The present invention relates to sensing apparatus.

Examples of the present invention provide apparatus comprising: an elongate electrical conductor having an electrical characteristic which measurably changes in the event of a mechanical disturbance at a position along its length; a signal generator operable to inject a frequency modulated continuous wave signal into the conductor at an injection position; a detector operable to detect reflected signals received at the injection position and to use the injected signal and the reflected signal to identify the occurrence of a disturbance.

The detector may comprise a demodulator which, in use, demodulates the reflected signal against the injected signal. Alternatively, or in addition, the detector may comprise a demodulator which, in use, demodulates the reflected signal against a quadrature signal which is the injected signal shifted in phase by one quarter cycle. The detector may analyse the demodulated reflected signal and/or the demodulated quadrature signal to identify its frequency components. The detector may be operable to identify the position of the disturbance, based on the frequency components identified. The detector may comprise one or more frequency-dependent filters, such as bandpass filters, operable to identify frequency components in the or each demodulated signal. A Fourier Transform analysis may be performed on the or each demodulated signal and/or on the orthogonal combination of the demodulated signals. The detector may be operable to identify changes in one or more Fourier Transform coefficients as indicative of a disturbance. The detector may maintain a running average of the or each Fourier Transform coefficient for identifying changes from the average.

The or each demodulated signal may be provided for analysis as a sample train. The detector may define a window to identify a group of samples for analysis, and be operable to move the window along the sample train to identify further groups for analysis. The detector may synchronise the windows with the sweeps of the frequency modulated continuous wave.

The detector may repeatedly apply a Fourier Transform to samples of the sample train to derive a plurality of sets of Fourier Coefficients, each set being representative of a respective sample, and in which the corresponding Fourier Coefficients from each set are stored as a time sequence of corresponding Fourier Coefficients, the time sequences being used as respective encoded audio signals. The signal generator may be operable to inject a plurality of signals into the conductor, a sample train being provided for reflections of each of the injected signals, and a Fourier Transform being applied to the samples of the sample trains to derive sets of Fourier Coefficients, the Fourier Coefficients from each set being stored in time sequences containing interleaved coefficients from sets derived from each of the plurality of injected signals, the time sequences of interleaved coefficients being used as encoded audio signals.

The signal generator may provide the injected signal directly for use by the detector. The signal generator may inject the frequency modulated continuous wave by means of a directional coupler which directs reflected signals to the detector.

The signal generator may be operable to inject a plurality of signals into the conductor, each signal being a frequency modulated continuous wave signal. The detector may be operable to detect reflections of each of the injected signals for use in identifying the occurrence of a disturbance. In another aspect, examples of the present invention provide a method of identifying a mechanical disturbance in an elongate electrical conductor having an electrical characteristic which measurably changes in the event of a mechanical disturbance at a position along its length, the method comprising: injecting a frequency modulated continuous wave signal into the conductor at an injection point; detecting reflected signals received at the injection position and using the injected signal and the reflected signal to identify the occurrence of a disturbance.

Examples of the present invention will now be described in more detail, by way of example only, and with reference to the accompanying drawings, in which:

Fig. 1 is a highly simplified diagrammatic overview of an example apparatus;

Fig. 2 is a cross-section through the conductor of Fig. 1 ;

Fig. 3 is a block diagram of a first example apparatus;

Figs. 4, 5 and 6 are plots of frequency against time for signals occurring within the apparatus;

Fig. 7 is a block diagram of a second example apparatus;

Fig. 8 is a simple spectrum of Fourier coefficients arising from the analysis performed by the apparatus;

Fig. 9 is a block diagram of part of a third example apparatus; Figs. 10, 11 and 12 are plots of frequency against time for signals occurring within the apparatus, when a fourth example is implemented; and Fig. 13 is a block diagram of part of a fourth example apparatus.

Overview

Fig. 1 illustrates apparatus 10 which comprises an elongate electrical conductor 12, a signal generator 14 and a detector 16.

The conductor 12 has an electrical characteristic which measurably changes in the event of a mechanical disturbance such as is indicated at 18, at a position along its length. The conductor may have a length of 1 km or more. The conductor 2 may be incorporated within a structure such as a fence or attached to another fixed structure such as a wall, or may be buried in the ground. This allows the conductor 12 to be used in a manner which will be described, to provide a form of perimeter protection in order to identify a disturbance which may represent intrusion, tampering or (in the case of a penal environment) an escape attempt.

