Active-passive bistatic surveillance for long range air defense.
From the UK Chain-Home (CH) network at the outbreak of World War 11 to the current NATO inventory, radar has continued to offer the most effective means of air defense (AD) surveillance, yielding information essential to efficient use of combat air patrols and other defense resources. Denial or degradation of this information is a priority for any air attack, while the performance of the AD network, in the face of vigorous hostile countermeasures, is crucial to the defense.
The importance of functional survivability was recognized from the start. The UK CH network constituted a chain of transmitting stations pulsing in staggered synchronism, each with a separate receiving station able to operate with a more distant partner if its local transmitter was attacked or jammed. Therefore, supplementary bistatic operation was an established ECCM technique even before the term radar had been coined.
In the CH system, as in most early radars, the functional and physical separation of transmitter and receiver was a matter of technical necessity. Development of the duplexer and the shift from decametric to centimetric wavelengths soon led to universal adoption of monostatic radar systems. Bistatic cooperation has been largely neglected because its complications were avoidable as long as progressive improvements in monostatic performance could match the developing countermeasures threat.
This paper examines some of the threats faced by today's AD radars and shows strong reasons for reconsidering both supplementary bistatic operation, using sensors at well-separated sites, and passive jammer location (PJL) provided by the same sensors cooperating simultaneously in passive modes.
Evolution of the Threat
Earth curvature limits the low level coverage of radar surveillance and modern terrain-following systems make long distance missions feasible at low level. However, sustained low level operation is unattractive. Low altitude flight paths stress aircraft and crew and are difficult to coordinate. Speed decreases but fuel consumption rises, limiting range and payload. Also, ground radars can exploit terrain features and low level incursion is not secure from short range weapons. The arguments for and against flying below radar cover have been debated since WWII without undermining the leading role of long range AD radar in NATO defense. AD radar denies the prized middle air space and an attacker must rely on countermeasures to win it back.
The efficacy of wideband noise jamming against radars has long been appreciated and, up to the mid-1960s, the perceived threat against surveillance radars consisted of large numbers of stand-off (SOJ), escort (ESJ) and self-screening jammers (SSJ). To counter this threat, fixed monostatic radars were built with large mean-power aperture dwell (MPAD) products and with improved sidelobe levels. Frequency agility was adopted to counter spot-jammers attempting to match their radiated bandwidth to that of the radar.
However, during the last two decades, the perceived threat has shifted from soft kill to mixed soft/hard kill. Soft kill is still perceived as multiple SOJ/ESJ/SSJ deployment, using a combination of broadband, responsive narrowband and deception jamming, but added to this are threats, such as cruise missiles and antiradiation missiles. The new threats have influenced the development of AD radars. New generations of primary radars of increasing size and power can be neither expected nor afforded. A priority in AD radar is now survivability. Major performance advances must be obtained in some other way.
A future aggressor must jam powerful radars that may use electronic beam scanning with adaptive pattern control and time-energy management. In long range AD scenarios, only airborne jammers will normally be suitably placed and these will need to maximize effective penetration of their targets. This may be done by radiating more power, by concentrating power into the instantaneous bandwidth of a target radar or by increasing antenna gain. Due to the limited prime power available to an airborne jammer and to difficulties in following agile radar emissions in a many-on-many scenario, the use of high gain jamming antennas is to be expected, despite the need for threat-assessment suites to direct the jammer beams.
The previously discussed threat could overwhelm a monostatic radar. However, in practice, surveillance is shared by a number of radars that pool their detections to form a recognized air picture.
Multisite radar cooperation will attract a numerically greater jamming threat than a single site, and so the need and the value of PJL support will increase. Also, it provides the geometrical basis of simple PJL, triangulation of measured jamming strobe directions to give plots of jammer locations. This might locate both the raiding ESJ/SSJ platforms and the more distant SOJS, although simple triangulation has limitations when there are many jammers.
Such advantages of multisite operation apply only when the sensors are linked by a communications network and when their coverage areas have sufficient overlap. Because of the increasing cost of major radars, a natural tendency is to space them widely apart. The familiar concept of autonomous monostatic radars pooling information leads on to consideration of true bistatic or multistatic cooperation, where sensors at different sites are more tightly coupled.
