When Computers Went To Sea

Militay Technology 2007년 5월호에 실린 글입니다. ... 표시는 생략된 부분을 의미합니다.   1. 얕은 바다에서도, 저주파 소나도 살 길이 있다. 빔 조향~ 하여튼 얕은 바다에서는 고~저주파 모두 동원해서 액티브 핑 때리는 것이 장땡.   2. 호위함이든 잠수함이든 이제 혼자서만은 안되고 이제 곳곳에 센서를 뿌려 놓거나 무인플랫폼 조종해서 주변을 휘젓는 것이 대세 - 초계기나 항모 닮아 가나?   3. 음향측정용 폭탄 - 지난 역사를 공부해두는 것은 역시 필요하다. 꺼진 불도 다시 봐야 하기에...   4. 신호 처리에 백지이다 보니 펄스 압축도 이해를 못해서 불통인 부분도 있습니다. @.@   5. 오타는 알아서 소해 작업을... ㅠ.ㅠ   Sonar Technology Review - Where We are, Where We Are Likely To Be Going   The Impact of Computer Revolution   The Single most dramatic technological development of the last decades has been the explosion of computer power, the typical estimate(Moore's Law) being that the power available in a chip doubles every 18 months. Increased power makes it possible to use more complex waveforms and to treat each element, or at least each stave, in a sonar on a more individual basis. At the same time the volume and cost of electronics needed to support a given type of sonar ahve declined dramatically, bringing lower-frequency sonars within the financial grasp of many more navies. Computing power also probably drastically reduces teh cost of the outputs of the many elements of a towed linear array, and it makes longer arrays practicable.   Much better signal processing clearly improves sonar performance. It also makes possible a kind of supplemental performance, because given such processing it is now possible to operate acoustic data links over substantial distances to pass sensor outputs and even imagery. At present most such transmissions are overt, but it is possible to transmit at low rates(about 100bits/sec, in future up to 1,000 bits/sec) covertly. ... Such communication would make it possible for multiple underwater platforms, such as submarines and UUVs, to exchange data to form a network-centric underwater tactical picture, with immense implications. The ability to co-ordinate several platforms, and to share their sensor data, would seem to dwarf any other near-term sonar development. It should be emphasised that, at least for the US Navy, this is a current or very neat-term proposition.   Another major development enabled by expanded processing power has been a redesign of submarine combat systems to that data from all the ship's sensors ― non-acoustic as well as acoustic ― is fed into a common fiber optic bus for processing by the ship's combat system. In the past, non-acoustic sensors such as radar and ESM were handled as supplements to the primary ship's sensors, her sonars, and the same would have applied to the fruits of UUVs and to data obtained over underwater data links. Now such data to help create a tactical picture for the command.   But paradoxically enough, quite aside from the question of added performance, there also are significant netative aspects. The very speed of development in computer technology makes it more and more difficlut to maintain older electronic systems, or thost incorporating components not currently in the civilian market. Even older civilian computer chips often cannot be obtained. This is of course a problen far beyond the sonar arena. The usual solution is described as adoption of open architecture, and some firms ahve prospered by providing a hardware and software path ot upgraded chips. Even so, the rapid continuing advance of computer hardware is difficult to reconcile with ships and aircraft intened ot operate for several decades. For example, at some point each family of chips, upgrades within which are relatively straightforward, is abandoned althgether in favour of chips which cannot easily accepts software wrieen for the earlier family. Such developments have real consequences, as demonstrated e.g. by the problems that plagued the combat direction system of the Royal Austrailian Navy's COLLINS-class submarines.   US developments have dramatised this trend. The most important sonar development programme, both for submarines and for surface ships, is en effort to insert commercial computing technology as rapidly as possible, with little or no reference to changing hull sonar arrays. The object of grastic inernal modifications under programmes such as the submariners' A-RCI(Acoustic Rapid COTS Insertion) is to keep changing signal processing and other computing hardware to keep up with the current state of practice in the civilian industry, on the theory that new computers must be purchased every two years. Machines are becoming so reliable that they can be used throughout a submarine patrol without any maintenance.   ... However, for the moment the US Navy has clearly grasped the central fact that the greatest payoffs in sonar technology are to be had in improved signal processing ― which may include using new waveforms ― rather than in buying new arrays.   The only important exception to this reality is that navies may choose to buy new kinds of arrays offering dramatic improvements in acoustic performance. The main current examples are lower-frequency active sonars using array receivers, the ships' hulls being too small to accommodate the active arrays. Some current interest in array ASW combines dramatically better signal processing with new array types.   