EuroFIRST PIRATE

from Wikipedia, the free encyclopedia
Typhoon with PIRATE, on the port side in front of the cockpit. In the background heat shimmer from the waste heat from the aircraft.

The EuroFIRST PIRATE ( English Passive Infra-Red Airborne Track Equipment ) is the infrared target system of the Eurofighter Typhoon . It is produced by the EuroFIRST consortium, consisting of SELEX Galileo , Thales Group and Tecnobit. The sensor was developed explicitly in order to be able to locate Soviet stealth aircraft over great distances even under heavy electronic countermeasures . Due to the demanding performance requirements - the sensor should be able to locate, identify and track targets in three dimensions like a radar in track-while-scan mode , and have an observation range comparable to the radar - it took a very long development period. It took more than 15 years from the beginning of development to delivery of the first sensor. PIRATE is the first electro-optical sensor with this capability outside of Russia.

PIRATE can work both as a Forward Looking Infrared (FLIR) and an infrared aiming system (IRST). When it was delivered in 2007, the performance was limited by the computing power available. In 2010, PIRATE was evaluated against stealth fighter aircraft of the type F-22 Raptor, which could be located at a “significant distance”. Up until 2013, PIRATE's tracking range could be further increased through software updates . For cost reasons, not every Eurofighter is equipped with this component.

development

Performance requirements

Analyzes that were carried out on the European Fighter Aircraft (EFA) at the end of the 1980s showed that the detection range of the CAPTOR radar would be reduced to less than 9 km through well-known Soviet stealth technology and distance-effective jammers. It was also expected that the opponent would outnumber them, at least 2: 1. Since in a dogfight the side that can throw more aircraft into the fight usually wins, the number of opponents had to be decimated on the approach. In order to still be able to conduct Beyond Visual Range battles, an infrared aiming system and sensor fusion were considered necessary. The tender demanded that the IRST would work in conjunction with radar and should therefore have approximately the same observation area. Recorded targets should be automatically tracked, characterized and the flight path determined. The targets should be shown on a display like a radar. Furthermore, it should be possible to display an infrared image of the target to the pilot in order to visually identify it. The system should also be able to work as a Forward Looking Infrared to give the pilot a landing aid. The main difficulty of the performance requirements was clutter suppression; it is estimated that the infrared aiming system would have to reject up to 1000 false targets per hour.

Spain joined the Eurofighter consortium in September 1985, but the Memorandum of Understanding for the development of the Infra-Red Search and Track (IRST) system was not signed until 1988. In 1989 two industrial consortia applied for the contract: FIAR and Thorn-EMI, as well as GEC Avionics and Pilkington and Ferranti. Germany decided in 1991 not to take part in the development of the sensor for reasons of cost, but reserved the right to use it later. At that time, PIRATE was already three years behind schedule. While Germany exited, Spain, the UK and Italy insisted on the IRST system and were unwilling to accept lower performance in order to cut costs. The British company Thorn-EMI (now Thales) finally won the tender for the infrared sensor of the Eurofighter, which is based on the Air Defense Alerting Device (ADAD), in September 1992 with the Italian FIAR and Spain's Eurotronica. The EuroFIRST consortium founded for this later consisted of FIAR, Eurotronica and Pilkington Optronics, which emerged from Thorn-EMI. So in 1992 the development of the sensor began.

Basic research

Since the mechanics were developed on the basis of the ADAD, the challenge lay in the development of the software through which the radar-like capability of the infrared target system is guaranteed. Practically, the experiences that were already available in the field of radar processors could be transferred to the IRST.

Stormer HVM of the British Army in the Sauerland . The ADAD's sensor dome is located on the tower.

The Air Defense Alerting Device (ADAD) was developed by the British company Thorn-EMI for the British Army and put into service in July 1993. The small Infra-Red Search and Track (IRST) sensor was supposed to help instruct MANPADS systems in their targets by means of horizontal 2D scans in the wavelength range of 8–13 µm . Between 1999 and 2000, a number of investigations were carried out in the UK using the ADAD sensor to address the problem of stealth and jammers in air defense. During the Kosovo war , too , every NATO air mission was accompanied by Grumman EA-6s , which made the Serbian air defense difficult. A sensor fusion of radar and infrared data was investigated as a solution. Furthermore, targets could also be located using stealth technology , since IRSTs at wavelengths of 8-12 µm have good sensitivity to targets at room temperature, and the outer skin of an aircraft can be located. The aircraft skin also heats up during supersonic flight due to air friction.

In tests by Pilkington Optronics, BAE Systems and the Defense Evaluation and Research Agency (DERA), stealth aircraft optimized for radar could easily be located by IRST sensors. Interestingly, it was found that the stealth coating of some aircraft increased air friction and thus heat radiation. Since locating UAVs and cruise missiles is particularly difficult because they are very small and have little radar reflecting surface and infrared emissions, the team's practical experiments focused on these goals. First, a BQM-74 drone simulating a cruise missile was tracked by a modified ADAD. He didn't keep her eyes on her until the drone opened the parachute at the end of its flight path. From then on it no longer behaved like an airplane, so that it was ignored by the signal processor and disappeared as a target on the display. In the next step, the discrimination of targets against a warm background was improved, specifically tests were carried out with a special ADAD to locate helicopters despite clutter . The algorithm only presented the helicopter to the operator, other infrared sources were ignored.