A basic passive-bistatic role is jammer location. The task of PJL in support of AD radar is to detect, locate and track all in-band jamming platforms affecting radar surveillance of the intended coverage area.
Since ESM systems are specifically designed to identify and make strobe angle measurements on emissions over a broad band, well-separated but linked ESM systems might seem to meet the requirement. However, ESM equipment is not effective in separating multiple radar jammers whose noise-like emissions have substantial spectral similarity and overlap. The quasi-omnidirectional ESM antenna beams cannot resolve a distribution of jammers in angle.
Another difficulty is caused by the multiplicity of ambiguous intersections found when attempting to locate a number of jammers by strobe triangulation. ESM systems usually cannot provide enough information to resolve this type of ambiguity. Similar problems occur with jammer strobes obtained from radars.
An effective method of pairing jamming strobes is by cross-correlation of the strobe waveforms. Consider two omnidirectional sensors on different sites, whose receiver outputs are simultaneously recorded and subsequently cross-correlated. Any emission received in the common passband should result in at least one peak in the cross-correlation function. This will occur when the differential processing delay between the waveforms being correlated compensates for the differential propagation delay from the jammer to the sensor sites. Measurement of this delay enables the jammer to be located on a hyperbolic surface having the two sites as foci. If, instead of being omnidirectional, one of the sensors had a narrow beam (directed at the jammer), a positive location would be defined at the intersection of the beam with the hyperbolic surface, as shown in Figure 1. Also, a pencil beam or other elevation measurement at either site would define height.
Cooperating sensors must be synchronized in both frequency and time to ensure that compatible data sets are recorded for cross-correlation. If either sensor uses a narrow beam, this must be scanned to intercept the jamming emissions and its pointing direction recorded. In order to carry out the cross-correlations, some form of link is required between the sensors.
The output of the cross-correlation is tested against a threshold. The delay and pointing angle parameters associated with a threshold crossing define a plot. These plots can resolve jammer groups, can distinguish incursive jammers from the SOJ background and can form a separate track for each jammer. Thus, passive-bistatic correlation can provide effective PJL support whenever raiders use their ESJ and SSJ resources to confuse radar detection.
While the outlined system can perform well against jammers emitting continuous broadband noise, there are other types of jamming that are likely to be encountered. The responsive jammer poses particular problems to our simplified system, since the jamming is radiated in a relatively narrow spectral band centered on the last measured frequency of an agile radar target. The problem is to ensure that synchronized compatible recordings at the two sensors are taken in the right beam directions in the spectral band being recorded. This must be done before the jammer's location has been determined and perhaps when its strobe heading cannot be resolved from others.
A solution is to use the receiver of a target radar as one of the passive sensors. A responsive jammer will be triggered when the radar's main beam transmission is directed towards it. The jammer then must respond in-band or risk being located directly. Any successful jammer response penetrates the radar receiver at good strength, and this meets the PJL requirement for a narrow beam receiver pointed at the jammer and on the right frequency at the time of its enforced emission. The arrangement avoids the need to provide a specialist PJL sensor at a radar site.
Before accepting the case for using a radar as one end of the PJL baseline, what to use for the other sensor must be considered. A second radar cannot be used because two adjacent radars would not illuminate the same jammer simultaneously on the same frequency and could not be directed to do so until after the jammer had been located. The second sensor could be slaved to receive on the radar's frequency, but it must have broad azimuth coverage if it is to intercept unlocated jammers as they respond to illumination by the scanning radar.
At first sight, the second sensor could have a simple omnidirectional antenna, since the intersection of the radar beam with a hyperbolic surface should be sufficient to define jammer location. However, in practice, solution of realistic multi-jammer scenarios requires antenna gain at both ends of the baseline.
The system is required to locate the position of each individual jammer (a target), typically in the presence of many other jammers distributed in space. At both sensors, the signals from each target will be accompanied by signals from the other jammers, defining a target-to-total signal power ratio, which will depend on jammer power and deployment and on the sensor beam profiles. A given correlation process will provide reliable detections of a target signal only if the target-to-total ratio is large enough.
Although a well-designed correlation system, using sufficiently long dwell times, can accept target-to-total power ratios well below unity, it can be shown that in all but the most modest of jamming scenarios an antenna without angular discrimination often gives unacceptably low ratios, leading to detection failure even when the jamming signals are strong. A narrow-beam antenna is needed at both sensors to exclude some of the competing jammer signals so that the correlator can resolve the remainder without requiring an excessively long dwell time.