Evolving Requirements   Over the last twenty years a radically changibng strategiic situation has transformed sonar requirements and systems. During the Cold War, the main threat to Western Navies was presented by Soviet nuclear submarines. Perhaps the single most important Cold War sonar discovery was that such submarines could often be detected passively, because most often they had to keep noise-making machinery running simply to keep their reactors from melting down. Accordingly, narrow-band acoustics(LOFAR) became the single key NATO sonar technique. Although the Soviets did eventually learn to silence their submarines, by the end of the Cold War Western passive sonars relying on "non-traditional" sound signatures were apparently redressing the balance in NATO's favour. The hallmark of this period was the development of a wide range of passive sonars, including towed arrays and passive sonobuoys.   NATO learned, moreover, to use a form of network-centric ASW, in which long-range passive arrays of the SOSUS system detected submarines which were then picked up by maritime patrol aircraft. In effect, te aircraft were viable because the submarines could be detected remotely. Here network-centric means that the operators rely on a tactical picture created by a network of remote sensors(which may include them). It is contrasted with the more conventional platform-centric approach, in which tactical decisions rely overwhelmingly on the platform's own sensors. One major advantage of separating detection from prosecution is that the victim is generally taken by surprise. ...   On the other hand, long-range detection is not always possible; SOSUS always had a percentage probability of finding a submarine on any given day, and its effect would have been cumulative. Through the Cold War, opinion varied dramatically as to whether remote sensing should or could be the primary means of detectin submarines. Several European NATO navies focused on ASW worhsips which would, in wartime, have escorted convoys, whereas the US Navy and, to a somewhat lesser extent the Royal Navy emphasised SOSUS and such prosecution tools as long-range MP aircraft and nuclear attack submarines. for example, the British Type 23("Duke" class) frigates were apparently conceived as platforms to deploy towed passive arrays in the Greenalnd-iceland-UK Gap, mainly as a means of cueing long-range aircraft and SSNs.   Now the world had been transformed, at least for the time being. The most likelyh venue of Western naval acitivity is the Third World, where the submarine threat in any conflict will be small numbers of diesel-electric(plus possibly AIP) boats. This main threat has characteristis very different from those of a nuclear submarine. Diesel submarines are not always inherently quiet, but unlike a nuclear boat they have multiple operating modes, some of which are extremely quiet. If they operate in waters shallower than crush depth, which generally means anything less than 1,000ft or so, they can sit on the bottom, making no noise at all because they are not running their engings, yet quite capable of detecting surface units above them. How often a diesel submarine must risk passive detection by running her deisels depends on her battery capacity; the new air-independent propulsion(AIP) units can prolong the period between snorkelling to weeks.   Moreover, when a diesel submarine does make noise, it may be at frequencies below those typical of nuclear submarines. For example, such submarines may produce blade-rate noise. A five-baladed propeller turning at, say, 180rpm produces a lowest harmonic at only 15Hz. This noise modulates the flow noise produced as the submarines passes through the water. Because flow noise is inherently broadmand, it does not lend itself to narrowband detection using techniques like LOFAR. It is, however, possible that some alternative basis ofr analysis, such as the use of wavelet functions, will be effective, and that greatly increased computer power makes such analysis practicable.   Water conditions are also likely to be radically different from those of the past. The Third World is overwhelmingly shallower than, say, the central Atlatic or the Norwegian Sea. Bottom topography may often be comples, making it easier ofr a submarine commander familiar with it to hide from searchers. In many places the surface ducts quite shallow, so hull sonars are likely to be effective mainly in bottom bounce mode, with all the difficulits that entails. In shallow waters the convergence zones which made ofr extremely long sonar ranges during the Cold War may be assent. Moreover, a force approaching a Thirld World operating area may be confronted with local conditions, such as river outlets, with which its sonar operators are unfamiliar.   Consequnces have included a dramatic decline in interest in towed passive arrays, at least for major surface units. Later US Navy ARLEIGH BURKE-class destroyers lack the SQR-19 towed array of the earlier ones. In the Royal Navy, the Type 2031 array of the "Duke" class has been superseded by the low-frequency Type 2087 active sonar.   