The next step was to test track fusion by Pilkington Optronics and DERA. For this purpose, two ADADs were operated as an IRST / IRST sensor network, BAE Systems took part in the tests where an ADAD was combined with a radar. Both sensors located targets independently of each other, the tracks were combined by sensor fusion. During tests, the IRST / IRST network was able to successfully locate approaching helicopters and trace the flight path of three aircraft in space. Due to the high resolution compared to radar, the position determination was more precise. The IRST / radar fusion was tested against helicopters, the target tracking proved to be more robust and the elevation determination to be more precise. The findings of the ADAD trials were incorporated into the Eurofighter's IRST.

Integration and delivery

In 1998, Pilkington Optronics was swallowed up by Thomson-CSF, later renamed Thales Group , and FIAR's position has since been taken over by SELEX Galileo . Spain's Eurotronica has been replaced by Tecnobit. In the summer of 2000 the PIRATE system was first mentioned in the published specifications of the Eurofighter. In laboratory tests, the sensor was mounted on a two-axis table to simulate the movements of the aircraft, and IR targets of various shapes were projected through a collimator into the mirror in order to check the parameters. Temperature differences of up to 0.01 K were tested and the noise in the sensor was reduced. A modified Dassault Falcon 20 was used for ( interview ) “ hack flight tests” to optimize the software , the Eurofighter test flights mostly took place with DA7 from 2002. The Eurofighter Typhoon was supposed to go into production at this time, but this was delayed by 2003. The delivery of the PIRATE should begin at the same time.

Flight tests began in 2001, when the FLIR function was tested on transfer flights between Turin and Sardinia. Objectives of opportunity were later also observed with the FLIR in order to coordinate the integration into the display elements of the Eurofighter. At first the image was too pale, especially with uniform backgrounds, which is why the histogram had to be changed. The prototype software was written so that it could be adapted during the flight. During hard maneuvers, the display also flickered brightly, which made software changes necessary. Furthermore, the temperature changes during the flight were greater than assumed in the laboratory, so that the algorithms for temperature compensation had to be completely rewritten. The sometimes different behavior of pixels on the detector led to streaks in the image, which were removed by improved software. For this purpose, the detector element was tested before delivery in order to mark bad pixels. Finally, landings with PIRATE as FLIR were carried out to demonstrate emergency landings in hostile areas, swampy or sandy areas, and landings on improvised landing sites . Due to flight safety regulations, however, these landings never took place in complete darkness, as civil airports were used.

Since there was practically no literature on the problem of the infrared aiming system at the time, a simple approach was chosen: the sensor was simply activated in FLIR mode and you wanted to see what it could detect. It was found that look-up scenarios (i.e. the targets are above the horizon) were much more pleasant, since look-down scenarios (i.e. the targets are below the horizon) require the target to be extracted from the IR background. As a result, work began on the look-up problem first. In 2004 the sensor was used for the first time as an IRST: A target that was found with CAPTOR was followed by PIRATE in STT mode for 200 seconds. However, at the time the false alarm rate (FAR) was too high, targets were also reported that did not even exist. After hard and controversial discussions, the decision was made to increase the threshold value and slowly work towards (original sound) "extreme detection ranges". The new threshold was set by mounting the sensor on the roof of the manufacturer's building and tracking traffic from the nearby civil airport. The threshold obtained by this method was so conservative that in subsequent flight tests with a cooperative goal this could only be pursued to a handful of miles. In return, the false alarm rate (FAR) was practically zero. Now it turned out that tracking was difficult when the target was between two scan strips and was maneuvering. A more precise production, a repositioning of the derotation prism and a revision of the algorithm were the solution.

DA7 climbing. MMX603 carried the brunt of the test flights for PIRATE

The signal-to-noise ratio (SNR) has now been gradually reduced and a diagnostic option has been integrated via the bus. Automatic search volume adjustment and auto-centering were also demonstrated. In order to reduce the FAR, the declaration time has now been increased slightly, i.e. the time in which a target is pursued before it is upgraded to a track. The stability of target tracking and movement prediction have also been improved. Furthermore, the classification, already included in the software but deactivated, was added, which turned out to be useful. The CAPTOR now served as a reference, whereby PIRATE achieved comparable positioning ranges and a better angular resolution. Very often targets could be tracked with both radar and the infrared aiming system, which improved the sensor fusion. Initially, the passive determination of the range by PIRATE proved to be too imprecise. In the test stand, the optimal maneuver was sought to improve the distance measurement, but the successes in the flight tests were mixed. The problem also affected the sensor fusion, since the declared radar and IRST contacts were precise enough at an angle; however, the difference in distance between PIRATE and CAPTOR was sometimes too great for the contacts to merge into one entity.