The signal generator 14 injects a signal into the conductor 12 at an injection position 19. Reflected signals are received at the injection position 19. The detector 16 uses the injected signal and the reflected signal to identify the occurrence of a disturbance 18.

Electrical conductor

The conductor 12 has an electrical characteristic which measurably changes in the event of a mechanical disturbance, as noted above. For example, the conductor 12 may include a wire 20 (Fig. 2) which extends along the conductor 12 alongside a substantially continuous body 22, the conductor 12 being free to move relative to the body 22 in the event of a mechanical disturbance such as vibration or impact. The body 22 may be a magnet or magnetic material so that movement of the wire 20 creates magnetic effects which locally change the electrical characteristic of the wire 20. In another example, the body 22 is a piezoelectric body against which the wire 20 is in contact so that movement of the wire 20 creates voltages which locally change the electrical characteristics of the wire 20. In another example, the body 22 is a tribo-electric body across which the wire 20 can rub as it moves to create capacitative effects which locally change the electrical characteristics of the wire 20.

When an electrical signal is propagated along the conductor 12 by means of the wire 20, the change of electrical characteristic caused by the mechanical disturbance 18 will result in a proportion of the propagating signal being reflected from the disturbance 18. In a practical application, it is likely that appropriate termination would be provided at the far end of the conductor 12, to prevent spurious reflections arising. Passive or active termination devices or techniques could be used.

Signal generator and detector (first example)

The signal generator 14 and detector 16 are illustrated in more detail in Fig. 3, in a first example. In this example, the signal generator 14 provides a signal to the injection position 19, through an amplifier 24. A directional low loss bridge 26 is provided at the injection position 19 to inject the output of the signal generator 14, amplified by the amplifier 24, into the conductor 12. Signals injected in this way into the conductor 12 will propagate away from the injection position 19, along the conductor 12. Reflected signals will propagate back towards the injection position 19, where the bridge 26 directs them through an amplifier 28 to a demodulator circuit 30. The demodulator 30 also receives the output of the signal generator 14, at 32, allowing the demodulator 30 to demodulate the reflected signals against the injected signals, to provide a demodulated signal for analysis by an analyser 34. The significance of this can now be explained with reference to Figs 4, 5 and 6.

Fig. 4 shows the frequency against time for a signal 36 of the type generated by the signal generator 14 for injection into the conductor 12 at the injection position 19. In this example, the injected signal 36 is a frequency modulated continuous wave (FMCW) signal. That is, the signal 36 is a continuous wave sinusoid having a frequency which is modulated as a repeating cycle of linearly increasing frequency, up to a peak frequency 38, after which the cycle repeats. When the signal 36 is injected into the conductor 12, it will be partially reflected from a disturbance 18, as noted above, because of the change of electrical properties of the conductor 12. Fig. 5 illustrates the reflected signal 36' in solid lines, and the injected signal 36 in broken lines, both signals 36, 36' being shown on the same time axis. In the interests of simplicity of this explanation, Fig. 5 assumes that reflection takes place from only one disturbance position 18, and that no other reflections or signal degradations take place. Other situations are described further below. As can clearly be seen in Fig. 5, the reflected signal 36' is delayed relative to the injected signal 36. This delay is created by propagation delay required for the signal 36 to propagate out to the disturbance 18, and for the reflected signal 36' to propagate back to the injection position 19. At any point in time, there is a frequency difference between the two signals 36, 36'. This frequency difference is a function of the slope of the signal 36 (the rate of increase of frequency), the speed at which the signal propagates along the conductor 12, and of the distance between the injection position 19 and the disturbance 18. Thus, the distance between the injection position 19 and the disturbance 18 represents the unknown variable from which the frequency difference arises.

The demodulator 30 demodulates the reflected signal 36' against the injected signal 36, as noted above. Accordingly, this demodulation recovers a frequency which is the frequency difference between the signals 36, 36', at any point in time. The demodulated signal is illustrated in Fig. 6, as signal 40. In this simple example, the demodulated signal 40 is a constant value at 42, except at periodic peaks 44 arising between the peaks 38 of the signals 36, 36'.