One other point is worth haking here. It is often suggested that poor sonar conditions are a double-edged sword, in that they limit a submarine's ability to target surface units passively. The sumarine is therefore forced to use her periscope and perhaps even her radar, in which case a pericsope-detection radar becomes a vital ASW sensor. That is true as long as the submarine operates entirely independently, which would probably be th case in the open ocean. However, in an inshore area the submarine might be albe to rely on the outputs of external sensors to reach a position to attack with long-range homing weapons. Information might well be tansmitted to the submarine by LF or even VHF radio, which penetrates the water surface.   The Return of Active Sonars   The effective silencing of diesel submarines has made for a resurgence of interst in active sonars. In the past, they have suffered from two great drawbakcs. One is that a submarine with the usual acoustic intercetp set can use the pulses she receives to localise thier source, either for direct attack or to map out the escorts around a high-value unit. The other great problem is that a shllow bottom limits effective range. In calm conditions, moreover, sonar pulses may bounce off the bottom and the surface, so the longer the effective range of the sonar, the more it can be confused by multipath effects. That is, the single pulse which the sonar emits splits into sub-pulses following a variety of path, which take slightly different times. The receiver is confused, as the results from the different paths interfere constructively and destructively, sometimes wiping out indications that a target is present. The longer the pulses, which normally means the greater thier effective range, the greater the effect. Only very short pulses, which normally would not go very far, can be distinguished, because they are shorter than the the differences between pulses following the different paths.   The main solution th this problem is pulse coding. As in radar, a long pulse(for range) can be compressed into, in effect, a very short one if the pulse frequency (or even phase) changes as the pulse is prodeced. In radar this concept is pulse compression; in sonar it is frequency modulation(FM), and the wider tha bandwidth, the narrower the equivalent pulse. Pulse compression is an antidote to the multipath problem. In some cases it can also be used to help identify an underwater object by, in effect, imaging its elements. Current hull sonars such as the upgraded form of Raytheon's DE 1160 or the US Navy's SQS-53 use pulse compression, and can also reduce multi-path effects by detailed control of the reception beam(s) both in elevation and in train. However, it seems possible that the most important application of pulse compression is to much lower frequency towed sonars, such as that in the US sonar surveillance ship(T-AGOS, with Low Frequency Adjunct), or the big active towed sonars(with linear array receivers) now being tested by several NATO naives.   As indicated above, the lower the frequnecy, the longer the sonar range. The lower the frequency, too, the larger the transducer which produces the signals. Below about 3.5k Hz, transducers are clearly too large for conventional hull installation; they can, however, be towed. The Low Frequency Adjunct operates at 100 to 500Hz, i.e. just above the typical cold War passive range - but Cold War passive systems often used harmonics, which would be multiples of this frequnecy. A really large low-frequnecy sonar can "ensonify" a whole ocean basin, such as the Eastern Mediterranean, and in effect it can map all underwater acitivity there. Submarines can be prosecuted by ships or aircraft vectored on the basis of such a map, just as during the Cold War, P-3s were vectored by the passive indication collected by SOSUS. The most dramatic difference is that, because the big sonars are active, they are far less likely to miss targets(they may miss bottomed submarines, but there are reports that even they are typically detected). One or two big active-array ships thus offer a real possibility of neutralising Thirld World submarine fleets, while operating a few hundres miles off-shore, presumably safe from counter-attack.   The main reason such sonars are not widely used is a question raised by environmental groups. They aruge that the sonars operate at frequnecies which affect marine mammals such as whales; they sometimes claim that incidents of whales making suicidal groundings are due to confusion generated by low-frequency sounds. At least in the United States, naval scientists have rejected such arguments. A compromise solution in the United States allowed the Navy to operate low-frequency arrays in peacetime only in a restricted area. The problem with such a limitation is that the boundary between peace time and wartime is less and less well-defined; there are areas where the US Navy will very badly want to keep track of potentially hostile submarine activity, not least as a war warning measure.   A Question of Frequnecies   ... Too, towing the sonar (and the receiver) decouples the sonar from hull motion, so that a relatively small ship can benefit from low-frequency operation.   