GB2232315A - Electromagnetically detecting a dielectric interface in a region below an aircraft - Google Patents

Abstract

The presence of one or more dielectric interfaces (8, 9) in a region below an aircraft (e.g. a helicopter (1)) having a rotor (main helicopter rotor (5)) is detected from the aircraft by transmitting radar pulses from a transmitter (2) downwardly towards the region and detecting, with a detector (3), radar pulses reflected upwardly from the region. The detector output is repeatedly sampled and the presence of the or each interface (8, 9) is determined from a corresponding spike (R1, R2) in the radar returns at the detector output. In order to remove or eliminate the effect of radar clutter contribution from the aircraft rotor which can mask the reflection spike(s) (R1, R2), the generation of the radar returns is effected in synchronism with the rotation of the rotor (5) such that the rotor clutter contribution to the radar returns is substantially invariant from one radar return to another. As a result, the radar clutter component can be at least partially separated, for example by high pass filtering of the sampled detector output. In this way, the spike(s) (R1, R2) can be detected more readily. The apparatus and method disclosed herein find particular application to measuring the thickness of an ice layer (6) in a polar region, which layer gives rise to two spikes (R1, R2) in the radar return. The ice thickness is determined from the time lapse between detection of the two spikes. In this way aerial surveys of thicker ice layers are made possible. <IMAGE>

Description

ELECTROMAGNETICALLY DETECTING A DIELECTRIC INTERFACE IN A REGION BELOW AN AIRCRAFT This invention relates in general to electromagnetically detecting a dielectric interface in a region below an aircraft, more particularly where the aircraft has a rotor producing electromagnetic clutter.

A dielectric interface arises where two materials of differing dielectric constant interface with one another. For example, water, ice and air have significantly different dielectric constants. Therefore, in a polar region, an ice layer on top of a body of water will provide two dielectric interfaces, namely an air/ice interface and an ice/water interface. Other examples of dielectric interface arise in the case of buried pipelines or subterranean water reservoirs in the desert. In each case, the or each dielectric interface causes a portion of incident electromagnetic radiation to be reflected by the interface and the present invention relies upon detection of reflected electromagnetic radiation for detecting the presence of the interface(s).A preferred, though nonlimiting, application is in determining the thickness of an ice layer in a polar region by detecting reflected electromagnetic energy from the air/ice and ice/water interfaces, and calculating the ice thickness from the time period between detection of a first reflection from the air/ice interface and a second reflection from the ice/water interface.

"Airrraft" in the context of this specification refers to any vehicle capable of flying but relates in particular, though not exclusively, to helicopters. "Rotor" as used herein refers to any rotatable component which is used in an aircraft and which can reflect electromagnetic radiation. In the case of a helicopter which in general has both a main rotor assembly and a tail rotor assembly, "rotor" refers to whichever rotor produces the greater clutter. This will depend upon the siting of the transmitting and receiving elements of the radar system and the geometry of the helicopter but usually it is the main rotor blade assembly generating the lift force enabling the helicopter to fly which is the main contributor to radar clutter. This is because the tail rotor assembly is considerably snaller in size than the main rotor assembly.It is further remarked that in a helicopter, the rotation of the tail rotor assembly is often synchrnnised with that of the main rotor assembly. Where this synchronisatron is on a one-to-one basis, adoption of the invention as described belch to reduce or overcome radar clutter produced by whichever of the main ard tail rotor assemblies is the major clutter contributor will automatically have corresponding advantageous effects as regards radar clutter produced by the other such rotor assembly. From the foregoing, it will be appreciated that in the context of a helicopter, "rotor" will be taken to refer usually to the main rotor blade assembly. For a propeller driven aircraft, "rotor" relates to the or each propeller blade assembly.

An airborne impulse radar system is known that remotely measures the thickness of ice. With reference to Figure 1 of the accompanying drawings, the system is installed on a helicopter 1 and comprises two antennas 2, 3 that are mounted on the outside of the helicopter, as shown in Figure 1, and an instrumentation package 4 that is mounted inside the helicopter. Of course, instead of two antennas, a single, transmit-receive, antenna could be used. One of the antennas (antenna 2) is a transmitter that radiates electromagnetic (radar) energy, while the other 3 is a receiver that detects electromagnetic energy. The rotor blades of the helicopter's rotor assembly are denoted by reference number 5.

Based on electrical signals generated by the instrument package 4 inside the helicopter, the transmitting antenna radiates a short duration, e.g. 10 nanosecond, burst of radar energy down towards the top surface of the ice layer (6) along a transmission path 7. Some of the energy in this pulse is reflected from the air/ice interface 8. The reflected energy travels along first reflection path 10 and arrives back at the receiving antenna 3. The remainder of the energy penetrates through the air/ice interface and propagates through the ice to the ice/sea water interface 9, from which a portion of the energy reflects off the ice/sea water interface and travels along second reflection path 11. The remaining portion of the energy passes through this interface and is rapidly attenuated by the subjacent conductive sea water.The instrumentation package 4 measures radar energy versus time and monitors the arrival at the receiving antenna 3 of energy reflected from the air/ice and ice/sea water interfaces 8, 9. Knowing the velocity of the radar signal in ice, the thickness can be computed using the formula: Thickness = F x Velocity of signal in ice x Time delay between reflection arrivals.

An idealized voltage signal that is generated by the receiving antenna 3, based on the reflected electromagnetic energy resulting from the transmission of a single radar pulse towards the ice layer, is shown in Figure 2. This plot of radar energy versus time is called a radar return.

