Surely the most spectacular recent advance in naval sensing has been the rise of unmanned vehicles – air, surface, and underwater – carrying sensors, whose output they can record or transmit back for analysis.
Moving sensors off board can greatly expand the footprint of a ship or submarine (or naval airplane). For example, to navies such as the United States Navy, submarines are invaluable for electronic reconnaissance because they are, in effect, invisible in most places. Thus the locals do not know that their signals are being intercepted, and they tend not to shut down equipment and communications. Since more and more of the world now relies on cell phones – on radios – for routine communication, reconnaissance can now scoop up far more information than in the past, when so much traffic was transmitted by landline. However, submarines are expensive, and they cannot be in more than one place at a time. A few years ago, the U.S. EDO Corporation displayed a drawing of an unmanned submarine-launched vehicle carrying intercept antennas. A submarine could, at least in theory, launch several of them and thus monitor a considerable area off a country’s coast. Moreover, the submarine, which was processing what they obtained, did not have to come close inshore.
Unmanned vehicles are to be the basis of the mission modules conceived for the littoral combat ship (LCS). For this purpose, they may exploit another advantage. In many cases it is vital that whatever carries a sensor has a carefully controlled signature. An anti-submarine ship, for example, should be as silent as possible. A minehunter should have the smallest possible magnetic and acoustic signatures, so that it does not set off mines it is trying to detect. Putting the sensors in devices that operate well away from the LCS itself dramatically reduces the need to control LCS signatures, hence considerably reduces the cost of the LCS itself. There is also another advantage, similar to that a submarine deploying unmanned underwater vehicles (UUVs) might enjoy. A ship with a sensor, such as minehunting sonar, in her hull can be in only one place at a time. She must go through a narrow swath of a minefield, one mine-like object at a time. However, there is no reason why the same ship cannot release multiple mine-detection UUVs that can separately examine swaths of the supposed minefield. The Royal Norwegian Navy has been operating the Hugin mine detection UUV from its minecraft for several years. Although typically a Norwegian minehunter employs only a single Hugin at a time, the concept certainly has the potential to support multiple ones. Mines detected by Hugin are to be destroyed by mine-killing mini-torpedoes sent to the positions of those mines.
Much the same might be said of an LCS operating multiple unmanned surface craft, each supporting its own sonar. A ship with a hull or towed sonar can be in only one place at a time, but multiple devices can cover a much wider area.
There are, of course, issues to be resolved. To make standoff sensing effective, the unmanned vehicle has to know where it is. That is not difficult for an unmanned surface craft, such as the projected anti-submarine module for the LCS. It is far more difficult for an unmanned underwater vehicle, which has no way of maintaining contact with, for example, the satellites that provide so much of the world with GPS navigation. A linked issue is propulsion. UUVs tend to be slow – but they operate for very long periods at that slow speed. The combination of power level and endurance is not easy to maintain, and probably new types of power plants, such as fuel cells, will ultimately be needed. If – a big if – a really compact plant offering high power and long endurance can be created, new kinds of underwater vehicles become possible, such as anti-submarine craft that might trail hostile or potentially hostile submarines out of their ports.
Unmanned surface craft certainly are easier to power, because they have access to air, but it is difficult to imagine operating craft small enough to be affordable (in numbers) in sea states that much larger ships find difficult. That may ultimately mean that even the surface ship sonar role must be filled by an underwater vehicle (i.e., one not subject to rough seas) – with all the problems of the underwater vehicle. Possibly a semi-submersible with a more or less permanent snorkel will be used. Alternatively, the problem may be seen as so difficult that instead of giving the LCS an off-board anti-submarine module, she will have an on-board processor and a towed sonar (to keep the sonar free of the motion of the LCS herself in a seaway). In that case, it is not clear to what extent the towed sonar will suffer from the noisiness of the LCS, which was accepted both because it seemed irrelevant (if separate unmanned sonars were used) and in order to keep LCS affordable.
There are also questions about how effective entirely unmanned sensors can be. Minehunting often requires a human operator’s judgment, to decide whether a possibly mine-like object is really a mine. In recent years, some manufacturers have claimed that advances in sensing and in processing what the sensor obtains have made automatic mine detection feasible. At the least, UUVs can conduct mine reconnaissance, determining that an area of interest does or does not appear to be mined. If there is enough sea room, and if an enemy’s mine stocks are limited, that may be enough to keep a fleet out of trouble. It is not, of course, enough to clear a mined area through which traffic must pass.
All of this is aside from the use of unmanned air vehicles to extend a ship’s horizon using existing kinds of sensors. Many navies are currently interested in unmanned shipboard helicopters for exactly that purpose, in some cases also with weapons on board.