After the problem could be solved satisfactorily, the fine-tuning was started. Infrared aiming enemies usually fly near clouds to hide in their glow. The AN / APG-77 , for example, has a weather map mode as a special feature; certainly not for making weather forecasts. This problem is to be countered within the framework of what is physically feasible through improved classification algorithms. At the end of the test phase, Eurofighters, tornadoes and phantoms could be tracked by PIRATE with angles and distance information for a long time and the data transferred to the bus. In the case of FAR and passive range determination, improvements were still considered to be desirable. The thermal cueing (TC) function worked without any problems, only the prioritization of the targets was an open question. Following suggestions from pilots, the targets were weighted so that priority was given to the proximity to the aircraft, the flight altitude of the Eurofighter above the ground, and the position of the target in relation to the horizon.

During development, EADS Ottobrunn was responsible for the robustness and the interface to the aircraft, while the flight tests mostly took place in Italy in Turin under Alenia. Since the integration of PIRATE into the Attack and Identification System (AIS) was the task of EADS, the sensor was connected to a laboratory setup of the AIS with radar, IFF, MIDS, LINS, radar altimeter and cockpit with symbol generator and HUD. The scenarios were simulated on the computer, whereby the entire avionics and their interaction could be tested. Air and ground targets were located by PIRATE in different scenarios against different backgrounds (land, sky, sea, forest, city), and fed into the bus. A stand-alone version was also available for the tracking tests. The data exchange between PIRATE, AIS and other avionic computers went smoothly. The CAPTOR radar has a better range resolution, while the PIRATE-IRST has a better angular resolution.

Although only five PIRATEs were in use in Warton, Manching and Turin during the test phase, and these systems were very experimental, there were hardly any hardware failures, so that work could concentrate on the software. The first PIRATE sensor was finally delivered to Aeronautica Militare on August 2, 2007 in a Tranche 1 Block 5 Eurofighter . In 2010 the US Air Force relocated four F-22 Raptor fighter jets to RAF Lakenheath . In subsequent encounters, PIRATE was able to locate the machines at (interview) “significant distances”. Up until 2013, PIRATE's tracking range could be further increased through software updates .

overview

OLS-27 of a SU-27UB

The installation of IRST systems in combat aircraft has a long tradition: Even aircraft in the Century series such as the Convair F-102 were equipped with an infrared sensor. At that time sensors like the AN / AAA-4 of the F-4 Phantom II consisted of only one pixel, whereby the sector was scanned mechanically. The scan image was presented to the pilot as C-scope and the range was negligible. The system was therefore quickly replaced by the AN / APR-25 radar direction finder.

Only with the appearance of the MiG-29 and Su-27 were usable systems with an acceptable location range mounted on aircraft with the OLS-29 and OLS-27. In these machines, the C-scope is abstracted and projected onto the head-up display . When choosing the infrared aiming system, the pilot sees short horizontal lines on the HUD, which mark the position of IR contacts. The leftmost and rightmost HUD positions each mark the maximum azimuth angle of the scan area, the same applies to the elevation. When the pilot locks on a target, the distance is measured with a laser, so that a fire control of guided missiles is possible. All other goals are lost. A mapping identification is therefore not possible. Modern systems such as the Optronique Secteur Frontal (OSF) of the Rafale work on the same principle, but have a more sensitive detector array, which enables greater ranges in BVR combat and visual identification.

The manufacturer of PIRATE makes no secret of the fact that the system was developed explicitly to locate LO aircraft and to neutralize jammers. In order to undermine infrared warning measures, PIRATE does not look for heat directly, but for changes in the background clutter . A target that suppresses its heat radiation would appear as a "black hole" in relation to the natural background and would be easier to locate. The only way to go undetected would be to send out the same infrared radiation as the background obscured by the aircraft, which is practically impossible. Even if it were possible to adapt the IR signature of an aircraft, it would not be clear to which background the signature would have to adapt, since the position of the IRST is unknown, and thus also the background image it sees.

The main problem today, however, is more jammers. By taking jamming pods like the AN / ALQ-184 with you, the small fighter aircraft radars on the other side can be practically neutralized. Not only did the Serbian MiG-29s experience this during Operation Allied Force , but also the American F-15s before the Red Flag maneuvering exercise in 2008 . During preliminary exercises at Mountain Home AFB , Indian MiG-21s were able to neutralize the AESA radars of the F-15C with modern Israeli jamming pods. As a result, all modern combat aircraft are equipped with IRST, and existing models such as the F-16E / F or F / A-18E / F are retrofitted. The capabilities of the infrared aiming system for the Super Hornet are essentially a copy of PIRATE; the location ranges should be AMRAAM-compatible. With the AIM-9X Block III, the US Navy is also developing a range-enhanced variant of the IR guided missile. This is also to be understood as a move against Chinese LO aircraft and interferers, which could subvert the radar-centered battle Beyond Visual Range with AMRAAM.