The plot shows voltage versus time as it is sensed at the output terminals of the receiving antenna. This return is recorded in the instrumentation package 4 mounted inside the helicopter 1. It could be stored in analogue format but in this example it is stored as a series of digitised values, called radar samples. In this case, a radar return is a series of radar samples collected over time. Thus, each sample of a radar return has a time delay associated with it where the time delay is the elapsed time between the instant the radar pulse is fed to the transmitter and the time the radar sample is collected at the receiver.

The two spikes (R1, R2) in the radar return represent the reflected radar energy from the ice layer top surface 8 and the ice layer bottom surface 9. The elapsed time between these two pulses is proportional to the thickness of the ice layer 6.

In actual operation, the radar return is not as simple as in the idealized case depicted in Figure 2. This is because the radar transmitter 2 does not radiate a narrow beam of electromagnetic energy solely along the transmission path in Figure 1. The antenna pattern is much broader and energy is radiated in all directions, including towards the helicopter fuselage and the main helicopter rotor assembly. An approximate pattern showing the amplitude of the radar energy radiated by the antenna in every radial direction is shown in Figure 3. The length of each arrow in this plot represents the peak amplitude of the energy radiated in the radial direction from the antenna 2. Measurements used to generate this type of plot are collected at a fixed distance from the antenna.

Likewise, the receiving antenna is more sensitive to incoming radar energy in one direction. However, incident radar energy from all directions influences the power level detected at the antenna terminals.

The directional sensitivity of the receiving antenna can also be depicted as shown in Figure 3.

Since the transmitting antenna 3 radiates energy towards the helicopter fuselage and the rotor blades and the receiver collects data from all directions, the actual return signal at the receiving antenna 3 includes components due to energy that is reflected from the helicopter body, its main rotor assembly, tail rotor assembly, the ice layer interfaces, and broadcast transmissions. These signals superimpose with each other resulting in a radar trace that resembles that shown in Figure 4. The unwanted reflections from the helicopter and rotor are termed radar clutter, and are a noise effect. Clutter usually is larger in amplitude than the ice layer top surface reflection or the ice layer bottom reflection. As a result, the air/ice and ice/sea water interfaces are difficult to locate on the radar return trace.

Fortunately, signal processing can be used to remove the clutter generated by the helicopter fuselage. The reflections from the fuselage are constant from one radar return to another because the radar paths between the antennas 2, 3 and the helicopter fuselage remain the same at all times. However, the distances between the antennas and ice surfaces are constantly changing. By using a filter to compare successive radar returns, the clutter generated by non-moving objects, that is, clutter that is constant with respect to the radar returns, can be subtracted from the radar returns while changing radar reflections caused by moving objects are preserved, as described in the next paragraph. This is a common filtering technique called high pass filtering.

On the other hand, rotor blades are moving relative to the antenna. Thus in general, the rotor radar clutter is not constant from return trace to return trace. Figures 5a and 5b show two possible rotor positions and paths for the radar reflections off the main rotor. Thus, common filtering techniques applied between successive radar traces, whilst effective to remove radar clutter due to the helicopter fuselage, cannot eliminate the clutter caused by the rotating rotor blades.

Rotor clutter limits the ability of the radar system to measure thick sea ice. The ice is slightly conductive and, as a result, absorbs radar energy. Consequently, as the ice becomes thicker, the power level of the ice/sea water reflection decreases. Geometric spreading of the radar energy as it travels also reduces the power level of the bottom reflection as the ice thickness increases. Eventually with ice over 30 to 40 ft.

thick, the power level of the clutter becomes larger than the bottom reflection: thus, the bottom reflection is lost in the noise. By reducing the rotor clutter, lower amplitude bottom reflections from thicker ice could be detected. Prior to the present invention, it was not known how this might be done.

One way to solve the rotor clutter problem would be to design a transmitting antenna that does not radiate any energy toward the rotor blades or design a receiving antenna that accepts only radar energy incident from directly below the helicopter. Unfortunately, there are sound physical reasons why such a set of antennas is difficult, if not impossible, to develop and use on the present arrangement.

Anther possible method of increasing depth of radar penetration into the ice and increase the maximum measurement thickness would be to increase the power of the radiated pulse. However, as radar power increases, the radar returns from the reflection increase in approximately the same prc > ortion as the clutter. Thus, the ratio of return signal to clutter remits approximately constant.

The present invention seeks to provide an alternative and effective solJtion to the problem of removing the effect of rotor radar clutter, which can be readily realised in practice.

According to the present invention there is provided a method of electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagnetic clutter, comprising the steps of: (a) transmitting electromagnetic pulses from the aircraft towards said region:: (b) detecting in the aircraft reflected electromagnetic pulses from the direction of said region, by means of a detector, said detector also detecting electromagnetic rotor clutter; (c) generating radar returns of which the instantaneous magnitude represents the time varying strength of the detector output at a particular delay following the trans mission of an electromagnetic pulse from the aircraft;; (d) synchronising the generation of the radar returns with the rotation of the rotor such that the instantaneous magnitude of the radar returns at each particular time point during the duration of the radar returns corresponds with substan tially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns is substantially invariant from one radar return to another; (e) at least partially separating the substantially invariant electromagnetic rotor clutter contribution from the radar returns to provide relatively clutter-free returns; and (f) detecting in at least one of the relatively clutter-free radar returns the presence of an interface reflection spike resulting from the presence of said dielectric interface in said region.

A preferred way of performing the invention is one wherein each radar return is generated by repeatedly sampling the detector output at different time delays after the transmission of a respective corresponding number of successive radar pulses, the respective such samples combining to form said radar return, and wherein each sample corresponding with a respective said time delay is subjected to respective high pass filtering, so as to at least partially separate the invariant electromagnetic clutter contribution.