The analyser 34 receives the demodulated signal 40 from the demodulator 30. As noted above, the frequency 42 of the demodulated signal 40 is a function of the distance from the injection position 19 to the disturbance 18. In the absence of any disturbance (and therefore in the absence of any reflected signal in this simple example), the frequency of the signal 40 will change as the frequency of the injected signal changes, and will change to a constant value 42 in the event that a disturbance arises to create a reflected signal. Accordingly, the analyser 34 is able to determine the occurrence of the disturbance by identifying the existence of a constant value demodulated signal, and is also able to determine the distance from the injection position 19 to the disturbance 18 by considering the frequency of the demodulated signal 40, if the slope of the signal 36 is known to the analyser 34.

Signal generator and detector (second example)

Fig. 7 illustrates a second example which has many similarities with the first example described above. Consequently, the same reference numerals are used for corresponding components, which are not fully described again, in the interests of brevity. The components of Fig. 7 which correspond with those of Fig. 3 are delineated by a broken line in Fig. 7. Additional components are a trigger circuit 46, a second signal generator 48 and a second demodulator 50. The reflected signal 36' is provided to both demodulators 30, 50, The analyser 34 receives the demodulated signals from both demodulators 30, 50. The trigger circuit 46 triggers the start of each cycle of the injected signal 36, produced by the signal generator 14. The trigger circuit 46 also triggers the second signal generator 48 to produce a signal which is at the same instantaneous frequency as the output of the generator 14, but with a quadrature (90°) phase shift. Accordingly, the demodulator 30 demodulates the reflected signal 36' against the injected signal, whereas the demodulator 50 demodulates the reflected signal 36' against the quadrature shifted form of the injected signal 36.

In the simple example of a reflected signal being a reflection from a single disturbance 18, with no other reflections or degradations, either of these demodulated signals will be sufficient for the analyser 34 to determine the position of the disturbance 18, in the manner described above in relation to the first example. However, in a practical situation, many reflections are likely to occur for many reasons (apart from the disturbances being detected), so that there is likely to be a discontinuous spectrum of frequencies generated in the outputs of the demodulators 30, 50. Various techniques can be used to identify the reflected frequencies of interest (representing disturbances) from others. Several examples will now be described. Other examples can be envisaged. In one method, the analyser 34 performs a Fourier Transform on the output of one of the demodulators 30, 50. This recovers Fourier Coefficients for the component frequencies of the demodulator output. In stable conditions, with no disturbances and with all other factors remaining consistent, this spectrum of component frequencies will remain substantially constant in frequency and amplitude. In a practical example, the Fourier Coefficients may differ over time, for example due to the lie of the conductor 12, or the elements within the conductor 12, after a disturbance, or some drift may occur, for example as a result of thermal effects within the system. Accordingly, the analyser 34 keeps a running average of the quiescent magnitude of each of the Fourier Coefficients. A highly schematic and simplified example is illustrated in Fig. 8 (a). In the event of a disturbance arising, the demodulated signals will include components at a frequency determined by the position of the disturbance 18, as noted above. Accordingly, there will be a change in the amplitude of the corresponding Fourier Coefficient or alternatively, a new Fourier Coefficient will appear in the spectrum of Fig. 8 (a), as illustrated in Fig. 8 (b) at 52. This change allows the analyser 34 to identify that a disturbance has occurred. This change also allows the analyser 34 to identify the position of the disturbance by reference to the frequency corresponding with the Fourier Coefficients which the analyser 34 has identified as changing. The analysis may include the use of scaling factors applied to the difference between the instantaneous value of the coefficients, and the corresponding quiescent average value recorded by the analyser 34. Appropriate choice and adjustment of scaling factors for the various coefficients allows the analysis to take into account that the conductor 12 may be non-linear in its frequency response, and that the response of the conductor 12 to a physical stimulus may not be consistent along the length of the conductor 12.

In the above method, the amplitude of the demodulated signal is dependent on the phase relationship between the injected and the reflected signals. One special case will arise if the signal arising from the reflection is exactly in quadrature phase at the injection point with the signal being used for demodulation (whether the injected signal or the quadrature signal). In that situation, no output would be recovered from the corresponding demodulator 30, 50. Thus, in a second method, the analyser 34 treats both demodulator outputs in the manner just described, to ensure that the analysis can be completed even if the reflection is in quadrature phase with one or other demodulation signal. In a third method, the demodulated signals may be combined orthogonally to provide a representation of the reflected power. The Fourier Transform is then performed on this orthogonally combined signal for analysis of the Fourier Coefficients as described above.