If the second sensor has a single steerable beam, then, in a search mode, it will need to be scanned across the field of view during the beam dwell time of the radar sensor, as shown in Figure 2. This limits the correlation dwell time and, hence, the processing gain available at each intersection cell.
Nevertheless, by using large antenna apertures and slow radar scan rates, useful performance can be achieved in search mode. Also, further local reductions in scan rate can be used to improve sensitivity once probable jammer locations have been determined. However, progressive developments in the jamming threat demand corresponding improvement factors in PJL performance.
A large improvement factor over the single beam system is available if the second PJL sensor has multiple simultaneous receiving beams. This allows a dwell time improvement factor basically equal to the number of beams servicing the area of threat, which can be 100 or more.
The technology for generating multiple receiving beams[5,6] with low sidelobe levels has been demonstrated as part of the UK bistatic program[7,8] and is already in operational service in AD radar receivers.9 By using IF techniques to process the outputs of a multi-element array, almost any desired number and arrangement of the receiving beams can be formed, subject to the dimensions and quality of the array aperture.
For long range PJL applications, discrimination in elevation contributes little to the improvement factor, so the requirement is for multiple narrow beams in azimuth only A horizontal strip-like array aperture with a linear beamformer is adequate. Since IF techniques are used, one beamformer and PJL processing system can service several strip-like array apertures covering different radar bands and directions. This solves the problem of providing comprehensive PJL support in distributed AD networks where the radars operate in different bands.
Supplementary bistatic radar modes are considered in which the baseline, transmit/receive spacing, is substantial, perhaps 50 to 200 km in AD applications. If the target range is much greater than the baseline, the performance is not very different from that of an equivalent monostatic radar. At shorter ranges, the capabilities of a bistatic radar differ significantly from those of a monostatic radar having the same MPAD product.
The bistatic receiver is well removed from the transmitter and from the unwelcome attention that powerful transmissions may attract. If the attacker does not know the receiver location, he does not know how to deploy SOJ resources to cover his line of attack. His difficulties will be multiplied if the transmitter also has its own monostatic receiver (as in current AD radars) or if there is a second bistatic receiver elsewhere.
It has been shown that bistatic geometry has the potential to reduce the effects of high gain directable-beam jammers. By steering its emissions towards the transmitter, such a jammer will reduce its effectiveness in the direction of the covert receiver. However, the receiver can use its own directivity to its advantage. Jammers not in the same beam as the wanted targets will be attenuated by receiver sidelobe protection and, if the antenna is of the phased-array type, the receiver could additionally use adaptive null steering.
If the attacker cannot rely on SOJ cover alone, his raiders may use ESJ/SSJ resources, or chaff. The defense might use PJL to locate and count the jammers, but this does not help the radar to see targets in chaff. Although a competent PJL system should have no difficulty in seeing through chaff, this is no use if the chaff has enabled raiders to maintain jamming silence.
However, in active-bistatic radar the transmitter can operate usefully at a much greater pulse repetition frequency (PRF) than an equivalent monostatic radar[11,12] because the bistatic receiver can use angular filtering to separate signal returns from the individual pulses, as shown in Figure 3. High PRF operation can increase the mean power of target illumination and, importantly, it also can improve the detection of targets in chaff. Conventional long range monostatic radars using low PRFs find that turbulence and velocity shear due to atmospheric motion can spread chaff returns to fill the unambiguous Doppler spectrum. Higher PRFs extend the range of unambiguous Doppler spectrum, reducing the probability of targets being masked. Suitable bistatic receiving techniques allow transmitter PRF to be increased by an order of magnitude or more, offering a significant detection improvement factor in chaff.