technology

sensor

PIRATE is a passive, rotatable, cooled infrared sensor with high resolution, which is also known as imaging infrared. It is based on the ADAD, but is largely a new development. The dome visible from the outside is only the azimuthally rotatable sensor head, which contains the tiltable mirror. Both elements are stabilized, the maximum azimuth and elevation range is probably 150 ° × 60 °. The sensor only consists of a Line Replaceable ltem (LRI), which in turn contains four subsystems: The sensor head module with the stabilized sensor head, the telescope with double magnification, the detector, the servo control processor and the electronics for signal adaptation. The data processing computer for discovering and tracking targets, the video processor for displaying the infrared image on the HUD, HMD or an MHDD, and the interface to the aircraft with BITE and data link. The sensor is connected to the MIL-STD-1553 and the EFA bus . Commands are mostly sent through the MIL bus, data through the EFA bus. PowerPC processors were used for signal processing whenever possible. The waste heat from the system is given off to the aircraft via a liquid cooling circuit. PIRATE also has an internal gyro instrument , the data of which are combined with those of the inertial system of the Eurofighter in order to improve the quality of stabilization. The dimensions of the sensor are 680 × 591 × 300 mm (L × W × H), and it weighs 48 kilograms. From the outside only the smallest part of the sensor is visible, most of the volume is in front of the cockpit.

Sensor dome with aerodynamic cladding. Most of the device is in front of the cockpit.

The infrared radiation passes through a window made of zinc sulfide with a protective coating made of BP and DLC , which is 80% transparent. Behind it, the radiation is diverted downwards by the sensor head stabilized in two axes with puff gold plating. This scans very quickly; Due to the purely passive working method, the search volume can be searched more quickly than with an active phased array antenna. A telescope with two magnification options is passed underneath in order to adjust the field of view . The focused field of view is used when the sensor works as an infrared target system, the wide field of view when used as a FLIR. The beam is then deflected backwards by 90 ° by a folding mirror and sent through a derotation prism, which on the one hand removes the periscopic effect and on the other hand rotates the field of view to match the detector field. The beam is then deflected again by around 90 ° by a mirror that tilts at the frame rate in FLIR mode . Now more lenses follow to focus the IR beam on the detector element. A total of more than 90 optical components are passed through, e.g. T. from germanium , gallium arsenide , etc., before the IR energy hits the radiation detector . PIRATE has an autofocus in which the temperature of all optical components in the beam path is measured and the position of the telescope lens is adjusted. For focusing, pressure and temperature are also measured in the beam path, and a number of lenses are moved directly in front of the detector. Print data are delivered via the aircraft's data bus. The angular precision for target tracking is 0.25 mrad (0.0143 °), whereby the angular extent of each pixel is taken into account as a function of the current position of the telescope lens.

The long, linear detector array with 768 channels and eight TDI readout circuits was newly developed for the Eurofighter. The time delay and integration (TDI) improves the signal-to-noise ratio . The mercury cadmium telluride (CMT) quantum detector, according to previous publications (2002 and 2006) in the wavelengths of 3-5 microns and 8-12 microns or work microns 8-11. A scientific publication by the manufacturer consortium from 2008 mentions a CMT with 3–5 µm and 8–14 µm, and a publication by the RAND Corporation in the same year mentions a quantum well infrared photodetector as a CCD. The detector is glued directly to a CMOS - ASIC , which works in the multiplex process . Both are housed in a Dewar vessel , which is kept at a cryogenic operating temperature of 70 K by an engine with the aid of a Stirling cycle .

The data processing is because these very high computing power and a lot of special challenges of memory required to the front-end - data transfer rate of 24 million pixels to cope per second. The data processing was programmed in Ada , and consists of more than 400,000 lines of code. The signals at the detector are converted into a 14-bit data stream and read out with a bandwidth of up to 400 MHz. PIRATE works like a radar in track-while-scan mode with look-up or look-down capability, only without emitting emissions. The distance can be determined purely passively by means of sequential triangulation ( kinematic ranging ). Up to 200 targets can be tracked and prioritized at the same time, only some of which can be tracked and sent via the EFA bus. The angle to the target, distance, signature and characteristic data of the target, dynamic data and measurement errors are sent via the bus. In FLIR mode, the tiltable mirror behind the derotation prism is activated in order to be able to generate 575 image lines per FLIR image.

Signal processing

Note: The software is constantly being improved. The status shown corresponds to about 1996 to 2008. In scientific publications from 2002/2003 z. B. There is talk of using a terrain database in the future when discovering and tracking targets; to generate a synthetic view for the pilot with PIRATE and the terrain model, or to expand the system to a kind of AN / AAQ-37 by integrating further fixed IR sensors .

The data stream is read from the detector field at over 300 MB / s. Non-Uniformity Correction (NUC) of all 768 detector channels is guaranteed by the Gain and Offset Correction (GOC) processor before the video signal is processed for target location and tracking. GOC improves the pixel data through real-time calibration over a wide temperature range, and removes the stray IR light that the detector throws at itself through lenses in front. Furthermore, the data is provided with a time stamp, as well as the scanning position of the sensor head. An automatic gain control based on the scene image and an offset correction follow. The data stream, now 16 bits “wide”, is now ready for target location, tracking and classification.