According to a second aspect of the invention there is provided a method of electromagnetically measuring the thickness of an ice layer in a polar region from an aircraft having a rotor producing electromagnetic clutter, comprising the steps of: (a' transmitting electromagnetic pulses from the aircraft towards the ice layer; (b detecting in the aircraft reflected electromagnetic pulses from the upper and lower surfaces of the ice layer by means of a detector, said detector also detecting electromagnetic rotor clutter; (c' sampling the detector output at a first given time delay after the transmission of an electromagnetic pulse and at a particular angular position of the rotor;; (d repeating step (c) a number of times where, each time, the sampling is effected at a respective, different, time delay after the transmission of an electromagnetic pulse, so that the samples from steps (c) and (d) cover a range of sequentially increasing time delays, said samples from steps (c) and (d) combining to form a radar return of which the instantaneous magnitude represents the time-varying strength of the detector output at a particular delay following the transmission of an electromagnetic pulse from the air craft; ; (e repeating steps (c) and (d) to provide further radar returns such that the instantaneous magnitude of the radar returns at each particular time point during the duration of the radar returns always corresponds with substantially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns is substantially invariant from one radar return to another; (f at least partially separating the substantially invariant electromagnetic rotor clutter contribution from the radar returns to provide relatively clutter-free returns; (g) monitoring the relatively clutter-free returns for the presence of interface reflection spikes indicative of the detection of reflected electromagnetic pulses from said upper and lower surfaces, respectively, of the ice layer;; and (h) determining the ice layer thickness from the time difference between detection of the spikes.

Because all the samples collected at the same time point in the radar returns are collected substantially in the same angular position of the rotor, the rotor clutter will be substantially identical between successive samples at the same time points, and so can be eliminated from the return trace, for example by high pass filtering, to enable the ice interface spikes to be more readily recognisable.

The invention may be put into effect by transmitting rapidlyrepeated electromagnetic pulses from the aircraft towards the ice layer, receiving composite electromagnetic signal pulses comprising static clutter from the aircraft, variable clutter from the rotor and reflections from the upper ice surface and from the lower ice surface, electronically converting these received high-frequency signal pulses into a lower frequency range radar signal or return suitable for digitization, processing and display, digitizing the lower-frequency radar returns, and processing and displaying the radar signal to indicate the delay between upper and lower reflections.

The electronic conversion of the received high-frequency signal pulses into a lower frequency range may be accomplished by a sampling procedure, where the rapidly-repeated received signal pulses are respectively sampled, in each case at sequentially-increasing values of delay after transmission of the corresponding pulse. If a single sample is derived in this way for each successive pulse, the sampler output reconstructs, step by step, the shape of the received high-frequency signal or return, from the several samples.

Without the improvement arising from the invention, the low-frequency signal would change as the variable rotor clutter changes.

The rate of change of the low frequency return depends on the relative speeds of rotor rotation and of reconstruction of the low-frequency radar output signa (or 'radar scan'). This yields two methods for eliminating the effect of the variable rotor clutter.

The first is to ensure that the radar scan is only built up from signals derived at a single rotor position. Thus if, for example, 512 samples are required to construct the low-frequency signal, then 512/N rotor rotations will be required, where N is the number of symmetrical blades making up the rotor. Thus throughout the radar scan there is no effective change in rotor position.

A second, faster, method is to ensure that the rate of radar scanning is an integer sub-multiple of the rotor rotation rate. In this case though the position of the rotor varies during the scan, the rotor position for each point of the scan is the same from scan to scan. Thus though the rotor clutter contribution varies from point to point in each scan it is static for each point from scan to scan. Digital filtering techniques allow the rejection of signals which are constant or vary slowly compared with those which are to be detected. Thus both static clutter, and rotor clutter which has been made static as a result of the techniques disclosed herein, can be rejected. This method requires continuous measurement of rotor position to ensure that the scan bears a constant phase relation to it.

A variation of the second method exists, whereby the rotor position is marked at a particular orientation, and the radar scan is begun as this position is passed. Provided that the rotor speed is nearly constant (which is a reasonable assumption for helicopters), the rotor position at any point during the scan is predictable to within narrow margins, and the effect of the second method above will be approximated to.

These methods require that the delay between transmission and sampling should be determined exactly or approximately from the position of the rotor. Provided that the high-frequency pulse repetition rate (PRR) is very high compared with the radar scan rate and rotor rotation rate, the absolute timing of transmission and sampling instants are of no consequence. However, if the PRR is sufficiently slow that the rotor position changes significantly from one pulse to the next, then the absolute transmission and sapling instants must be locked to the radar scanning process.

Acceding to the invention from a third aspect, there is provided apparatus for use in electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagnetic clutter, comprising: - cleans arranged to transmit electromagnetic pulses from the aircraft towards said region and to detect in the aircraft reflected electromagnetic pulses from the direction of said region;; - oceans arranged to generate radar returns of which the Instantaneous magnitude represents the detected electro magnetic pulse strength at a particular delay following the transmission of an electromagnetic pulse from the aircraft, said radar returns generating means including synchronising means arranged to synchronise the generation of the radar returns with the rotation of the rotor such that the instant aneous magnitude of the radar returns at each particular time joint during the duration of the radar returns always corresponds with substantially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns, resulting from detection of electro magnetic rotor clutter by the transmitting/detecting means, is substantially invariant from one radar return to another; - weans arranged at least partially to separate the substan tially invariant electromagnetic rotor clutter contribution from the radar returns to provide clutter-free returns; and - means arranged to detect in at least one of the relatively clutter-free radar returns the presence of an interface reflection spike resulting from the presence of said dielectric interface in said region.