Current examples of such sonars are the Royal Navy's Type 2087(by TUS), the EDO 980 adopted for the Singaporean FORMIDABLE-class frigates, and the BAE/Thales ATAS adopted by the Pakistan and Taiwan navies. For its new ZUMWALT-class DDGs, the US Navy is buying a new bow sonar and a low frequency wideband VDS(LBVDS), with provision for the two to operate bi-statically as the sonar portions of the Integrated Undersea Warfare System(IUWS). The unique configuration of the IUWS sonar incorporates both high frequency(HF) and medium frequency(MF) sonars mounted in close proximity to each other. The HF sonar is designed for in-stride mine avoidance, while the MF sonar is predominantly for anti-submarine and torpedo defence operations.   If the water is not too shallow, low-frequnecy signals can travel great distances, particularly if the angles at which they graze surface and bottom are small. The key is the ability to control the elevation of the sonar beam, and to limit its vertical extent. Presumably a towed "pinger" with a vertical array of transducers, such as that of the British Type 2087 offers just such a potentials. At least in theory, such operation offers considerable potential to low-frequency helipcopter dipping sonars. The two most advanced current models arguably Thales' FLASH(in use by, among others, The French, US, and Royal Navies) and L3's HELRAS(adopted by, among others, Germany, Greece, Italy, the Netherlands, and Turkey) Both have a potential ability to steer their beams in the vertical and thus to create the desired nearly horizontal long-range beam. At least in the case of FLASH, the beam elevation was originally set to provide bottom-bounce performance, but the sonar can easily enough be modified. These sonars operate at much the same frequencies as the lowest-frequency current hull sonars(e.g. SQS-53).   ... The longest claimed active range for such a set is the two convergence zones reported for SQS-53C in deep water. Current development, moreover, seems to be directed towards providing more sophisticated wave forms rather than towards adding power - which is, in any case, limited by what can be transmitted without causing cavitaion.   E2R Systems   Explosive echo-ranging or E2R is another kind of low-frequnecy technology. JULIE was conceived by a US Navy developmental squadron(at Wanninster, Pennsylvania) in the mid 1950s as a way of overcoming a perceived threat to the passive sonobuoy system then in service, the silencinf of diesel submarines. ... The corresponding sonobuoy technique was to set off practice edpth bombs, whose sound would bounce off the submarine. As it turned out, in anything but very deep water, eht echo of the bottom would swamp any echo prodeced by the submarine. Thus, JULIE could be used only in water more than 8,000ft deep, which would hardly apply to most current Third World conditions. In the 1970s-1980s the Soviets developed a form of JULIE using elaborate arrays of explosives to create beams which overcame the bottom problem.   As originally developed, JULIE entailed no signal processing: the operator simply compared the time lags between the explosion and its first two large detections, one from the direct path between explosion and sensor, the other including reflection off the target. As computing power exploded in 1990s, the question was whether a more detailed examination of the time trace of the echoes from the submarine and those from the bottom. ...   The US DARPA research agency sponsored a project, DISTANT THUNDER, in which several arrays(either shipboard linear ones or sonobuoys) received signals from an explosion, and a central processor(at that time, roughly the size of a shoebox) compared the pressure traces at each in detail. At the time, it took about 20 min. for the processor to deduce a detailed map to bottom topography - including any bottomed submarines - from the combination of traces. ... DISTANT THUNDER was incorparated on board US destroyers operationg off Korea, where the water is shallow and the submarines are diesels, and it is part of recent version of the SQQ-89 undersea warfare combat direction system. DISTANT THUNDER is, incidentally, a network-centric as opposed to platform-centric approach to inshore ASW, and as such it may be a pointer to the future. The US Navy calls the revived JULIE techniques Enhanced Echo Ranging, rather than the Explosive Echo Ranging of the past.   What About Submarine Sonars?   The same concept clearly applies to submarines. At present virtually all submarine sonar are passive, because the penalty ofr active pinging is quick detection and almost certain attack. ... Diesel submarines present the same problems to existing submarines that they do to surface ships and maritime aircraft: their signatures vary, and at times they can be extraordicarily quiet. What to do?   Given the echo-ranging techniques of the past, one possiblility is to use a UUV to ping, in effect to illuminate the underwater area. Oppoents of these techniques argue that any such illumination is as likely to illuminate the searching submarine as well as the potential victim, and of both are similarly equipped perhaps it is not at all cleat that illumination is worthwhile. Moreover, it can be argued that a Western nuclear attack submarine presents a far larger echoing area than a Third World diesel craft, hence that a source of illumination may offer greater advantages to the smaller submarines.   