The signal processing consists of four components. First, the data stream is loaded into a large digital ring buffer for buffering . The data is then loaded into the Detection Processor (DTP). The desire to carry out BVR battles with the IRST requires the highest possible resolution and a large aperture . Since both are limited in practice and the scan area should not fall below a certain size, aerial targets must be located at great distances in the subpixel area. The target's signal will therefore always be “contaminated” with random noise and clutter. A spatial matched filter increases the signal-to-noise ratio (SNR) by declaring a pixel point that exceeds the threshold value to be the target based on a local threshold value. To limit the false alarm rate, it is assumed that the target is among the “hottest” objects in the sky. A two-dimensional optimal filter rejects area targets and optimizes the signal-to-noise ratio of the point target. As a result, the majority of the pixels that are assigned to the background are discarded. Since the background is highly structured, the filter is adapted locally adaptively. The process is optimized to be able to locate targets at great, medium and close distances. Three filters are used for this purpose: Channel 1 locates targets at great distances, channels 2 and 3 at medium and close distances by reducing the resolution by calculating the mean value of 2 × 2 or 4 × 4 pixels. The three channels work simultaneously and independently of each other. The spatial size of the area around a located target is limited to 7 × 7 pixels in order to be processed by the kernel. A duplicate detection assigns detections at different distances (aka channels). Each contact located in this way is assigned a scan position and time so that the contact can also be assigned to consecutive scans. The signal processing divides the scan volume into different elevation areas based on details, flight altitude and height above ground. Each of these areas has different background and goal characteristics. An elevation code is now assigned to each pixel and is updated during a scan. Each elevation area is now assigned an optimal filter based on background clutter and target IR signature, which is calculated between the scan passes of the elevation area. The filters are also adjusted when the detection load or track load command is invoked. Objects that are assigned to the clutter are treated equally separately.

The location processing now works with the locations of the signal processing filter. Close groups of contacts are either separated into different clusters or, if these can be assigned to the same position in consecutive scans, are combined into one contact. This makes use of the fact that a real target is spatially consistent, which is not the case with random noise. The target coordinates are transformed into the aircraft's own system based on the position of the sensor head and the position of the detotation prism. For this purpose, weighted mean values ​​are used, whereby calibration data of the sensor head, statistical inaccuracies of the IRST structure and optics, errors due to the acceleration forces, temperature, mixed values ​​of the internal gyro instrument and the laser gyro of the aircraft and the vibration of the structure of the aircraft and sensor are taken into account. Each cluster is assigned an estimated distance based on its spatial expansion and intensity.

In the next step, the computer tries to extract tracks from the clusters. A movement model of the celestial bodies recognizes the moon and the sun and removes them from the signal processing. Since the spatial extension of these objects is relatively large, precautions are taken to avoid deleting covered targets. It is also examined whether the new contact belongs to a target that has already been located. Assignment can be difficult with densely packed target clouds, since contact could also be between two tracks. For this purpose, Bayes' theorem is used to resolve conflicts through subsequent measurements (i.e., scanning passes). To do this, Reid's multiple hypothesis tracking (MHT) algorithm is used. A hypothesis tree is set up for each contact in which the solutions “false alarm”, “first location of new track” and “location belongs to track no. X” are entered. With every scan and contact, the hypothesis tree branches out further, so that the computer would quickly be overloaded with clusters that are close together, so that they have to be considered separately. The problem is exacerbated because the signal-to-noise ratio is very low due to the desired high location range, and thus a lot of false targets put a strain on the system. For groups that have a linear distribution, a check is carried out to determine whether they are soil objects. If the cluster fulfills a certain persistence, a track is initiated from the cluster, provided that it fulfills the dynamic and infrared signature of a target, in order to avoid hypotheses from getting out of hand. A cluster counting function largely eliminates the generation of hypotheses with noise. The range estimate from earlier is now used to predict the target's movement, with the target also being tracked in elevation and azimuth. The most probable assignment of a position point to measurement points from previous scanning cycles ("association") is realized by a detection window ( gate ), the size of which is calculated dynamically from noise, assumed maximum target speed and acceleration. For example, at long distances to the target, the lateral movement will be negligible, so any hypotheses based on it can be eliminated. At the same time, the probability of each hypothesis is calculated by placing a 7 × 7 pixel gate on each future contact position, in which each pixel point is weighted according to a Gaussian function . The tracks are tracked using a Kalman filter with three states in elevation and azimuth, and a target track is established using MHT. If the target maneuvers, the models are adjusted accordingly. If a target is lost and cannot be detected again even after four consecutive scans, the track is terminated. If a contact can be recorded again, it will be associated normally like the others. The track file is retained when PIRATE changes to a mode other than standby in order to continue with the target accompaniment after switching back. However, if a time limit is exceeded, the track will also be terminated.