The means arranged at least partially to separate the invariant electromagnetic clutter contribution from the radar returns preferably comprises high pass filtering means. The cut-off frequency(ies) of the high pass filter means is selected so as to remove from the radar return trace the clutter component but pass the higher frequency component containing the return spikes.

In a preferred arrangement adapted for digital sampling, the signal processing means comprises a clock pulse generator arranged to cause the transmitting/detecting means to transmit said electromagnetic Xulses towards the said region, a sampler, a delay generator responsive to the clock pulse generator output and arranged to cause the sampler to sample the detected electromagnetic pulse strength repeatedly at different time delays after the transmission of a respective corresponding number of radar pulses, the respective such samples combining to form said radar returns, and a plurality of high pass filters, each arranged to subject a said sample corresponding with a respective said given time delay to high pass filtering, so as at least partially to separate the invariant electromagnetic clutter contribution.

In order to correct for errors due to small timing variations in the detector output sampling (variations in time delay), the delay generator is arranged to cause the sampler to sample the detected electromagnetic pulse strength a predetermined number of times at each different given time delay, and the signal processing means further comprises an averager for averaging the plurality of detected electromagnetic pulse strengths at each different given time delay, these average values combining to form a radar return.

For a better understanding of the invention and to show how the same may be carried into effect, reference will now be made, by way of example, to the accompanying drawings, in which Figures 1 to 4, 5A and 5B have already been described and in which: Figure 6 shows a typical digitised return trace obtained according to one preferred way of performing the present invention; Figure 7 indicates the sampling procedure used to assemble a return trace of the kind shown in Figure 6; Figure 8 is a block diagram of a preferred circuit arrangement for generating the return traces; and Figure 9 is a block diagram of a preferred circuit arrangement for processing the return traces so as to remove or reduce the radar clutter component from the return traces.

Firstly, a practical example will be discussed in which all sampling is effected at the same rotor angular position. In this example, the radar system is deployed in a helicopter, in the same manner as described above with reference to Figure 1. The helicopter has a twinbladed rotor system which rotates at 5.4 Hertz (324 r.p.m.). This rate of rotation is maintained by an automatic throttle control on the jet engines powering the helicopter rotor. This automatic governor maintains a fairly constant rate (+/- 10 > ) of rotation under most flight conditions.

Since the radar cannot distinguish between one blade and the other blade, the blades pass over the transmitting antenna at a rate of 10.8 Hertz (2 x 5.4). The exact number of blades and the rate of rotation varies from one helicopter design to another, but is constant within one design. The twin-bladed rotor has been used as an example only: other rates of rotation and number of blades could of course also be used instead.

The radar system emits radar pulses and samples radar reflections at a rate of 50,000 Hertz. Since the radar is pulsing approximately 4600 times faster than the 10.8 Hertz rate that the rotor blades are passing over the helicopter, a one-to-one correspondence between the rotor blade rotation rate and the pulsing rate (or sampling rate) would slow down the rate at which the radar data is collected by approximately 4600 times.

Even with the radar pulsing and sampling at 50,000 Hertz, the helicopter must fly slower than 5 knots to collect enough data to generate a complete measure of the ice thickness. If the pulsing rate were reduced to 10.8 Hertz, the survey speed would have to be decreased by a factor of 4600 to .001 knots. Such a slow surveying speed is not desirable.

While a 10.8 Hertz pulsing rate would increase stability of the clutter, it also would decrease the quality of the sampled radar data. Due to the way in which the data is sampled and digitized in this particular example of the radar system, samples from 3042 pulses of the radar are used to generate one sampled and digitized radar return. If these 3042 separate pieces of a radar return are not collected over a small patch of ice, the assembled trace would represent a range of ice thicknesses and the data quality would be poor.

As the helicopter flies faster, the 3042 portions of 3042 different radar trace will be collected over a larger section of ice. If the ice thickness, surface topography, or the helicopter altitude changes approximately one foot during the time period required to collect one assembled radar return, the reassembled radar trace will not represent a unique section of ice and, as a result, will not yield a measure of ice thickness than can be easily interpreted. The precise data sampling procedure used will be discussed in more detail hereinbelow.

With rotor synchronization set such that the tramsitter fires only with the main rotor always in one position and with a 10.8 Hertz radar pulsing rate. the radar system would require over six minutes to sample 3042 pulses and reassemble one digitized radar trace. During a six-minute period, the helicopter pilot cannot maintain altitude or position over the ice within one foot so the data quality would be poor even though clutter stability has been improved. Therefore, making the clutter constant during each radar pulse of the radar system by synchronizing the pulse transmission or sampling with a single fixed rotor position may not improve data quality, even though clutter stability has been improved.

A trace practical embodiment of the invention will now be described in which clutter can be stabilized and radar data quality preserved by synchronizing the data collection process with the rotor angular position in a manner that enables the rotor clutter to be predictable but not constant from each sample to the next. This is accomplished by timing the pulsing of the radar based on the position of the rotor blade. Before a more practical synchronization method can be described, an understandine of how the radar system collects or samples the reflected voltage signal from the receiving antenna 3 is required. As shown in Figure 6, eac radar trace or return is composed of 507 digital samples.