However, the more sophiscated Western units can be given soma important opportunities to even the odds. One is that they can deploy a series of sensors on board UUVs, with the target being localised by the joint attentions of the dispersed sensing field. Another is that the big submarines has much more signal-processing power. Coding the transmission by the UUV may make it far eaiser ofr the large boat than for a smaller one(which will not, in any case, have the appropriate template for detection) to detect echoes.   Problems in the Littorals   Overall, littoral conditions limit sonar range. To start with, the noise level in the littoral is usually quite high. Sources include heavy inshore shipping, pleasure craft, and even many industrial installations. In the past, the typical solution to a submarine threat has been screening, the object being to present the submarine with a barrier which it must breach before it can attack its targets. The shorter the effective sonar range, the more anti-submarine ships are needed to create a viable or even credible barrier. ...   However, the one obvious trend in modern Wetsern navies is the steady decline in numbers as ships become more expensive. For the US Navy in particular, the question is how to maintain an effective sea base off a hostile shore. Hostile submarines probably will be albe to find the sea base, simply because it will be operated so close to enemy reconnaissance. How can they be prevented from disrupting it? The current US approach recalls both the Cold War SOSUS concept and the newer DISTANT THUNDER. It recognises that the reason large numbers of ASW craft are demanded is that large numbers of sensors are needed to make up for their limited ranges. The sensors themshelves need not be attached permanently to ships; it is enough if the ships(or the Fleet Command) should be able monitor the sensors so as to detect approaching submarines. ...   The necessary sensors are now being tested. for shallow water, the most important are probably small linear arrays which can be laid on the bottom, in some cases adding magnetic sonsor. The new Littoral Combat Ship will distribute then and monitor then; probably it will also support the craft, likely unmanned, which will deliver weapons to attack the submarines they detect. One interesting feature of such a system would be that the surveillance element would locate the submarine well enought that the attacking weapon would not have to be very large. In one formulation of the US programme, the same 6.75"-dia. body could be used both to prorecute submarines and to attack inbound torpedoes, though presumably with different innards.   Future Possibilities   Obviously even a net of sensors will not detect every submarines; no sonar system in history has been perfect. ... Thus the likely complement to the sensors will be shipboard torpedo defences, including active ones. It also seems likely that sensors detecting the firing or run of a torpedo will be used for quick localisation of the firing submarine, so that counter-attack will be almost instantenous. Whether or not such sensing killed many submarines, surely the knowledge that any attack would result in a quick counter-attack would tend to discourage many submarine commanders.   The US programme to develop unmanned underwater vehicles offers another possibility. A UUV might be developed to approach enemy or potentially enemy submarines. If it had enough endurance, it might simply trail them. Upon the outbreak of hostilities, the UUV might provide won distant submarines or other forces with sufficiently presice information to attack the hostile boats. In either case, the key would be that the UUV would operate very close to the target submarine, so it would not need long range sonar capability. Another possibility of current interest would be for the UUV to tag the submarine, perhaps with a device incorporating a sonar transponder or a coded reflector. In that case the submarine would lose her effective stealth. If the sonar transmission were sufficiently coded, the target might not even be aware that she was being tracked. ...   Clearly the distributed bottom sensors would not be effective in an open-ocean context, if only because the units to be protected wold soon move out of their tange. However, it is still possible to imagine that sensors could be distributed around a force in the form of sonobuoys. That idea was tested in the 1970s int he context of the abortive US Sea Control Ship, whereby her VSTOL fighters would have spread sonobuoys ahead of a force to make it possible to form a sanitised corridor. Now, with very copact sonobuoys availalbe, small helicopters operation from Littoral Combat Ships can, presumably, strew enough expendable sensors around a moving force to detect submarines over a wide area. That the force is moving probalby compels the submarines, moreover, to move and so to generate flow noise, which the sensors can detect (that may not be the case in shallow water). A moving DISTANT THUNDER carpet would, presumably, even catch submarines hiding on the bottom in shallow water.   This sort of area approach to ASW, relying very heavily on signal processing and on large numbers of inexpensive point sensors, seems to be the most likely Western response to the drastically reduced effective sonar range of many Third World areas.