GB2344021A - Passive transponder tagging system - Google Patents

Abstract

An electronic tagging system suitable for tagging persons or articles, comprises a plurality of tags, each for association with a person or an article to be tagged, a base station having a transmitter, for illuminating the tags with r.f. energy, and a receiver tuned to receive r.f. energy re-radiated from the tags, wherein the tags each comprise an antenna, which serves to receive r.f. energy radiated by the transmitter, a rectifier, which produces d.c. power from the received r.f. energy, a pulse generator which is driven by the d.c. power and synchronised with changes in the received r.f. energy to produce a 'direct sequence spread spectrum' (DSSS) signal providing a unique code for each tag and wherein the base station receiver includes a decoder which is responsive to the re-radiated coded DSSS signals from the tags for tag detection and identification purposes.

Description

Improvements in or relating to Tagging Systems This invention relates to electronic tagging systems and more especially but not exclusively it relates to a system for tagging passengers in an airport environment.

Tagging systems are known which use'passive transponder' technology and which comprise a base station having a transmitter for illuminating tags with r. f. energy at a suitable frequency, to provide power for the tags which is used to re-radiate distinctive reflections, and a receiver (not necessarily co-located with the transmitter) tuned to receive the reflections so as to identify individual tags. The tags each normally comprise an antenna, a rectifier, which serves to produce d. c. power from the received r. f. energy, an r. f. switch operative to change the aperture of the antenna so as to provide the distinctive reflections and a pulsing circuit which controls operation of the r. f. switch and which is driven by the d. c. power.

In these known systems the pulsing circuit might normally comprise a simple crystal oscillator of the kind used in many watches.

Although this may be suitable for a single tag, it becomes problematic for use in systems requiring a large number of distinctive tags. This is because it requires the manufacture of a large number of special crystal oscillators each having slightly different resonant frequencies, and given the frequency inaccuracy of the oscillators, it would not be possible to generate enough unique frequencies. Moreover, the cost of a resonator and its bulkiness do not make it an attractive proposition.

It is an object of the present invention to provide a tagging system in which the aforementioned disadvantages are largely obviated.

According to the present invention as broadly conceived a tagging system comprises a plurality of tags, each for association with a person or an article to be tagged, a base station having a transmitter, for illuminating the tags with r. f. energy, and a receiver tuned to receive r. f. energy re-radiated from the tags, wherein the tags each comprise an antenna, which serves to receive r. f. energy radiated by the transmitter, a rectifier, which produces d. c. power from the received r. f. energy, a pulse generator which is driven by the d. c. power and synchronised with the received r. f. energy to produce a'direct sequence spread spectrum' (DSSS) signal providing a unique code for each tag and wherein the base station receiver includes a decoder which is responsive to the re-radiated coded DSSS signals from the tags for tag detection and identification purposes.

By using direct sequence spread spectrum (DSSS) a number of unique codes can be radiated and detected efficiently using digital processing, the technique providing discrimination against various forms of interference and enhancement of signal to noise performance.

Since the coded signals are synchronised, they may be 'orthogonally'or'near orthogonally'coded which affords the advantage that a set of codes having zero mutual interference may be chosen.

In the case of near orthogonal codes, which as the name suggests are almost completely orthogonal, each of the tags is arranged to transmit a different near orthogonal code, produced for example from a set of codes defined by rotations of a common Msequence. This provides a suppression by a power factor equal to the length of the code squared, the provision of uniquely defined codes being facilitated because transmissions are synchronised.

Alternatively, the unique codes may comprise a rotation of a common M-sequence with a trailing bit in such a way as to provide full orthogonality.

As a further alternative, the unique codes may be Gold codes taken from a set of near orthogonal Gold codes, or from a set of orthogonal extended Gold codes.

The unique codes may be Walsh codes which may be covered modulus 2 by an extended M sequence.

It will be appreciated that the use of pseudo random codes, which to some degree interfere with one another, rather than orthogonal or near orthogonal codes, would not be practicable since the suppression of interference from one code to another is related to a power factor approximately equal to the length of code, and in a situation of widely differing signal strengths this suppression may not be adequate to allow the weaker of two (or more) signals to be detected. This effect is known as the'near-far effect', which in mobile radio systems using DSSS, is overcome by means of automatic power control. This is clearly not possible with tags and so it is essential to use orthogonal or near orthogonal codes and as herein before mentioned, this in turn imposes a requirement for synchronisation of the tags'code generators.

One way of facilitating synchronisation is to generate a synchronising pulse in the illuminating transmitter. This will reset the codes in each of the tags so that they all start from the same point in their sequence. The code generators would then each run asynchronously under control of their own crystal oscillators and in this case all crystal oscillators would operate on the same frequency.