All tracks that are persistent are classified. The signal processing tries to classify the track as an air target, cloud or ground clutter. To improve performance, a clutter map of the target area is created, with the help of which the shift in the field of view can be corrected electronically. In this way, the line of sight in the subpixel area can be stabilized, so that future contact is expected at nine possible pixel positions (original position and directly next to it) during the next scan. This method can also be used to verify an initial location, as the signal-to-noise ratio can be lowered, which almost doubles the location range. If the classification cannot be made in one scan, the next one is tried, etc. The classification is supported by the last up to five false associations, the history of the cluster details, dynamic properties of the target, signal characteristics, distance estimation, sensor fusion data from the AIS, altitude of the typhoon, altitude of the track, a feature extraction and movement models of the planets are taken into account. The feature extraction supports the indication of the direction of movement of the target. If the Detection Processor (DTP) sends a command to extract features to the Feature Extraction Processor (FEP), the following process is started: The FEB accesses the image of the target cluster in the ring buffer and loads it after the IR image was recognized. The image spot is processed to extract the object from the background. An algorithm now looks for the corners of a moving object in order to logically connect them with closed contour processing . Since the corners do not always touch, the algorithm connects them if the direction, intensity and shape match. The edges themselves are recognized as a discontinuity in the video image by a 3 × 3 pixel mask. The corner points are stored in a two-dimensional array of numbers, which changes when the target maneuvers. In the next step, it is determined which orientation the flight object is adopting in the image, and thus an approach or distance is present. The result is passed on to the tracker and classifier. The FAR is also depressed by the feature extraction. New objects with closed contours are compared with the clutter map and assigned either to the background or to a target category. Streaming vectors are used to represent the distance to objects and their direction of movement. Since the map is updated with each scan and the aircraft is moving, the coordinates of the world model must be continuously adjusted as the sensor's viewing angle changes. The whole process requires brutal computing power, especially with quick scene changes. This process runs continuously and requires a large part of the computing power of the system. If there are many inquiries in a short time, prioritization takes place, since the image in the ring buffer is overwritten after a certain time. If the processor is overloaded, a warning is issued. Classifying a target takes less than a second. At the same time, it is checked in the case of moving targets whether they could be covered by the background on their way through the world model. Since the distance measurement is not precise enough to determine “in front of” or “behind”, it is simply expected that an object will disappear when two come close. Since the dimensions of the obscuring object are known, it is possible to predict where the target will reappear in order to be recorded again. The principle can be refined if a terrain model is available to determine the spatial depth.

At a distance of 2 km to less than 200 m, the targets are so large that they cover the field of view of the sensor. During the ADAD tests it was already established that a different method of target tracking had to be used in this distance range. Image correlation is used to ensure target tracking . A specific feature of the target, e.g. B. recorded and tracked the cockpit. In subsequent scans, the section of the target video is correlated with new images of the target. This new position is then used to update the track. The picture of the breakpoint is renewed periodically.

The Track Processor (TKP) is the last step in the chain. It creates a target list from the classified and permanently updated tracks, in which target details and priority are noted. The details for each target are updated with each scan. The prioritization is changed when a new target is added or external commands are added. As a rule, the prioritization is carried out by the AIS and its data is adopted. The internal prioritization only takes into account the distance, elevation angle and rate of approach of the targets. In addition, a sequential triangulation ( kinematic ranging ) is now carried out. Angular data and speed to the target, the flight speed and position of the Eurofighter and the measurement times are converted into a Cartesian position of the target and its estimated speed using an adjustment calculation. The obtained 3D track of the targets is transferred to batch processing . The MTT mode continues even if PIRATE switches to another operating mode with the exception of standby , whereby tracks can only be deleted by a time limit if the last position measurement was too long ago. Up to 200 targets can be tracked and prioritized at the same time, only some of which can be tracked and sent via the EFA bus. Angle data, distance, signature and characteristic data, dynamic data and measurement errors are sent via the bus.

In FLIR mode, the image is downscaled from 14 bits to 8 bits using histogram equalization. The contrast is increased and temperature peaks are smoothed out. This is followed by sharpening to strengthen edges. Depending on whether the image is displayed in MHDD, HUD or HMD, PIRATE uses different tables for gamma correction . In the HUD display, a few lines of image have to be cleverly added and deleted to fill the pilot's field of view. A display is possible as a light-hot or dark-hot image. In STT mode, the IDENT function can also be selected, which provides an electronic zoom for a 3 ° × 3 ° or 6 ° × 6 ° field of view. This increases the frame rate to 40 ms (25 fps), and a still image of the target can be displayed. Furthermore , an air-to-ground mode was implemented with Thermal Cueing (TC) in order to be able to track several ships, cars, trains, etc. at the same time. For the pilot, the targets are marked with a “v” on the FLIR image if the signal-to-noise ratio (SNR) exceeds a certain threshold value. In test flights, ships, boats, cars on roads, trains on rails, etc. could easily be located. Upon the pilot's suggestion, the targets are weighted so that the proximity to the aircraft, the flight altitude of the Eurofighter above the ground, and the position of the target in relation to the horizon determine the priority. Test flights were carried out over the sea, over agricultural land, forests, mountains, etc. The results of the passive distance measurement were also verified here . In thermal cueing (TC) mode, angle data of the targets, their distance, intensity and size are passed on via the bus. If a target is lost, it will be adjusted for a certain time after the last known movement before the track is terminated.

