connection with the threshold circuits 17 ϑ₁ and 17 ϑ₇ for neighboring directional characteristics, which provide bearing ungen₁ and ϑ₇ , is shown. At the outputs of the threshold circuits 17 ϑ₁, 17 ϑ₇ appear frequency spectra of the noises received with the associated directional characteristic, which exceed the detection threshold set in the threshold circuits 17 ϑ₁, 17ϑ₇. There are spectral lines of the frequencies f 1 and f 2 checked in this example as to whether the directional characteristics have a new understanding of the bearings ϑ 1 and ϑ₇ or whether, in order to stay with the detection example described above, more than one vehicle was detected at time T 3 and at time T₄ there is a target separation. A signal appears at the output H of the logic circuit 224 when a target separation has been recognized.

0037The logic circuit 224 consists of two AND gates 225, 226, the four inputs of which are connected to the outputs of the threshold circuits 17 ϑ₁, 17 ϑ₇, an adjoining OR gate 227, one arranged in parallel therewith, with the threshold circuits 17 ϑ₁, 17 ϑ₇ connected logic stage 228 and an AND gate 229 connected to OR gate 227 and logic stage 228.

0038For the detection example given at the time T₃ under the bearing ϑ₁ at the frequencies f₁ and f₂ spectral lines at the outputs A and B of the threshold circuit 17 ϑ₁ are found. These outputs have L signals. No spectral lines appear at the outputs C and D of the threshold circuit 17 ϑ₇, their outputs have O signals. An O signal then appears at the output E of the OR gate 227. The combination of the two AND gates 225, 226 and the OR gate 227 performs the following combination of the signals at their inputs A, B, C, D:<maths id="math0002" num=""><img file="EP0213541A2_D0003.tif" /></maths>

0039At time T₄, when the spectral line of frequency f₂ disappears at bearing ϑ₁ and is detected under bearing ϑ₇ and the other spectral line of frequency f₁ remains below bearing unter₁, an L signal appears at output E since the inputs A and C each have an L signal and inputs B and D have an O signal.

0040The logic stage 228 consists of two AND gates 2281, 2282, which only output an L signal at their outputs when an L signal is present at both inputs, a downstream NOR gate 2283 and a flip-flop which can be switched by the clock generator 223 2284. At time T₃, the AND gate 2281 has an L signal, the other AND gate 2282 has an O signal, since no spectral line was detected there. At the output of the downstream NOR gate 2283, an L signal appears, which prepares the subsequent flip-flop 2284 for an L signal at its output D. At time T₄, the NOR gate 2283 then outputs an O signal and the flip-flop 2284 tilts to the L signal. An L signal now appears at output H of logic circuit 224, which identifies the target separation at time T₄.

0041Here, the logic circuit 224 is given by way of example only for two spectral lines with two threshold circuits 17. It goes without saying that all other spectral lines at the output of the threshold circuits 17 in FIG. 4 are also examined in the detection display unit 216 in the same way by corresponding logic circuits.

0042Another type of representation of detection results is indicated in FIG. 4 by the display unit 221 in the detection display unit 116. On the screen of the display unit 221, five display screen units 230 to 234 arranged one below the other can be seen, on each of which frequencies f are displayed in a Cartesian coordinate system depending on the bearing ϑ, the brightness of the points appearing there being dependent on output signals of the threshold circuits 17.

0043All of the display screen units 230 to 234 are connected with their bright control input to the threshold circuits 17, with their input for the independent variable ϑ to the angle generator 222 and with their input for the dependent variable f to the frequency outputs of the frequency analysis circuits in the processing channels 202. The display screen units 230 to 234 are controlled one after the other by the clock generator 223. At time T₁ on the top display screen unit 230, the frequencies determined at time T₁ in the processing channels 202 are displayed above the bearings ϑ₁, ϑ₂ and ϑ₃. At time T₂, a frequency spectrum per processing channel 202 is again determined and displayed on the display screen unit 231 underneath. It can be seen that the three frequencies indicated under bearing ϑ₁ have remained unchanged, as has the frequency indicated under bearing ϑ₂, but that the two frequencies indicated under bearing ϑ₃ at time T₁ can now be found under bearing ϑ₄. A vehicle detected under the bearing ϑ₃ at time T₁ thus has a tangential speed component with respect to the panorama sonar system and can now be found under the bearing ϑ₄.

0044At time T₃, the image has changed again, as can be seen from the third display screen unit 232. The three frequencies under the bearing ϑ₁ and the one frequency under the bearing ϑ₂ have been preserved. Under the bearing ϑ₂ two frequencies have been added, which were detected at the time T₂ under the bearing angle ϑ₄. The vehicle continued to move at a constant speed on its course.

0045At the time T₄ one can see on the display screen unit 233 that two of the frequencies previously detected under the bearing nach₁ are still under the bearing ϑ₁, but the third frequency under the bearing ϑ₇ is detected. This shows that at least two vehicles have been displayed under the bearing ϑ₁. Target separation is now also possible here.

0046The frequencies shown at time T₂ under the bearings ϑ₂ and ϑ₄ can be found at time T₄ under the bearings ϑ₂ and ϑ₅. You can see that the bearings of the two vehicles have crossed.

0047The vehicle detected under the bearing ϑ₂ at time T₁ has a fixed bearing until time T₄, it rests or travels radially with respect to the panorama sonar system.

0048At time T₅ the following picture emerges: The two vehicles that were last bearing at the angles ϑ₁, ϑ₂ and ϑ₇ have kept their bearings. The vehicle that was displayed under bearing ϑ₅ at time T₄ continued to move on its course and can be found under bearing ϑ₆.

0049The same situation is shown on the display unit 220, only that the two vehicles, which were detected under the constant bearing ϑ₂ and the bearings ϑ₃, ϑ₄, ϑ₅ and ϑ₆, can only be separated here with the aid of the logic circuit 224. With the display unit 221 it is clearly visible on a screen that in the area around the bearing ϑ₂ two targets with crossing bearings were detected and in the area around the bearing ϑ₁ at time T₄ a target separation was possible.

0050Instead of displaying the time bearing plot in Cartesian coordinates, as on the display unit 220 in FIG. 4, the same situation can also be represented in polar coordinates in which the bearing Winkel is assigned to the angular deflection. Such a detection display is sketched in FIG. 6. At time T₁, a point appears under the bearing ϑ₁ on the display screen, which receives a circle after multiple confirmation in time, and from the time T₄ on the bearing erhält receives another point, which is provided with a train at time T₅ to indicate that that there was a target separation into two targets under the bearings ϑ₁ and ϑ₇. In addition, it can be seen from the display screen that a vehicle under the bearing ϑ₂ was detected from the time T₁ to the time T,, since there the target was provided with a circle, while a second vehicle which was detected at the time T₁ under the bearing ϑ₃ , the bearing ϑ₆ now shows, as the train shows on the display screen.

0051FIG. 7 shows a further possibility of displaying the time bearing plot, in which a polar coordinate display is selected. Bearing ϑ is associated with angular deflection and time T with radial deflection. It can also be seen here that at the angles ϑ₁ and ϑ₂ was detected with a fixed bearing. The target separation in the area of ​​the bearings ϑ₁, ϑ₇ at the time T₄ and the target crossing in the area of ​​the bearing ϑ₂ at the time T₃ can be seen on the display unit 220.