The unmanned vehicles can also strew fixed sensors on the sea bottom, which may be particularly important in shallow littoral waters. Conventional sonar performs very poorly in such places, particularly when the water is warm. However, a simple array on the seabed, looking up, can function perfectly well, using what is called the reliable acoustic path (RAP). A submarine passing over such a sensor will be detected. A line of sensors can form a fence, and an array of fences strewn over the bottom will indicate a submarine’s course and speed. Since the entire system is passive, the submarine is unlikely to be aware of having been detected, and will probably continue on its course and speed. A system monitoring the arrays can use that data to predict the submarine’s future position, and to arrange an attack.
The U.S. Navy has been interested in exactly this approach to shallow-water anti-submarine warfare (ASW) for some time. At least in theory, the combination of unmanned vehicles planned for the LCS is the ideal way to distribute the arrays while making sure they are in known positions (so the tracking can work). Again, in theory, tracking and prediction can be precise, so an ASW weapon dropped on the submarine’s predicted position need not spend much of its time searching for its target. That is the basis of current U.S. interest in an ultra-lightweight 6.75-inch torpedo (which will also function as an anti-torpedo weapon).
For radars, the great development in the past decade has been the rise of relatively inexpensive active-array radars. New fighters already have active electronically scanned (radar) arrays (AESAs); this is the naval equivalent. In many applications active arrays are superseding passive ones like the antenna array of the U.S. SPY-1 radar used in the Aegis system. Both kinds of array consist of a large number of separate elements, each of which transmits and receives radar signals. Each small element creates a broad beam, which by itself provides little or no information on the direction to a radar target. However, if the signals passing through the elements are properly timed in relation to each other, they add up into a radar beam. Since the timing (phasing) can be computer-controlled, the direction of that beam can also be controlled. Instead of swinging about as a mechanically scanned antenna turns, the beam moves as desired. In SPY-1, for example, the radar registers a target its beam detects. It then quickly generates beams around the last known target position, to find the next target position and hence to calculate target speed and course.
In the Aegis radar, the radar signal, with its pulse structure, is generated by a single power tube deep in the ship. The elements in the array apply the desired timing. The array is called passive because it does not create the signal. The radar creates a single, powerful radar beam that can be maneuvered electronically to detect fast-moving targets. However, that single power tube cannot create multiple beams, doing different things, simultaneously.
The next step, the active array, consists of elements that produce their own signals. As in the passive array, they receive timing commands, but unlike the passive array, an active one does not have a single source of radar power. Instead, normal electric power is fed in, and radar signals emerge. Such an array can create a number of different beams, with different signal structures, simultaneously. It can also be set to ignore (null out) a jamming signal, something a passive array apparently finds much more difficult.
An active array also offers the advantage of scalability: It is much easier to make larger or smaller than a passive array. To make the array more powerful, the producer has merely to add more panels. A passive array can certainly add elements, but in order to add power it has to change power tubes, because that is where the power originates. This kind of scalability may be of particular importance for ships intended for ballistic missile defense. Dealing with such distant targets requires both a different waveform (which is not a great problem) and a great deal more power in the beam.
There are, to be sure, problems to overcome. First, the active array produces heat, which has to be dissipated. Thus when the Japanese Maritime Self-Defense Force adopted a rotating (single-face) active array some years ago, it had an attached radiator. Photographs of the demonstration version of the current Chinese active array showed water pipes; indeed the presence of the pipes helped indicate that the radar was an active rather than a passive array.
Second, the different elements of the array had better be precisely matched, or the beam will not point in the desired direction. For example, all of them had better radiate at exactly the same power, pulsing in the same way.
A third problem is building small enough array elements (transmit/receive elements) at the desired frequency. The higher the frequency, the shorter the wavelength, and the easier to build usable elements. That is why active fighter arrays, which operate at higher frequencies, are more common than warship active arrays. For a ship, lower frequency is associated with longer range. The multi-function active phased array radar (APAR) arrays on Dutch and German frigates, for example, operate at higher frequency than SPY-1, which is why these ships (but not most U.S. Aegis ships) have separate long-range target detection radars. The British Sampson, on Type 45 destroyers, does operate at SPY-1 frequencies, but its elements are massive enough that instead of using fixed arrays it uses a rotating two-faced array.
Raytheon’s SPY-3 is to be the first U.S. active-array radar. It was conceived for the Zumwalt-class missile destroyer and was successfully tested at sea in May 2006. These ships were designed to combine several radars, operating at different frequencies, with a single back-end processor. Thus the short-wavelength SPY-3 was intended to work with Lockheed Martin’s longer-wave active-array Volume Search Radar. The latter was later canceled to reduce the cost of the Zumwalt program, but a new Air and Missile Defense Radar (AMDR) is planned. It may revive the earlier combination.