Operating modes

IRST attack format : An enemy machine (red) at 35 nm and 3000 ft, and a group of unknown contacts (yellow) at around 50 nm and 1000 ft. The gray target comes via the data link.

PIRATE is usually an infrared targeting system with a track-while-scan mode, but it can also act as a FLIR. In air-to-air operation, PIRATE locates, accompanies, classifies and prioritizes multiple air targets in all target positions, look-down and look-up, and at the same altitude. The sensor is fully integrated into the Attack and Identification System (AIS) . As with the CAPTOR radar, the scan volume of PIRATE in MTT mode is automatically determined by the AIS. By default, PIRATE also serves as a passive missile warning in the front sector of the Eurofighter . In addition, seven operating modes are available, which can be used both in the air-to-air and in the air-to-ground role:

  • Standby mode of the sensor:
    • Standby: mode to save power and delete memory.
  • IRST modes for operation as an infrared target system. The IRST can be used with the radar to improve precision and reduce the false alarm rate. Or to search a volume other than the radar to improve the pilot's situational awareness:
    • Multiple Target Track (MTT): A certain volume is searched for air targets, the targets found are tracked and prioritized. When selecting a destination, the system switches to STT mode.
    • Slaved Acquisition (SACQ): Conception of a target whose position was received via a data link. If the target is detected, the system changes to STT mode.
    • Single Target Tracking (STT): In this mode only one target is tracked. The update rate increases, the target can also be identified in the IDENT submode.
  • FLIR modes to improve landing and visibility of the pilot, as well as for air-to-ground location of potential targets with a (O-Ton 2003) "simplified tracking function". Targets can also be located here on the subpixel level.
    • Landing Aid (LAAD): The FLIR image is projected onto the HUD for take-off and landing aid. 575 lines and 690 pixels per line are displayed.
    • Flying Aid (FLAD): The FLIR image is projected onto the HUD. The Thermal Cueing (TC) function can be activated to locate and track ground targets.
    • Steerable IR Picture on Helmet (SIRPH): The IRST sensor is coupled with the pilot's head movement. The sensor then looks where the pilot is looking, and the FLIR image is projected onto the helmet display . In this mode, the pilot can also have the field of view searched by TC.

Range

The range of the sensor is a well-kept secret of the manufacturer consortium. The RAND Corporation speaks of 50 nm (93 km) against a subsonic target from the front. Troop service speaks of 50 to 80 kilometers, but also considers 150 kilometers to be possible. According to the manufacturer consortium in the International Society for Optical Engineering (SPIE) in 2003, the range is compatible with the air-to-air guided weapons carried. In a publication from 2008, the consortium in SPIE stated that during the test campaigns it was shown that PIRATE possessed similar capabilities as the radar in tracking targets and a similar range. The information fits well with the tender, where an observation area comparable to the radar was required. The USAF also required an infrared target system for the Advanced Tactical Fighter with a detection range of up to 160 nm (288 km), which later fell victim to the budget (AIRST). However, weather conditions have a significant impact on the performance of infrared-based target search and tracking.

The following is an overview table that compares the specified detection ranges of the OLS-35 and AIRST, and the literature and manufacturer information on PIRATE. Since the infrared radiation of the target types is identical, the ratio of the detection ranges does not change. In this way, the location range can be calculated if the distance is specified for only one target type. The launch of a missile is most recognizable because the exhaust plume of the rocket engine is almost 1000 ° C, and a Mach 4-faster missile reaches a temperature of 650 ° C at the stagnation point . A supersonic target of Mach 1.7 still reaches 87 ° C.

The information in Military Avionics Systems (2006) contradicted the specialist literature available in 2003, which indicated range compatibility with the air-to-air guided missiles carried. The information from the US-American RAND Corporation from 2008 is higher, but also contradicts the technical literature available in 2008, which now announced a range similar to radar for the first time. Furthermore, the range was further increased through software updates until 2013. The truth for PIRATE will probably lie somewhere between the information from RAND and the previously planned AIRST for the ATF.

reference Subsonic target (front) Supersonic target Mach 1.7 (front) Mach 4 supersonic target (front) Subsonic target (rear) Launch of a guided missile
RAND Corporation, 2008 ( OLS-35 ) 50 km (27 nm) 54 km (30 nm) over 81 km (45 nm) 90 km (50 nm) over 90 km (50 nm)
Military Avionics Systems, 2006 ( PIRATE ) 72 km (40 nm) 78 km (43 nm) over 117 km (65 nm) 130 km (72 nm) over 130 km (72 nm)
RAND Corporation, 2008 ( PIRATE ) 90 km (50 nm) 98 km (54 nm) over 146 km (81 nm) 163 km (90 nm) over 163 km (90 nm)
Flight International, 1986 ( AIRST ) 160 km (89 nm) over 173 km (96 nm) 259 km (144 nm) almost 288 km (160 nm) 288 km (160 nm)