These 507 14-bit numbers represent the voltage signal measured at the receiving antenna 3 over a time period of approximately 250 nanoseconds (250 millionths of one second). Therefore, the elapsed time between each sample is approximately .5 nanoseconds. This implies that the radar would have to digitize and store on magnetic tape 25 million 14-bit numbers every second if every pulse of the radar system were completely sampled (50,000 pulses per second x 507 samples per pulse). This sampling rate far exceeds current echnlogy: consequently, a slower sampling method iE; utilized.

Rather than collecting 507 samples each time the radar is pulsed, the radar system samples only one portion of each pulse. At each successive pulse of the radar, the sample is collected at an increasing delay from the instant the radar is pulsed. Assuming that successive radar pulses do not change drastically, 507 samples from 507 different radar pulses can be assembled to form one sampled radar trace. The assumption of slow changes in successive radar pulses is valid if the helicopter flies slowly, at approximately 5 knots. At a speed of five knots, the helicopter moves forward less than 2 ft during the time period that one complete sampled radar trace is generated.

The sampling process proceeds as follows and is illustrated in Figure 7. The radar pulses the transmitter 2 (pulse 1) and then collects a sampled data point (point 1) .5 nanoseconds later at the receiving antenna 3. The radar transmitting antenna 2 is then fired one more time (pulse 2) and after a delay of 1 nanosecond, another data point (point 2) is collected at the receiving antenna by the sampler. This is repeated 507 times and the delay between the pulsing of the transmitter and activation of the sampler is increased each time by .5 nanoseconds. This is repeated until the last data point (point 507) is collected at a delay of 253.5 nanoseconds.

This sampling process differs from that of the preceding embodiment in that in that case, because of the one-to-one correspondence between the rotor blade rotation rate and the firing and sampling rate, only a single sample is collected from each group of approximately 4600 successive radar pulses. In contrast, in the sampling method disclosed with reference to Figures 6 and 7, 507 samples are collected from 507 successive radar pulses.

This sampling method of Figures 6 and 7 requires accurate timing control of the sampling process. Inevitably, some small errors are introduced by small variations in relative rotor position from return to return and some samples are collected several fractions of a nanosecond too early and some are collected too late. To reduce the effect of these timing errors, the radar system preferably averages several samples which are collected at the same delay. In one preferred arrangement, the radar system averages six samples before the delay is increased. Therefore, the radar collects six samples from six successive radar pulses at 0.5 nanoseconds delay, averages them, and digitizes the result.During the generation of the six transmitted pulses and the collection of the corresponding six samples, the rotor position hardly changes ( 0.23 , at a rotational speed of 324 r.p.m.) so that the change in rotor clutter contribution is small during the averaging. Then it collects six samples at 1.0 nanosecond delay from the next six successive pulses, averages them, and digitizes the result. This process of six collections is repeated 507 times with an increased delay of .5 nanoseconds after every sixth time the radar is pulsed. Therefore, each digitized or sampled radar trace contains 507 values that each have been produced from six pulses of the radar system. As a result, each digitized radar trace is reconstructed from 3042 separate pulses (507 x 6) of the radar transmitter.

The clutter contribution of each of the six samples in each set changes by a small amount because the rotor moves slightly between samples.

However, as successive sets of samples at the same delay (six samples per set) are collected as the rotor rotates, the clutter caused by the rotor in corresponding samples in each such set will be substantially the same. As a result, when the samples'in each such set of samples at the same delay are summed and averaged, the rotor clutter contribution to the average will also be substantially the same for successive radar returns and may be easily removed by filtering.

Assuming that no time is wasted storing the data, and every pulse of the radar is sampled once, the radar generates one complete radar trace every .061 seconds (3042 pulses to generate one trace with 50,000 pulses being generated every second or 3042/50,000 - .061 seconds). This corresponds to a rate of 16.4 reassembled radar traces generated every second. Due to internal housekeeping functions by the microprocessor that controls the sampling process, the actual rate in a practical embodiment will be lower, e.g. 11 radar traces per second.

This sampling process can be controlled so the noise effect of the helicopter rotor can be reduced significantly. As mentioned previously, this can be achieved by pulsing the radar when the rotor is in one fixed position. Unfortunately, as explained above, this is not the preferred arrangement because it slows down the data collection rate and results in poor data quality because the reassembled radar trace is a smear of data points collected from radar pulses generated at a range of helicopter heights and ice thicknesses. A more practical approach is to sample the radar returns in such a manner that the rotor clutter at each of the 507 sample points is constant between adjacent traces but not constant along the 507 points that constitute one trace.In one example, the first data point in all the radar traces would be sampled with the rotor offset zero degrees from the fore/aft axis of the helicopter. The second data point in all the radar traces would be sampled with the rotor at 5 degrees.

The third point could then be collected at 10 degrees. The clutter at position 1 would be constant on the reconstructed traces, yet the rotor clutter at data point 1 would be different than the rotor' clutter at data points 2 and 3.

The fact that clutter changes along the trace (from data position 1 to data position 507) does not reduce the ability of signal processing algorithms to eliminate rotor clutter. Filters to subtract steady clutter returns are implemented by analyzing data at constant delay.

Constant delay in this case implies constant data position. For example, at position 1, the clutter from the rotor is constant because the data located in position 1 is always collected with the rotor in one defined position. By using 507 digital filters to process the data collected at each of the 507 data positions, the constant clutter at each data position can be subtracted from radar data.