Several manufacturers in different countries are now working on active arrays. Two examples seem to be worth mentioning. Israel Aircraft Industries (IAI) developed a four-face fixed active array specifically to upgrade the Israeli Eilat-class corvettes. The radar also found an export customer. Like many array radars, it uses a configuration in which the individual transmit-receive elements are grouped, the groups receiving computer commands. When Israeli patrol boats began to suffer attacks from anti-tank missiles, IAI was able to provide a quick solution because each of the groups could function as a full-blown radar, albeit with a rather broad beam. There was obviously no great hope of shooting down the approaching missile, given limited time, but if the missile could be detected in time the boat could launch decoys and evade. IAI therefore produced a four-face radar, each face of which was a single group taken from the larger radar it had originally developed. This was downward scalability.
Thales Naval Nederland (formerly Signaal), the Dutch arm of the multinational Thales Group, developed the multifunction active radar operated by the Dutch and German navies some years ago. It proposed a single-face version for patrol ships. However, it has actually sold something different: a fixed mast with four radar array faces. The mast also carries higher-frequency arrays for navigation and surface search, electronic search sensors, and communications arrays. This arrangement solved a serious real estate problem. Normally the performance of all the electronics on board a surface ship is limited because superstructure elements, carrying other electronic arrays, block each array in some direction. Even if blockage is not obvious, signals reflect off the metal in the ship in complex ways, which cost performance. For years the U.S. Navy built expensive copper models of its ships specifically to measure and to solve, if possible, interference problems. The single mast appears on the new Dutch Holland-class patrol ships.
There are also attempts to cut the cost of active arrays. The most prominent seems to be the Australian CEA Technologies company, whose tile arrays are currently being installed on board ANZAC-class frigates in a kind of mini-Aegis configuration.
Another important line of radar development has been attempts to develop stealthy radar. There are two complementary requirements. One is to make the radar signal itself difficult to detect. The other is to make it difficult for an enemy radar to pick up reflections from the radar antenna, which (in a non-phased array) must be a reflector focusing radar energy for itself.
For the first requirement, the trick is to stretch out the radar signal. The most extreme example is the Thales Scout, a surface-search radar that operates, in effect, at FM rather than AM – it radiates continuously at a low power level, changing its frequency. The broad-band receiver associates frequency with time the way a normal radar associates the pulse of radar energy with time (hence range). Many radars use a less extreme version of this technique – pulse compression. Search receivers generally require energy above a threshold to trigger them. Stealthy radar works because radar detection depends not on the power at any one moment (peak power) but on average radar power, which is much lower. Scout in effect operates at average power all the time; pulse compression radars produce more power, but not as much as simpler pure-pulse sets.
Radars whose antennas are difficult to detect are a more recent development. It is possible to tune materials so that they pass only signals within a narrow bandwidth. If a mast is tuned to the frequencies at which the radar turning within it operates, that radar can look out but nearly all radars looking in are blocked by the mast. For example, the new Franco-Italian FREMM frigate will have its main radar inside its stealthy mast. Other navies have followed much the same approach.
The other side of radar is the interception of radar signals – electronic support. Probably the most important development is the ability to recognize – to fingerprint – particular radars, rather than radar types. At least in theory, a radar’s fingerprints ought to be related not only to the kind of signal produced, but to physical features such as nicks in a waveguide. Fingerprinting seems increasingly important in a world in which many navies (or others) use either the same type of radar or very closely related types. It also deals with an expanding problem. At one time radar signals were all generated by specialized tubes, such as magnetrons, whose dimensions and shapes determined the sort of pulses they produced. It therefore made excellent sense to associate particular patterns of pulses with particular radars. However, many modern radars use computer software to create their waveforms, which are then amplified by broadband devices such as traveling wave tubes. This sort of operation makes radar inherently multipurpose, with different kinds of signals for different purposes. The first such radar in naval service seems to have been the AWG-9 of the F-14 Tomcat, which had both air-to-air and surface-attack modes. When a French-built Iraqi Mirage attacked the USS Stark using a surface-attack radar, her electronic warfare operator thought the radar signal he saw had come from an Iranian F-14 operating in surface-attack (or search) mode. At that time, 1987, the flexibility of the AWG-9 radar was quite exotic. It no longer is. Software-controlled waveforms make it far more difficult for anyone to rely on a dictionary of radars embedded in an electronic countermeasures system. That is aside from the problem of selecting appropriate jamming signals.
For sonars, the most important lesson the U.S. Navy learned after the Cold War was that it could dramatically improve performance by upgrading processing. At the end of the Cold War, U.S. submarines were losing their edge against their Soviet rivals. The Russians were finally effectively silencing their submarines. The normal reaction would have been to develop a new generation of sonars, with larger (for higher gain) new arrays. That began with the Seawolf class, which had a massive new bow sonar and a new flank sonar. With the end of the Cold War, sonar funding collapsed; no massive new sonar was likely to appear. The new Virginia-class attack submarine was smaller than a Seawolf, hence could not accommodate its new bow array (it could be fitted with the flank sonar).