At high PRF, the bistatic receiver must gather and resolve target returns scattered from the several transmitter pulses simultaneously in flight. This is more difficult than in monostatic radar. One solution is for the bistatic receiver to generate an azimuth array of narrow beams spanning the angular cover. Alternatively, assuming timely knowledge of the transmission schedule and pointing direction, the bistatic receiver need generate only some smaller number of highly agile beams directed dynamically to intercept returns from the succession of pulses. Knowledge of transmission schedules normally requires a data link between transmitter and receiver, although unlinked bistatic exploitation of transmission from rotating-beam surveillance radars has been reported.[16,17]
Both multibeam and agile-beam reception have been utilized in an experimental bistatic radar test-bed that has been operating in the UK for several years.[7,8] Tests have shown that long-baseline bistatic working with a cooperative AD-type radar transmitter is fully practical and can be achieved within current phased-array technology. In these tests, the data link between transmitter and receiver, carrying pointing and schedule information at kilobaud rates, uses a single good quality telephone channel.
A bistatic sensor that never transmits can be optimized for reception. The sensor must generate multiple or agile azimuth beams with low sidelobes. Having no transmit function, these beams can be formed at intermediate frequency or at baseband, so pattern adaptivity can be incorporated using digital techniques at little extra cost. The receiving phased array of the UK bistatic radar test-bed is of this type, having a separate first mixer for each vertical group of array elements. It has been equipped to form a large number of multiple beams fixed in space, a smaller number of highly agile beams and an adaptive null pattern. These are available simultaneously but are used and adapted independently.
In summary, provided the architecture of array elements and associated front-end arrangements are well chosen, bistatic phased-array receivers can support a full range of techniques, including high PRF and advanced signal processing techniques, to improve radar performance in the presence of severe threats. Supplementary bistatic support within an AD radar network can make a major improvement in performance in difficult scenarios. The geometrical advantages of bistatic working have long been known. The other major performance improvements now available stem from recent advances in multibeam phased-array antennas and in digital signal processing.
Active-Passive Network Synergy and Compatibility
PJL and active-bistatic working can each offer substantial, but not individually complete, support to an AD radar network facing a varied countermeasures threat. For example, PJL can locate and count ESJ/SSJ emitters, but does not induce intruders to emit. Active-bistatic support can detect silent raids in chaff, but will encourage multiple ESJ/SSJ emissions with which neither it nor conventional radars may be able to cope. It is clear that PJL and active-bistatic support work synergistically. It has been argued that efficient dual support within a multiradar AD network could counter all the important threat combinations.
Today's AD networks include a mix of radar types operating in different bands, typically D-band, E/F-band and G-band. Ideal support would cover them all. For large improvement factors, both PJL and active-bistatic working require a sensor with multiple or agile narrow beams in azimuth. Moreover, both systems cooperate with the radars themselves and require instantaneous bandwidth and agile tuning range similar to the radars.
It is clear that both active and passive bistatic receiving systems can use a common antenna concept, which will be a phased array with a horizontal aperture comparable to those of the radars being supported. True broadband radar-type antennas covering several octaves are not yet practical. However, since Western AD radar is confined to defined bands and modest instantaneous bandwidths, a true broadband support antenna is not required.
A typical receiver would need azimuth beamwidths of 1.5 [degrees], corresponding to 5 m of planar aperture at E/F-band, 3 m at G-band and 12 m at D-band. Allowing modest vertical dimensions, these apertures stack conveniently into one face of a 12 m ISO trans-shipment container, as shown in Figure 4, or into a 6 m container if D-band is omitted.
Geometrical Aspects of Compatibility
The geometry of curved-earth radar intervisibility for a distributed sensor network is a complex subject. This paper takes a simple approach. First, the addition of PJL and active-bistatic support will not improve the geometrical radar visibility of targets, no matter where support antennas are placed; nor will it allow the main monostatic AD radars to be deployed further apart. This is obvious, but is sometimes misunderstood.
Second, almost all aspects of bistatic support and improvement factor, both passive and active, are strongest within a broad angle normal to the bistatic baseline, and become geometrically weak at angles close to the baseline. Where the AD network has the general form of a chain, a planar passive support antenna added to the chain between radars should be slewed to align with the general axis of the chain, as shown in Figure 5, and will then give good improvement factors normal to the axis, but not along it. True all-around coverage, giving good improvement factors in all directions, demands a denser deployment of sensors in depth.
Time and Frequency Management
Active-bistatic reception must be tuned to the transmission frequency of a cooperating radar. Jamming signals for correlation must be received simultaneously by a cooperating radar receiver and by the support sensor. These requirements are clearly compatible and require only a single set of schedule data sent by link, although cooperation with several radars would require timesharing.