The cut-off frequency of each high pass filter needs to be selected so that the filter will pass the expected high frequency component resulting from spikes in the return trace caused by the detection of reflection radar pulses from the upper and lower surfaces of the ice layer but filter out the low frequency component caused by the invariant radar clutter. It might be expected that when flying at a substantially fixed height above an ice layer of substantially uniform thickness, the filters would not be able to distinguish adequately between the higher frequency component in the return trace due to the interface return spikes and the lower frequency component due to the radar clutter. However, in practice, no pilot can fly a helicopter at a sufficiently constant height and/or the ice layer would not be sufficiently uniform, such that a cut-off frequency could not be selected for each filter that would distinguish adequately between the two signal components. Selection of an appropriate value for the cut-off frequency is a matter for routine design and experimentation within the ordinary skill of the competent man and need not be further described herein.

For synchronization, the rotor position for the sample at each time delay needs to be substantially constant, but the spacings between the several rotor positions corresponding with the samples making up each return need not be uniform, providing they are the same from radar return to radar return. Each sample at a particular delay must be taken at a substantiall constant position in the rotor rotation, which may for convenience be an integer submultiple of 360 degrees from a defined datum.

Each scan contains 507 separate range samples, and complete scan traces could be generated at periods which are integer multiples of the rotor position repetition period. In one constructional example, there are two rotor blades which rotate every .185 seconds (5.4 Hertz) and because of symmetry, the blade pattern repeats every .0925 seconds (10.8 Hertz).

Therefore, the radar could be configured to generate a sampled radar trace every .0925 seconds (10.8 Hertz), which is one-half of a blade rotation, .1850 seconds (5.4 Hertz), which is one complete rotation of the blades, .2275 seconds 3.6 Hertz), which is one and one-half rotations, and so on.

The rate at which the reconstructed radar traces are generated must be selected from one of these discrete frequency steps. 8y way of example, a radar system which pulses at 50,000 Hertz results in sampled radar traces being generated at a rate of 11 Hertz. Therefore, the system could operate at the 10.8 Hertz rate required for rotor blade synchronization.

The speed with which the impulse radar system can be flown while collecting data is dependent on how rapidly the system generates sampled radar traces. A doubling of the sampling rate would double the speed at which the radar can be flown. If the reconstructed trace generation rate is increased above 10.8 Hertz (period of .0925 seconds), the radar will be collecting more than one sampled radar trace before the rotor pattern is repeated.

In this situation, the clutter on successive radar traces cannot be synchronized because the rotor position needed for synchronization is not being repeated fast enough. One possible solution to this problem is to create two different rotor positions for the start of sample point number 1. One set of traces could be collected using a starting point of O degrees from the fore/aft axis of the helicopter and an ending point before the rotor reaches 90 degrees. The next trace could be started at 90 degrees and finished before 180 degrees. As a result, the clutter for the 0 degree traces would be different from the 90 degree traces. All the traces collected with the 0 degree starting point should have constant clutter and all the traces collected at the 90 degree starting point should have a different yet constant clutter pattern.During digital signal processing, one set of 507 digital filters would be used to process the 0 degree traces and another set of 507 filters, probably using an identical algorithm, would be used to process the 90 degree traces.

As the pulse repetition rate and sampling rate are increased even further, this concept could be extended to any number of starting positions. In the case of three starting positions based on 0 degrees, 60 degrees and 120 degrees, three separate sets of 507 filters using identical or different algorithms could be utilized, and the radar would have to generate reassembled radar returns at a rate of 21.6 Hertz. In this case, data could be collected when flying at 10 knots. An attempt to implement this increase in pulse repetition would require a faster digital tape deck to record the data and a higher speed computer to process the data for real time display. Increasing the pulsing and sampling of the antenna systems would not be a significant design problem.

As regards sensing the rotor blade position, there are two methods of controlling the radar system based on the rotor angular position. One method works even if the rate of rotor rotation changes, but it is more difficult to implement than the second method that requires a substantially steady rate of rotor rotation. In the first method, the rotor angular position is used to define the instant that each data point is pulsed and/or sampled. This requires a sensor to monitor rotor blade position with respect to a given orientation. For example, if the fore/aft rotor blade position were the reference position, this sensor could generate an output measured in degrees of rotation from that reference position. As the rotor blade passes through a set angular position, the radar would pulse and sample.Every pulse and sample of the radar would be controlled by the rotor position sensor.

One possible sensor that meets these measurement specifications is an optica' encoder. This device would be geared into the rotor head and generates a digital ouput measured in degrees. In this case, since the radar could be pulsed and the return sampled based on the position of the blade, the rate of rotor rotation does not significantly affect the performance of the system.

The second method relies on the main rotor on the helicopter spinning at a substantially constant rate, which can be used to simplify the synchronization of the radar with the rotor blade angular position. In the previous case, each of the 3042 transmitted radar pulses and/or sampling events is initiated by the sensor that monitors blade position.

In the simplified case, the blade position is used to start the pulsing and/or sampling process. The sampling and pulsing then continues at a constant rate until one complete sampled radar trace is generated. When the sampled return is complete and all 3042 data points have been collected, the transmitter can be shut off. Another series of pulses from the transmitter will be initiated when the blade position is once again in the correct starting position. Alternatively, after all 3042 data points have been collected, the sampler can be shut down and the transmitter can continue to pulse. When the rotor has returned to the starting position, the sampler can then be restarted and the collection of 3042 data points repeated.

In every case, the elapsed time between the first data position in the return and each successive data position in the return is the same.

Therefore, if the first data point is collected with the rotor in a fixed position and the rotor is spinning at constant speed, all 507 data points will always be collected with the rotor blade in one of 507 different but constant positions.

The sensor for this application need not continuously monitor blade position, it must simply detect the instant the blade passes a particular reference position. Many noncontacting devices including eddy current and photoelectric sensors can accomplish this task.