Web links

Individual evidence

  1. ↑ a b c EFA warms to Rafale IRST. In: FLIGHT INTERNATIONAL. September 12, 1990, accessed June 12, 2014 .
  2. ↑ a b c Eurofirst reveals EFA IRST. (PDF) In: FLIGHT INTERNATIONAL. August 26, 1989, accessed June 12, 2014 .
  3. ↑ a b c UK gets new air-defense sensor. In: FLIGHT INTERNATIONAL. July 30, 1993, accessed June 12, 2014 .
  4. ↑ a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac Luigi Enrico Guzzetti, Livio Busnelli: EF2000 PIRATE test flights campaign . In: Proc. SPIE 7109, Image and Signal Processing for Remote Sensing XIV, 71090N . October 10, 2008.
  5. ↑ a b c d e f g h i j k l m n o p q r s t Pilkington Thorn Optronics: The Fusion of an IR Search and Track With Existing Sensors To Provide a Tracking System for Low Observability Targets . In: Multi-Sensor Multi-Target Data Fusion, Tracking and Identification Techniques for Guidance and Control Applications (AGARD-AG-337) . October 1996.
  6. ↑ a b c d e f g h i j k l m n o p q Manfred Scheuer: FLIR / IRST of the European fighter aircraft . In: Proc. SPIE 4714, Acquisition, Tracking, and Pointing XVI . July 1, 2002.
  7. ↑ a b c d U.S. Navy Follows UK Lead On Infrared Systems. In: Aviation Week & Space Technology. July 13, 2013, accessed June 20, 2014 .
  8. ↑ Eurofighter's final systems snag on cost. In: FLIGHT INTERNATIONAL. June 26, 1991, accessed June 12, 2014 .
  9. ↑ a b c The Market for Airborne Electro-Optical Systems . In: Forecast International . No. 10 , 2011 (English, forecastinternational.com [PDF; accessed June 12, 2014]).
  10. ^ Norman Friedman: The Naval Institute Guide to World Naval Weapons Systems, 1997-1998 . US Naval Inst Pr, 2007, ISBN 1-55750-268-4 , pp. 316 .
  11. ↑ a b c Optronics in Integrated Air Defense. (PDF) (No longer available online.) In: RD Hoyle; Business Development Manager Pilkington Optronic. May 22, 2000, archived from the original on June 10, 2015 ; accessed on June 12, 2014 . Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / ftp.rta.nato.int
  12. ↑ a b c d e Luigi E. Guzzetti; Livio Busnelli: Infrared search and track and imaging system: testing in the laboratory and during flight . In: Proc. SPIE 7300, Infrared Imaging Systems: Design, Analysis, Modeling, and Testing XX . April 22, 2009.
  13. ^ Doug Richardson: Stealth Warplanes: Deception, Evasion, and Concealment in the Air . Zenith Press, 2001, ISBN 0-7603-1051-3 .
  14. ↑ John A. Malas: F-22 radar development . In: Proceedings of the IEEE 1997 National Aerospace and Electronics Conference . July 14, 1997.
  15. ^ 1st Eurofighter with PIRATE IRST Radar Delivered to Italian Air Force. (No longer available online.) In: air-attack.com. August 2, 2007, archived from the original on October 22, 2014 ; accessed on May 30, 2014 (English). Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / www.air-attack.com
  16. ↑ USAF pilot describes IAF Su-30MKI performance at Red Flag-08. In: Flightglobal. November 5, 2008, archived from the original on December 19, 2008 ; accessed on June 12, 2014 .
  17. ↑ Raytheon Looks At Options For Long-Range AIM-9. In: Aviation Week & Space Technology. June 19, 2013, accessed June 20, 2014 .
  18. ↑ sensors. (No longer available online.) In: typhoon.starstreak.net. May 30, 2014, archived from the original on November 27, 2014 ; accessed on May 30, 2014 (English). Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / typhoon.starstreak.net
  19. ↑ a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad ae af ag Boyd Cook: PIRATE. The IRST for Eurofighter TYPHOON . In: Proc. SPIE 4820, Infrared Technology and Applications XXVIII, 897 . January 1, 2003.
  20. ↑ IRST (PIRATE) EUROFIGHTER TYPHOON INFRARED SEARCH AND TRACK. (PDF) (No longer available online.) In: Selex ES. 2014, archived from the original on June 26, 2014 ; accessed on June 20, 2014 . Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / www.selex-es.com
  21. ^ A b c Ian Moir, Allan G. Seabridge: Military Avionics Systems . In: John Wiley & Sons Ltd . 2006.
  22. ↑ Srivastaval et al .: Airborne Infrared Search and Track Systems . In: Defense Science Journal . tape 57 , no. September 5 , 2007.
  23. ↑ a b c d e Air Combat Past, Present and Future. (PDF) In: RAND. August 2008, accessed May 30, 2014 .
  24. ↑ a b c d The Eurofighter "Typhoon" (VII) Radar and self-protection . In: Troop service . No. 06 , 2008 ( bmlv.gv.at [accessed on May 30, 2014]).
  25. ^ A b Hughes and GE join forces on ATF. In: Flightglobal. June 11, 1986, accessed May 30, 2014 .