A suitably sited multibeam support receiver could cooperate with two or three radars. Many AD radars use antennas that rotate in azimuth and may scan sequentially through sectors where support is needed. Time and transmission constraints in a multisensor network can allow simultaneous passive and active support modes without compromising radar autonomy or requiring tight control for timesharing the support.
Within the support sensor, the receiving array(s) and multiple beamformer can be common to both active-bistatic and PJL modes without compromising simultaneous use. Processing of recorded jammer signals differs from active-bistatic processing, but apart from real-time functions, such as A/D conversion, all signal and data processing is digital and could be performed in either common or dedicated hardware as desired. A nondedicated approach to processing hardware allows the support functions to be adapted to requirements. Array, beamforming and processing architectures should be chosen carefully in order not to constrain adaptability. Experience shows that defense systems may be used in circumstances quite different from those originally envisaged by their designers.
Functional survivability must be considered when the active AD radars are under threat of attack. The covert support receivers, well away from transmitters, face little physical threat. Therefore, it is logical to locate as much as possible of the system at these receiver sites and to minimize equipment burdens at the radars. To cooperate with a support receiver, a radar should need no more than a small, easily-fitted unit interfacing with the data link.
An obvious potential weakness associated with the support receivers is their dependence on data communication with other network elements. The communications link could take a variety of forms depending on the data rate required, vulnerability and consequent constraints on the mobility of sensors. The overall work rate of a support receiver is reduced by poor communications, but the loss is graceful and can be controlled to maintain priority tasks. In appropriate circumstances, limited performance can be maintained without any links.[16,17]
Another weakness is dependence of the support on the vulnerable radars. Apart from their own primary monostatic radar functions, these radars provide illumination for active-bistatic reception, inducement of responsive jamming, seduction of directed jamming signals and schedule data for simultaneous bistatic reception. If these functions are removed, the support will be much reduced.
To reduce vulnerability, AD radars may be deployed well to the rear, may be moved frequently, may practice EMCON and may use a variety of agile transmissions. These imply that target detection and support requirements will be at longer ranges in a defined direction of threat rather than locally with 360 [degrees] coverage, and that sensors will be netted so that individual radars are not indispensable and can be shut down or moved. Therefore, the support receivers must be adaptable to a variety of different radars, transmissions and data link arrangements and also must be at least as mobile as the radars they support but could have slewable-sector, rather than full 360 [degrees] coverage.
If the radars fail, active-bistatic support will have no illumination, but jammer location and reporting can continue provided that there are jammers emitting within view of two support receivers. Even a single support receiver can find and report jammer strobes. However, although useful, this level of performance is not ideal. It is highly desirable to maintain some level of radar illumination in the AD network.
The vulnerability of active AD radars demands that active-passive bistatic support must be mobile, versatile and adaptive if it is to match the radar's own survival strategy. However, where suitable support receivers are provided, they can contribute substantially to functional survivability, as well as to performance, in an AD network.
In the struggle for the control of middle air space, developments in countermeasures have eroded the performance of air defense radar at a time when the considerations of cost and vulnerability preclude reequipment with radars of increased size, power and complexity.
This paper has considered how dual PJL/active-bistatic support of current in-service AD radars can offer important system enhancements. A mobile support receiver capable of serving several different AD radars within its deployment range has been proposed. It has been shown how multibeam receiver architecture supported by digital signal processing can substantially improve surveillance performance against chaff and jamming threats and contributes to survivability.
Based on experience with the UK bistatic test bed, a dual-mode support receiver concept was proposed that uses current phased-array technology, modular processing in industry-standard hardware and existing networks. With agreement on interfaces, the concept is applicable throughout NATO. Multistatic support is practical, affordable and available.
The work reported in this paper has been carried out with the support of the Procurement Executive Ministry of Defence and Marconi Radar Systems Limited. (*) This paper was presented at a symposium held by the Advisory Group for Aerospace Research and Development in October 1990.