However, it would be desirable to eliminate the use of crystal oscillators completely and this could be done by providing a clocking signal in the r. f. transmission itself.

Some embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which; Figure 1, is a schematic block diagram of a plurality of tags operatively associated with a base station; Figure 2a, is a schematic block diagram of a transmitter which forms a part the base station as shown in Figure 1; Figure 2b, is a schematic block diagram of a pulsing circuit forming a part of the transmitter of Figure 2a; Figure 3, is a schematic block diagram of a receiver which forms a part of the base station shown in Figure 1; Figure 4, is a schematic block diagram of apparatus which forms a part of a tag, and; Figure 5 is a waveform diagram; Referring firstly to Figure 1, a tagging system comprises a plurality of tags A, B, and C, and a base station BS, wherein the tags are associated with items or persons to be tagged (not shown). The base station comprises an illumination transmitter TX which serves to illuminate the tags with r. f. energy used, thereby to provide power which facilitates the transmission of coded identification signals from the tags A, B, C, to the base station BS.

The illuminating transmitter TX, comprises an oscillator 1 arranged to feed a mixer 2 fed also from a pulsing circuit 3 so as to provide an output signal which is fed to a transmitter antenna 4 via an amplifier 5.

As shown in greater detail in Figure 2, the pulsing circuit 3, comprises a look-up table 6 fed with clock pulses from a clock pulse generator 7, output signals from which are fed to a digital to analogue converter, (DAC) 8 which provides an analogue output signal fed to the mixer 2 (as shown in Figure 1) via a low pass filter 9. The clock pulse generator 7 is arranged also to provide a synchronising signal on a line 10 (shown also in Figure 3), which is used in the base station receiver RX. In the present example the transmitter TX is co-located with the receiver RX and accordingly the synchronising signal is carried by means of hard wiring but alternatively, if the TX and RX were not colocated, then the synchronising signal could be transmitted thereto by means of suitably modulated electromagnetic radiation chosen in accordance with the application in view.

Referring now to Figure 3, the receiver RX comprises a receiver antenna 11, which feeds received signals to a band pass filter 12, tuned to the oscillator 1 of the illumination transmitter, as shown in Figure 1.

Signals received by the antenna 11 and passed by the filter 12 are fed to an envelope detector 13 which feeds a controlled integrator 14 via an amplifier 15. The integrator 14 is controlled by means of a timing generator 16 fed on the line 10 with the synchronising pulsed signals originating from the oscillator 1 so that coded signals from the tags are stepped through a shift register 15 the contents of which is sampled by a multiple correlator 16 primed with the codes corresponding to the codes of the tags A, B, C, whereby in the presence of correspondence between one of the primed codes and a received code sequence, an output signal is delivered on an output line 17.

Each of the tags A, B, C, includes a receiver as shown in Figure 4, which is powered by radiation received from the illumination transmitter TX at the base station, so as to re-radiate in an appropriate identification code. The receiver comprises an antenna 18 which is coupled to an r. f. tuned circuit, comprising an inductor 19 and a capacitor 20, which is tuned to be responsive to the frequency of the illumination transmitter TX. Received signal voltages developed across the tuned circuit are fed to a d. c. rectifier 21 having an associated reservoir capacitor 22 which serves to store energy used by a code generator 2 3. The code generator is fed with clock and reset signals via a diode 24 coupled to a capacitor 25 which is much smaller than the capacitor 21, the relative time constants being chosen so that clock signals are fed to the code generator 23 on a line 26, reset signals on a line 27, and d. c. power on a line 28. Output signals from the code generator 23 on a line 28 are fed via a resistor 29, a diode 30 and a capacitor 32 to the antenna 18 for re-radiation, a PIN diode 33 being connected so that in combination with the capacitor it shunts the tuned circuit 19,20, whereby the tuning of the components 19,20, is altered significantly without preventing reception of the d. c. power pulses. The operation of this circuit will be better understood from a consideration of the waveform diagram of Figure 5.

Referring now to Figure 5, consider an r. f. envelope waveform as shown, wherein an initial pulse 34 allows the capacitor 25 to charge up.

This also serves to reset the code generator 23. Since the time constants as aforesaid, are chosen so that signals on the line 26 track the envelope changes of the waveform as shown in Figure 5, a first rising edge 35 triggers a first stage of a code to be produced (e. g. a binary 1 involves an r. f. switch (not shown) in the code generator being'on', whereas a binary 2 involves an r. f switch being'off). During a following peak 36 of the envelope, the receiver effectively integrates and dumps over a limited period during which it can be assumed that all tags have been triggered (the actual time of triggering for a particular tag will be a function of that tag's proximity to the illuminating transmitter). This ensures that full synchronisation is achieved from the viewpoint of maintaining (near) orthogonality. It should be noted that a tag must able to detect a following rising edge irrespective of the state of its r. f. switch. This could perhaps be achieved by means of a separate feed for the detector which produced similar results irrespective of the state of the r. f. switch.