Reference is now made to Figures 8 and 9 for a description of one preferred circuit arrangement for measuring the thickness of an ice layer.

Referring firstly to Figure 8, clock pulses generated by a pulse repetition frequency cloak 12 are supplied by a delay generator 13 to a transmitting circuit 14 which causes radar pulses to be radiated from transmitting antenna 2. Reflected radar pulses are converted by receiving antenna 3 into corresponding electrical signals which are amplified and sampled, at time intervals determined by delay generator 13, in amplifier 15 and sampler 16 of receiving circuit 17. The sampled detector output is passed to a buffer circuit 18, and the buffer output is averaged, in averager 19, over a given number (e.g. six) of sampled detector outputs at the same time interval between pulse transmission and reflected pulse sampling. Once six pulses have been averaged, the average value is converted to a digital signal in analogue-to-digital converter 20.

A control processor 22 serially processes the digital output signals from analogue-to-digital converter 20 in a manner which will be described in some detail hereinbelow. Additionally, it generates a delay request digital signal which indicates that it is ready to receive the next digital sampled detector output signal from analogue-to-digital converter 20. The delay request digital signals are passed to delay register 23 which provides an output digital signal indicating the duration of the next time interval for which the detector output is to be sampled. This digital signal is then converted into an analogue signal in digital-to-analogue converter 24 and supplied as a control input to delay generator 13. The timing of the initiation of the radar pulse transmission and/or return sampling is controlled, via control processor, by position sensor 21.The circuit elements 12, 13, 18, 19, 20, 23 and 24 together constitute a timing and interface circuit 25.

With reference to Figure 9, the control processor 22 includes 507 digital high-pass filters HPF1, HPF2 ... HPF507. Each filter is associated with a respective one of the digital samples making up the return trace shown in Figure 6 and has its band-pass frequency set to pass the spikes (R1, R2) corresponding with the reflected radar pulses from the upper and lower surfaces of the ice layer but to eliminate or reduce the magnitude of the clutter component at the corresponding angular position of the helicopter rotor.The digital samples from the analogue-to-digital converter 20 are routed, by control processor 22, sequentially into the respective high pass filter according to the appropriate time intervals between pulse transmission and detector output sampling, and the filtered digital signals are temporarily stored in respective memory positions 1 to 507 in memory 25 of display interface 26. This display interface also includes a shift register 27 having 507 inputs respectively connected to the 507 outputs of memory 25. The shift register sequentially inputs the 507 digital signals stored in the memory 25 to digital-to-analogue converter 28 whose output is connected to a visual display unit 29 or oscilloscope or the like, which provides a visual indication or recording of the processed sampled detector outputs.From this data, the time interval between detection of the reflected radar pulses from the upper and lower surfaces of the ice layer can be determined, and hence the thickness of the ice layer.

The foregoing description with reference to the drawings has been given in relation to an airborne radar system for measuring the thickness of an ice layer in a polar region. It will be appreciated, however, that it equally finds application to measuring the thickness of any layer or body of material exhibiting dielectric interfaces at its two principal faces. It will also be obvious to the skilled reader how to adapt the radar system described for detecting the presence of a dielectric interface. In that case, a single spike in the high-pass filtered radar return indicates the presence of the interface.

It is further remarked that it is not essential that the synchronisation between the generation of the radar returns and the rotation of the helicopter main rotor be exact. Since the radar pulsing and detector output sampling rates are very much greater than the rate of rotation of the main rotor, the rotor angular position will be only minimally different if the generation of the radar returns is slightly advanced or retarded. Therefore, the radar clutter contribution at any rotor angle will be substantially invariant and it is sufficient if the synchronisation is only approximate.

Claims (20)

CLAIMS:

1. A method of electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagnetic clutter, comprising the steps of: (a) transmitting electromagnetic pulses from the aircraft towards said region; (b' detecting in the aircraft reflected electromagnetic pulses from the direction of said region, by means of a detector, said detector also detecting electromagnetic rotor clutter; (c" generating radar returns of which the instantaneous magnitude represents the time varying strength of the detector output at a particular delay following the transmission of an electromagnetic pulse from the aircraft;; (d synchronising the generation of the radar returns with the rotation of the rotor such that the instantaneous magnitude of the radar returns at each particular time point during the duration of the radar returns corresponds with substan tially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns is substantially invariant from one radar return to another; (e) at least partially separating the substantially invariant electromagnetic rotor clutter contribution from the radar returns to provide relatively clutter-free returns; and (f) detecting in at least one of the relatively clutter-free radar returns the presence of an interface reflection spike resulting from the presence of said dielectric interface in said region.

2. A method as claimed in claim 1, wherein each radar return is generated by repeatedly sampling the detector output at different time delays after the transmission of a respective corresponding number of successive radar pulses, the respective such samples combining to form said radar return, and wherein each sample corresponding with a respective said time delay is subjected to respective high pass filtering, so as to at least partially separate the invariant electromagnetic clutter contribution.