[1.] B.T. Neale, "CH: The First Operational Radar," GEC Journal of Research, Vol. 3, No. 2,1985, pp. 73-83. [2.] I. Bardash, "Phased Arrays For ECM Applications," Microwave Journal, September 1982, pp. 81-92. [3.] G.N. Taylor and B.P. Scoffield, "Passive Location of Microwave Emitters," Military Microwaves 1990 Conf. Proc. pp. 484-488. [4.] F.J. Berie, "Mixed Triangulation/Trilateration Technique For Emitter Location," Proc. IEE, Vol. 133, Pt. F., No. 7, 1986, pp.638-641. [5.] J.R. Wallington, "The Role of Analogue Beamforming in Radar," GEC Journal of Research, Vol. 3, No. 1, 1985, pp. 25-33. [6.] B. Wardrop, "The Role of Digital Processing in Radar Beamforming," GEC Journal of Research, Vol. 3, No. 1, 1985, pp. 34-45. [7.] M.R.B. Dunsmore, "Bistatic Radars For Air Defense," IEE Radar 87, Conf. Proc. No. 281, 1987, pp. 7-11. [8.] T.A. Soame and D.M. Gould, "Description of an Experimental Bistatic Radar System," IEE Radar 87, Conf. Proc. No. 281, 1987, pp. 12-16. [9.] C. Latham, "MARTELLO: A Modern Three-Dimensional Surveillance Radar," GEC Journal of Research, Vol. 3, No. 2, 1985, pp. 104-113. [10.] B. Wadrop, "Bistatic Radar Using Adaptive Digital Beamforming," 1990 IEEE AP-S Symposium Digest, Vol. 1, pp. 392-395. [11.] K. Milne, "Principles and Concepts of Multistatic Surveillance Radars," IEE Radar 77 Conf. Proc., 1977, pp. 46-52. [12.] M.R.B. Dunsmore, "Bistatic Radars," Alta Frequenza, Vol. 58, No. 2, March-April, 1989, pp. 53-79. [13.] M.C. Jackson, "The Geometry of Bistatic Radar Systems," Proc. IEE, Vol. 133, Pt. F, No. 7, 1986, pp. 604-612. [14.] P. Matthewson, "An Experimental Pulse-Chasing Bistatic Radar Receiving System," 19th Euro. MW Conf. Proc., 1989, pp.275-280. [15.] P. Matthewson, B. Wardrop and D.M. Gould, "An Adaptive Bistatic Radar Demonstrator," Military Microwaves 1990, Conf. Proc., pp. 471-476. [16.] J.G. Shoenenberger and J.R. Forrest, "Principles of Independent Receivers For Use With Cooperative Radar Transmitters," Radio & Electronic Engineer, Vol. 52, No. 2,1982, pp. 93-101. [17.] P Mallinson, "A Bistatic Radar Tracker," Phillips Research Labs, Redhill: Annual Review, 1987, pp. 96-98. [18.] J.C. Old, D.J. Day and G.N. Harvey, "GASP: The GEC Array Signal Processing Test Bed," IEEE Radar 85, Conf. Proc., pp.220-225. [19.] C.A. Fowler, "Comments on the Cost and Performance of Military Systems," IEEE Trans., Vol. AES- 15, No. 1, 1979, pp. 2-9. [20.] D. Wooller and A.J. Poelman, "The ECCM Potential of a Passive Bistatic Complement to Radar Detection of Aircraft Targets," AGARD Symposium on Avionic Countermeasures, Munich, 1990.
Brian Wardrop received his PhD degree in 1970 from Loughborough University for research into stripline group-delay equalizers. Since 1970, he has worked in the Radar Research Lab of GEC-Marconi Research Centre. In 1988, he became manager Wardrop has conducted research into the areas of charge-coupled devices, radar signal processing, adaptive antennas and multireceiver phased-arrays for bistatic radar. He has served on committees for several IEE International Radar Conferences and is an IEE Fellow.
Bob Molyneux-Berry received his BSc degree in electrical engineering from Bristol University in 1965. Since that time, he has worked at GEC-Marconi Research, where his work has encompassed microelectronics, railway traction, nuclear fusion, electron tubes, broadcasting, laser pulse systems, radar transmitters, processing and displays. Since 1985, he has concentrated on developing novel approaches to passive sensing at microwave frequencies. He has originated more than a dozen patents.
|Printer friendly Cite/link Email Feedback|
|Author:||Wardrop, B.; Molyneux-Berry, M.R.B.|
|Date:||Jun 1, 1992|
|Previous Article:||Planar microwave electro-optic phase shifters.|
|Next Article:||An improved layout for Lange couplers.|