It is possible that the problem of activation of tags due to high power r. f. interfering pulses is insurmountable for some applications of the above proposed system. If this is the case, it will be necessary to revert to using watch crystal oscillators to run the codes. The codes will then need to be synchronised by having the illuminating transmitter send a short fast sequence of pulses to form a code which the tags can recognise and use for re-synchronisation.

As regards the choice of code used, a combination of CDMA and TDMA may be chosen. A TDMA element may be imposed by giving each tag an assigned time slot. After an initial pulse, each tag is arranged to count up pulses until it reached its time slot. Thus, suppose the code length was 1024. A tag which is assigned to operate in time slot 3 will count 3 x 1024 pulses before responding. Individual codes may be orthogonal or near orthogonal and several alternatives are possible.

For example Walsh codes may be used which are fully orthogonal whereby a set of length N provides N unique codes.

Because the balance of the codes is poor they would need to be covered by a common, more random code (e. g. an M sequence). The advantage of using Walsh codes is that the receiver could search for the entire set of codes very efficiently using the fast Hadamard transform. The disadvantage is that the tags would need to store the entire sequence. Thus, for a code of length 1024, one would need a store of 1024 bits length.

Alternatively as mentioned earlier, rotations of an M sequence which are near-orthogonal may be used which affords some limited potential for simplifying the complexity of the receiver. Correlation against all rotations of the code can be performed in the frequency domain using the discrete Fourier transform. Unfortunately the Cooley-Tookey fast Fourier transform (FFT) can only be used for frames of length 2Nwhereas an M sequence has length 2N-1. The Winograd transform is an implementation of the discrete Fourier transform which can achieve some reduction in complexity for sequences of lengths which are not prime. If the length of the code is 22N-1 it can be factored into (2N-1) (2N + 1). Thus 1023 = 31 x 33 = 31 x 3 x I l. The advantage of the M sequence for the tags is that it is very straightforward to implement. For a code of length 2 -1 all that is needed is two registers of length N, one to hold the start position (which uniquely determines the code), the other to hold the current position. Moreover, to set a new code into a tag requires the input of only N bits.

A further possibility is the use of augmented M sequences. As mentioned above, M sequences are only nearly orthogonal. Full orthogonality can be obtained by generating a set of codes based on rotations of an M sequence and then adding an extra zero at the end of the sequence for all rotations. This adds only modestly to the complexity of the tag. Although this now creates a sequence of length 2N (for which the Cooley Tookey FFT can be used), the codes no longer represent simple rotations of one another so multiple correlation via the discrete Fourier transform is no longer possible.

However, one could omit the last bit, apply the process described for M sequences based on the Winograd transform above and then add the correlation (the same for all) to the outputs of the process.

Thus M sequences could be augmented to provide full orthogonality at modest additional complexity in both the tags and the receivers. Whether this is necessary will depend on the differences in signal strength and the numbers of tags which appear in practice.

It is, therefore apparent that M sequences, with or without augmentation could be used. Let us now examine the number of codes available. Suppose we have a code length of 1024 and a TDMA frame length of 8. This provides 8 x 1024 unique codes. If the pulsing rate is 8.192 kHz this would allow all 8192 codes to be searched for in one second. This pulsing rate is a reasonable figure since it would allow the signal to sit comfortably inside a 25 kHz channel. The code length and number of TDMA slots may be altered freely at this design stage. The overall constraint is that M= X F where M is the number of unique codes, T is the overall polling period and F is the frequency of the pulsing.

In general we find that the number of unique codes is marginal.

By increasing the frequency and the polling period within practical limits we get up to a maximum of around 100,000 unique codes. This might be adequate for a single airport if all codes are assigned upon entry to the terminal building. However, it would not provide well for a large number of premium passengers whom we might wish to welcome upon arrival at the terminal. It would certainly not be suitable for a global/inter-airline system.

Upon re-examination of the coding concept we find that it is over designed in the sense that all 100,000 uniquely coded tags could turn up within range of a single receiver and still be unambiguously detected! By introducing a further layer on the coding and removing the above possibility we can generate an almost limitless number of unique codes and, at the same time, eliminate the need for the TDMA element.

The principle is to have each tag transmit several orthogonal or near orthogonal codes one after another, wherein each successive code is selected independently of the others. For example, we could have a basic sequence length of 256 and have each tag send 8 different codes.

This would give a total of 2558 = 1.788 x 10l9 unique codes. Of course we could not have all of them turn up at once! A pulsing waveform at 8.192 kHz could perform a detection in only ~ second.