In these methods, the demodulated signal (or signals) may be digitised into a train of samples, and windowing may be used (before or after digitisation) to split the samples into a series of windows before applying a Fourier Transform to the samples within each window. Windows may be synchronised with the start of each cycle of the injected signal, for example. Many other regimes for choosing and using windows can be devised. A windowing function may be applied to the window for reasons which are known in themselves in relation to Fourier analysis, such as to address problems of spectral leakage. In a fourth method, the analyser 34 may include a set of band-pass filters with respective passbands, centred at various frequencies of interest. The outputs of the demodulators 34, 50 are applied to each of these filters. Disturbances on the conductor 12 will result in variations in the amplitude of the output of these bandpass filters, allowing the occurrence of a disturbance to be identified.

Signal generator and detector (third example)

The arrangements described above in relation to the second example can be further extended to form a third example, here described with reference to Fig. 9. This example uses the arrangements of the second example, illustrated in Fig. 7 to provide at 54 the outputs of the demodulators 30, 50. That is, the inputs 54 to the arrangements of Fig. 9 are the demodulated output from the demodulator 30, and the quadrature demodulated output from the demodulator 50. These are orthogonally combined and digitised to create the sample stream indicated at 54. The sample stream at 54 is applied to a shift register 56, which periodically shifts its contents to receive further samples from the stream 54. Accordingly, the shift register 56 contains a time sequence of samples from the sample stream 54. The contents of the shift register 56 form the inputs to a Fourier transform function illustrated at 58, which may be implemented as dedicated hardware or as appropriate software or firmware, running on general-purpose hardware. The contents of the shift register 56, or part of them, form a window through which the Fourier transform 58 samples the sample stream 54. On each occasion that a Fourier Transform is performed by the function 58, a set of Fourier Coefficients 60 is produced, here illustrated as a column of rectangles. Each position in the column corresponds with a different Fourier coefficient, and therefore with a different frequency within the reflected signal 36'. Each position in the column also has an associated shift register 62. On each occasion that a new Fourier Transform is to be performed, the contents of the shift registers 62 are shifted and the previous Fourier coefficients at 60 are moved into the shift registers 62.

Accordingly, as the analysis continues, each of the shift registers 62 contains a time sequence of the corresponding Fourier Coefficients relating to a particular frequency in the reflected signal 36'. Each frequency recovered by a Fourier Transform corresponds with a particular position along the conductor 12, as noted above. Accordingly, each shift register 62 contains a history of disturbances (if any) at the corresponding position along the conductor 12. Each other shift register 62 contains a history of disturbances (if any) at another corresponding position along the conductor 12. The separation of these corresponding positions represents the resolution available from this analysis. Thus, the contents of each shift register 62 can be considered as a separate channel of output of the apparatus, each channel corresponding with a known length of the conductor 12 at a particular position along the conductor 12. The contents of the shift registers 62 can be subjected to an averaging of the quiescent magnitude, or to filtering or other techniques, in order to exclude spurious results and identify disturbances occurring in the corresponding length of conductor 12.

The contents of the shift registers 62 can be considered and dealt with as a form of digitised audio signal with a sample rate equal to the rate at which Fourier Transforms are implemented by the function 58. This audio signal can be subsequently processed in order to produce an output such as an alarm signal, indicating the occurrence of a disturbance, and also identifying the location of the disturbance. The audio signals may also be forwarded for audio surveillance purposes or for storage, allowing the signals to be accessed retrospectively, for reference or analysis. Signal generator and detector (fourth example)

The arrangements described above in relation to the third example can be further extended to form a fourth example. In this fourth example, multiple FMCW signals are injected into the cable. This can be done by generating each signal separately, using a plurality of the arrangements of the second example, illustrated in figure 7. Each of these arrangements generates one of the multiple FMCW signals. The multiple FMCW signals may then be injected independently into the conductor 12, or may be mixed prior to injection. The injected FMCW signals may each vary over the same range of frequencies, but other possibilities could be used.

For simplicity, and without suggesting any lack of generality, the following example uses two concurrent FMCW signals which are identical except that the start of the frequency sweeps are interleaved and distributed equally over time. This is illustrated in Fig. 10 which shows two FMCW signals 36 overlaid on each other. As each signal 36 reaches its peak value 38, the other signal has risen halfway up its frequency slope.