3. A method of electromagnetically measuring the thickness of an ice layer in a polar region from an aircraft having a rotor producing electromagnetic clutter, comprising the steps of: (a) transmitting electromagnetic pulses from the aircraft towards the ice layer; (b) detecting in the aircraft reflected electromagnetic pulses from the upper and lower surfaces of the ice layer by means of a detector, said detector also detecting electromagnetic rotor clutter; (c) sampling the detector output at a first given time delay after the transmission of an electromagnetic pulse and at a particular angular position of the rotor;; (d repeating step (c) a number of times where, each time, the sampling is effected at a respective, different, time delay after the transmission of an electromagnetic pulse, so that the samples from steps (c) and (d) cover a range of sequentially increasing time delays, said samples from steps (c) and (d) combining to form a radar return of which the instantaneous magnitude represents the time-varying strength of the detector output at a particular delay following the transmission of an electromagnetic pulse from the aircraft;; (e) repeating steps (c) and (d) to provide further radar returns such that the instantaneous magnitude of the radar returns at each particular time point during the duration of the radar returns always corresponds with substantially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns is substantially invariant from one radar return to another; (f) at least partially separating the substantially invariant electromagnetic rotor clutter contribution from the radar returns to provide relatively clutter-free returns;; (g) monitoring the relatively clutter-free returns for the presence of interface reflection spikes indicative of the detection of reflected electromagnetic pulses from said upper and lower surfaces, respectively, of the ice layer; and (h) determining the ice layer thickness from the time difference between detection of the spikes.

4. A method as claimed in claim 3, wherein the transmitting in step (a) is carried out by always pulsing a pulse transmitter in the aircraft at the same particular angular position of the rotor.

5. A method as claimed in claim 3, wherein the sampling in step (d) is carried out, each time, always at the same particular angular position of the rotor as in step (c).

6. A method as claimed in claim 3, wherein the transmitting in step (a) is carried out by initiating a series of pulses at a preselected angular position of the rotor.

7. A method as claimed in claim 3, wherein the sampling in step (d) is carried out, each time, at respective, different, angular positions of the rotor which are different from the rotor angular position in step (c).

8. A method as claimed in claim 6, wherein each individual sampling of the detector output in step (d) is effected at a respective, different, time delay after transmission of each respective pulse in said series, which time interval increases by a preselected time increment from one individual sampling to the next, the several such samples being combined to forum a single radar return.

9. A method as claimed in any one of claims 3 to 8, comprising the further step of: li) determining the average value of a predetermined number of successive samples corresponding with each different time delay, these average values for the respective time delays combining to form a radar return.

10. A method as claimed in claims 5, 6 or 8, or in claim 9 as appended to claim 5, 6 or 8, wherein a radar return is generated once only during the rotor position repeat time or an integer multiple thereof.

11. A method as claimed in any one of claims 1 to 9, wherein the radar returns are generated respectively over successive integer sub-"jltiples of the rotor position repeat time and step (g) is carried out by respective high-pass filtering of the radar returns.

12. A method as claimed in any one of claims 1 to 10, wherein the substantially invariant electromagnetic clutter- is at least partially separated from the radar returns by subjecting the radar returns to high pass filtering.

13. Apparatus for use in electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagneti clutter, comprising - means arranged to transmit electromagnetic pulses from the aircraft towards said region and to detect in the aircraft reflected electromagnetic pulses from the direction of said region;; - means arranged to generate radar returns of which the instantaneous magnitude represents the detected electro magnetic pulse strength at a particular delay following the transmission of an electromagnetic pulse from the aircraft, said radar returns generating means including synchronising means arranged to synchronise the generation of the radar returns with the rotation of the rotor such that the instantaneous magnitude of the radar returns at each particular time point during the duration of the radar returns always corresponds with substantially the same rotor angular position for that time point, whereby the temporal distribution of the instantaneous electromagnetic rotor clutter contribution across the radar returns, resulting from detection of electromagnetic ' rotor clutter by the transmitting./detecting means, is substantially invariant from one radar return to another; - means arranged at least partially to separate the substantially invariant electromagnetic rotor clutter contribution from the radar returns to provide clutter-free returns; and - means arranged to detect in at least one of the relatively clutter-free radar returns the presence of an interface reflection spike resulting from the presence C said dielectric interface in said region.

14. Apparatus as claimed in claim 13, wherein tme means arranged at least partially to separate the invariant electrorragnetic clutter contribution from the radar returns comprises high pass filtering means.

15. Apparatus as claimed in claim 13 or 14, wherein the synchronising means includes a rotor angular position detector associated with said rotor.

16. Apparatus as claimed in any one of claims 13 to 15, wherein the signal processing means comprises a clock pulse generator arranged to cause the transmitting/detecting means to transmit said electromagnetic pulses towards said region, a sampler, a delay generator responsive to the clock pulse generator output and arranged to cause the sampler to sample the detected electromagnetic pulse strength repeatedly at different time delays after the transmission of a respective corresponding number of radar pulses, the respective such samples combining to form said radar returns, and a plurality of high pass filters, each arranged to subject a said sample corresponding with a respective said given time delay to high pass filtering, so as at least partially to separate the invariant electromagnetic rotor clutter contribution.

17. Apparatus as claimed in claim 16, wherein the delay generator is arranged to cause the sampler to sample the detected electromagnetic pulse strength a predetermined number of times at each different given time delay, and wherein the signal processing means further comprises an averager for averaging the plurality of detected electromagnetic pulse strengths at each different given time delay, these average values combining to form a radar return.

18. Apparatus as claimed in claim 16 or 17, further comprising a unit for storing or displaying the relatively clutter-free radar returns, a memory arranged to store the outputs from said high pass filters, and a shift register arranged to read the stored data in said memory serially into said storage or display unit.

19. A method of electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagnetic clutter, substantially as hereinbefore described with reference to Figures 6 to 9 of the accompanying drawings.

20. Apparatus for electromagnetically detecting a dielectric interface in a region below an aircraft having a rotor producing electromagnetic clutter, substantially as hereinbefore described with reference to Figures 6 to 9 of the accompanying drawings.