The underlying compromise in this code approach is that several tags can, in principle, mimic another. For example, suppose a tag, A, used augmented M sequence codes: a, b, c, d, e, f, g, h. If there were eight other tags and one of them had'a'as its first code, another had'b'as its second code and so on, the presence of tag'A'would be mimicked. Of course, for the above design, the probability of this happening is infinitesimal. A combinatorial analysis will be needed to determine the optimum design. Note, however, that the orthogonal or near orthogonal codes themselves must not be made too short since they provide processing gain which protects against interference external to the systems as well operating CDMA resolution of tags. Intuitively a minimum length of about 256 seems appropriate.

Once the number of available tag codes becomes very large, the philosophy of tag operation changes. The original idea was to have reprogrammable tags which would be assigned with new codes at the airport terminal to avoid conflict with other codes already present there. With the new scheme the tags can be programmed with a unique code at the time of manufacture, e. g. as part of a credit card size unit also holding a magnetic strip for the customer's personal details. Clearly, with the large set of codes available it will not be possible for every receiver to search for all possible cards. The problem can be broken down as follows : Only premium passengers will hold personal cards. There will be a limited number of these in circulation. These can be searched for with a powerful processor upon entry into the terminal building.

The search size can be reduced by searching only for customers who are known to be flying around that time.

Once a passenger checks in he/she will be issued with a card (if not a premium passenger) whose code will be entered on the system he/she will relinquish the card on boarding the aircraft.

Internal to the terminal building the receivers will search only for cards that have been registered, either by recognition of a premium customer on entry or by issuing a card on check in.

It will not be possible for an illuminating transmitter to cover an entire airport terminal building and therefore a re-use pattern of illuminating transmitters will need to be established. They may or may not use the same frequencies in accordance with the particular application there may be some areas of non coverage at system boundaries. These boundaries are preferably chosen to be regions of low ambiguity about passenger location (e. g. corridors). It will also be essential that the pulsing waveform is fully synchronised (with low jitter) between all illuminating transmitters. This could be done, for example, using an off-air standard. The actual choice of pulsing frequency is not critical (to within a factor of 2) and could vary from one airport to another.

By using systems as herein described is possible to design low cost tags with an essentially limitless number of unique identities. The technologies identified here hold the potential for generating a universal tag for prisoner tagging, road tolling, test gear tagging, or airport luggage handling etc.

Claims (1)

1. CLAIMS 1. A tagging system comprising a plurality of tags, each for association with a person or an article to be tagged, a base station having a transmitter, for illuminating the tags with r. f. energy, and a receiver tuned to receive r. f. energy re-radiated from the tags, wherein the tags each comprise an antenna, which serves to receive r. f. energy radiated by the transmitter, a rectifier, which produces d. c. power from the received r. f. energy, a pulse generator which is driven by the d. c. power and synchronised with changes in the received r. f. energy to produce a'direct sequence spread spectrum' (DSSS) signal providing a unique code for each tag and wherein the base station receiver includes a decoder which is responsive to the re-radiated coded DSSS signals from the tags for tag detection and identification purposes.

   2. A tagging system as claimed in Claim 1, wherein the coded signals are'orthogonally'coded.

   3. A tagging system as claimed in Claim 1, wherein the coded signals are'near orthogonally'coded.

   4. A tagging system as claimed in Claim 2, wherein in order to provide the unique code for each tag, each of the tags is arranged to transmit a different rotation of a common M-sequence.

   5. A tagging system as claimed in Claim 3, wherein in order to provide the unique code for each tag, each of the tags is arranged to transmit a different rotation of an M-sequence with extension.

   6. A tagging system as claimed in Claim 2, wherein the unique codes are Walsh codes.

   7. A tagging system as claimed in Claim 2, wherein the unique codes are Walsh codes covered modulus 2 by an extended M sequence.

   8. A tagging system as claimed in Claim 2, wherein the unique codes comprise a rotation of a common M-sequence with a trailing bit in such a way as to provide full orthogonality.

   9. A tagging system as claimed in Claim 2, wherein the unique codes are Gold codes.

   10. A tagging system as claimed in Claim 3, wherein the unique codes are Gold codes with extension.

   12. A tagging system as claimed in Claim 1, comprising a sequence of orthogonal or near orthogonal codes as claimed in any of Claims 4 to 8, wherein the codes in the sequence are selected independently.

   13. A tagging system as claimed in any preceding claim, wherein synchronisation is facilitated by the generation of a synchronising pulse in the illuminating transmitter which serves to reset the codes in each of the tags so that they all start from the same point in their sequence whereby code generators which generate the codes each run asynchronously under control of their own crystal oscillators arranged to operate on the same frequency.

   14. A tagging system as claimed in Claim 1, and substantially as hereinbefore described with reference to the accompanying drawings.