Analysis suggested a way out. Existing sonar processing and post-processing (the stage at which sonar data is assembled to form meaningful patterns) was relatively primitive, often using fairly old computers. U.S. civilian computers were far more powerful, and they were developing very rapidly. The U.S. Navy mounted a new program of Acoustic Rapid COTS (commercial off-the-shelf) Insertion, A-RCI. Submarine sonars’ data were fed into a new fiber-optic bus that could be connected to enclosures in which commercial computers were installed. The A-RCI program envisaged a rolling program of hardware and software upgrades: new hardware every four years, new software every two. The hardware would not be “state-of-the-art,” because that was not yet reliable. Instead, it was “state of the practice.” Existing enclosures could be re-used again and again because new computers were generally smaller than their predecessors. A-RCI has proven remarkably effective; the edge has definitely been reclaimed and substantially extended. The idea has been extended throughout the U.S. fleet.
As submarines operated more in littoral areas, moreover, it became clear that the classic division between acoustics (sonar) and other sensors was unproductive. For example, a submarine hearing a surface ship might best identify that ship using her periscope or her electronic search receiver. In the Virginia class, all the ship’s sensors are tied to the same fiber-optic bus, their outputs available at all the combat system workstations. This development in turn was tied to the adoption of electro-optic periscopes in place of the usual purely optical ones.
An electro-optical periscope replaces the human eye of the operator with a camera or, more usually, with several cameras, including infrared ones. Instead of standing at the eyepiece, the observer uses a computer console. From a submarine designer’s point of view, the most dramatic effect of this development is that the periscope is no longer a tube that must be led into the submarine’s attack center. Thus the attack center need no longer be in the upper part of the hull directly under the submarine’s sail. In the Virginia class, it is lower in the hull, where the hull is wider and there is much more space. For that matter, the periscope no longer must penetrate the submarine’s pressure hull; only a cable need do that. Hull penetrations such as periscope openings have been a major structural problem in the past, limiting submarine diving depth, for example.
Perhaps the operational effect of such periscopes is even more radical. Exposing a periscope endangers a submarine, so in the past, submarine officers were trained to make a quick scan and instantly to understand the tactical situation. An electro-optical head can make a quicker scan but it can also send what it gets to a workstation at which the image is captured for review and analysis – and for comparison with other information the submarine may be collecting, such as that from sonars and electronic search sensors. It may even be possible to place the electro-optical sensor in a pod that floats off from the submarine, so that an enemy seeing it cannot know exactly where the submarine is.
The modern emphasis on relatively shallow-water operations against quiet diesel submarines has changed sonar. Although they may be quiet, nuclear submarines necessarily run their machinery constantly. For decades, NATO anti-submarine operations were based on using passive sonar, which was designed to pick up the constant sounds of Soviet nuclear submarines against the random shifting sounds of the sea. A diesel submarine is very different, because it has three distinct modes of operation: snorkeling on diesel power, running on its battery, and sitting on the bottom (a nuclear submarine will not bottom because it runs the risk of sucking mud into its condenser and thus being put out of action). The diesel submarine cannot be altogether silent, but on battery it is probably as quiet as a very quiet nuclear submarine, with much less regular noise to recognize. On the bottom it can indeed be silent.
The obvious answer is active sonar: If the submarine does not emit sound, then the hunter can produce a sound to echo off it. However, a submarine may well locate the pinger before she is located, and she may be able to attack before she is detected. Hence considerable interest in recent years in different forms of standoff active sonar. Initially that often meant using a small explosive charge to create the ping and existing passive arrays or sonobuoys as receivers, a technique called Enhanced Echo-Ranging or Explosive Echo-Ranging. It recalls a Cold War technique called Julie, adopted when it seemed that the Soviets would trump passive sonars by silencing. In modern form, it uses multiple receivers and powerful computer processing to solve problems like reflection off the bottom. This technique can produce an image of bottom topography good enough to show a bottomed submarine. The receivers may be towed arrays on board ships or they may be sonobuoys; the principle is the same in either case.
Small explosions are useful, but their sounds are somewhat irregular. The next step is a more controlled sound source, an example being Ultra’s ALFEA.
Another way to use active sonar is to ping from a distance at very low frequency, a method the U.S. Navy has used successfully from a sonar surveillance ship. Powerful low-frequency signals can disrupt marine mammals, so at present the U.S. Navy finds itself limited in where it can train with the system. However, several NATO navies are experimenting with their own low-frequency (but higher frequency than the U.S. Navy’s) systems, such as the Royal Navy’s Type 2087.
This article first appeared in the Defense, Spring 2011 Edition.