The two signals will each produce a respective reflected signal in the event of a disturbance arising on the conductor 12. This is illustrated in Fig. 1 1. Each injected signal 36 (shown in broken lines) is followed by a respective reflected signal 36' (shown in solid lines).

At any particular time, there exists a frequency difference between each of the injected signals 36 and the corresponding reflected signal 36', and also between the two reflected signals 36'. Thus, when the signal received back at the injection position 19 is demodulated against a single injected frequency (which can be either one of the injected signals), two beat frequencies 64 are produced. This is illustrated in Fig. 2. The two beat frequencies 64 change over time, as the relationship between the frequencies of the two injected signals changes.

The horizontal dotted line represents an upper threshold frequency 66 which can be applied to the beat frequencies 64, to represent a maximum distance from the injection position 19. Beat frequencies above the threshold frequency 66 are ignored in the analysis. Appropriate selection of the frequencies for the injected signals 36, and of the threshold frequency 66, allows the maximum distance to be chosen to correspond with the end of the conductor 12. The upper threshold frequency 66 can be imposed by a low pass filter, for example. A similar frequency/time plot is produced by demodulating against the other of the injected F CW signals.

The two demodulated signals produced by demodulating against each of the injected FMCW signals, are further processed in a manner similar to that described above in relation to Fig. 9. This is illustrated in a further simplified manner in Fig. 13. In view of the similarities with Fig. 9, the same reference numerals are used again, with an apostrophe suffixed. On each occasion that a Fourier Transform is performed by one of the functions 58', a set of Fourier Coefficients 60' is produced, here illustrated as columns of rectangles. In each column, each position in the column corresponds with a different Fourier coefficient, and therefore with a different frequency within the reflected signal 36'. On each occasion that a new Fourier Transform is to be performed by either of the functions 58', the contents of the shift registers 62' are shifted and the previous Fourier coefficients at 60' are moved into the shift registers 62'. Each of the shift registers 62' receives an alternating or interleaved sequence of coefficients from the two sets 60'. Thus, each position in each column 60' has an associated position in the other column 60', and these two associated positions have an associated shift register 62'.

Accordingly, as the analysis continues, each of the shift registers 62' contains a time sequence of the corresponding Fourier Coefficients relating to a particular frequency in the reflected signal 36'. Each frequency recovered by a Fourier Transform corresponds with a particular position along the conductor 12, as noted above. Accordingly, each shift register 62' contains a history of disturbances (if any) at the corresponding position along the conductor 12. Each other shift register 62' contains a history of disturbances (if any) at another corresponding position along the conductor 12. The separation of these corresponding positions represents the resolution available from this analysis. Thus, the contents of each shift register 62' can be considered as a separate channel of output of the apparatus, each channel corresponding with a known length of the conductor 12.

Furthermore, the interleaving of Fourier coefficients from the two functions 58' has the effect that each of the shift registers 62' contains a form of digitised audio signal with a sample rate which is higher than the rate at which Fourier Transforms are implemented by either of the functions 58'. In this example, the effective sample rate is double the rate achieved in the third example described above. The audio signals can be further utilised in the manner described above in relation to the third example. Concluding remarks

Many variations modifications can be made specific examples described above, without departing from the scope of the invention. Many different technologies can be used to implement the constituents of the apparatus described. In those examples which use Fourier Transforms, analysis may be based solely on the magnitude of the various Fourier Coefficients and these may therefore be normalised around zero by the analyser. Alternatively, the analyser may keep a running average, so that the normalised magnitude may be the difference between an instantaneous Fourier Coefficient magnitude, and the average. The average may be calculated as a mean, mode or in other ways. As an alternative to normalising the Fourier Coefficients, or in addition to doing so, the outputs of the shift registers 62 in the third and fourth may be subjected to a high pass filter in order to remove any DC components. The use of multiple injected signals has been described in relation to the fourth example. Multiple injected signals could also be used in the other examples. This may be done to provide redundancy in the outputs, for improved reliability. Alternatively, different frequency profiles could be used for the various injected signals, to provide different resolution outputs for various reasons.

Whilst endeavouring in the foregoing specification to draw attention to those features of the invention believed to be of particular importance it should be understood that the Applicant claims protection in respect of any patentable feature or combination of features hereinbefore referred to and/or shown in the drawings whether or not particular emphasis has been placed thereon.