US 20050173530 A1
Magnetic stripe documents are identified and authenticated on the basis of repeatable magnetic characteristics of a stripe recorded with digital data and waveform characteristics of the sensed digital data. Stripes are sensed to provide an analog signal, only the portion storing leading zeros being used. Cyclically, the leading-zero portion is sampled and digitized to provide a target number of samples using a frequency locked loop. The target number of samples is then reduced by a selection process that attains a predetermined number of digital samples from central locations in magnetized portions of the record specifically, spaces between data transitions. Selected groups, each of a predetermined number of digital samples, are then combined to provide magnetic characteristic data. Such data is then combined with waveform data (range) to provide an identification word. Documents are verified by correlating reference and sensed identification words. One or several reference identification words also may be tested for excessive similarity suggesting a fraudulent copy.
1. An identification system for identifying authentic documents bearing a magnetic stripe recorded with digital data and having a repeatable magnetic characteristic, comprising:
a magnetic stripe sensor configured to sense the magnetic stripe to provide an analog signal representative of the recorded digital data and the repeatable magnetic characteristic;
a digitizer configured to sample a portion of the analog signal to provide digitized samples indicative of the repeatable magnetic characteristic;
a waveform circuit configured to provide range data characteristics of at least the sampled portion of the analog signal;
a storage configured to store representations of the digitized samples and the range data as identification data to identify the document; and
a processor configured to determine whether the digital data recorded on the magnetic stripe has been copied from the authentic document using the digitized samples of the repeatable magnetic characteristic; and
wherein the processor is also configured to determine whether the repeatable magnetic characteristic has been copied from the authentic document using the range data
2. An identification system according to
3. An identification system according to
4. An identification system according to
5. An identification system for identifying authentic documents bearing a magnetic stripe recorded with digital data and having a repeatable magnetic characteristic that has unique range characteristics, comprising:
a magnetic stripe sensor configured to sense the magnetic stripe to provide an analog signal representative of the recorded digital data and the repeatable magnetic characteristic;
a magnetic characteristic circuit configured to provide magnetic characteristic representations indicative of the repeatable magnetic characteristic;
a waveform circuit configured to provide range representations indicative of the range characteristics of the analog signal generated from a portion of the magnetic stripe including the repeatable magnetic characteristic by the magnetic stripe sensor; and
a forming circuit to provide document identification representations based on the magnetic characteristic representations and the range representations to identify the documents.
6. An identification system according to
7. An identification system according to
8. An identification system according to
9. An identification system according to
10. An identification system according to
11. A system for use with a card bearing a magnetic stripe having a repeatable magnetic characteristic that has known range characteristics and recorded with digital data in the form of magnetic transitions, said system for providing a sensed characteristic identification for the card, comprising:
means for sensing said magnetic stripe to provide representations of digitally recorded data and representations of the repeatable magnetic characteristic in the form of digital sample signals;
means for selectively storing card identification words formed from the digital sample signals to manifest the repeatable magnetic characteristic of a card and known range characteristics of the repeatable magnetic characteristic.
This is a continuation-in-part of application Ser. No. 09/120,816 entitled MAGNETIC STRIPE CARD VERIFICATION SYSTEM filed Jul. 22, 1998.
Generally, the present development relates to systems and processes utilizing the repeatable characteristics of a magnetic stripe for authenticating stripe bearing objects, e.g., documents as credit cards.
Various forms of cards and other documents bearing a magnetic stripe have long been used for a variety of different purposes. Such magstripe cards are currently used in large numbers, for example in the forms of credit cards, debit cards, tickets, passes, I.D. cards and so on. Typically the magnetic stripes of such cards carry digitally recorded data, for example relating to the use of the card, the assigned user, and so on.
Although magnetic stripe documents are widely and successfully used in commerce and industry, various forms of copying have become common place. The aggregate losses from fraudulently copied magstripe cards are very large. Consequently, a continuing need exists for a practical and economical system and method to reliably identify, and thus verify or authenticate magstripe documents.
Over the years, there have been numerous proposals for verifying documents, including techniques for authenticating magnetic stripes. A substantial number of prior proposals have been based on a concept of using certain magnetic characteristics of the magnetic stripe to verify documents, as cards. In that regard, it has been determined that magnetic stripes possess inherent, substantially unique, magnetic characteristics that can be repeatedly sensed. Over the years, these characteristics have been recognized as contributing a signal component sometimes considered to be noise in the sensed analog signal. Thus, along with the sensed data, a repeatable signal component (noise) also is present. Essentially, just as the magnetic characteristics of individual stripes are distinct, the resulting “noise” signals are somewhat unique and can characterize or identify magnetic stripes. Thus, it has been proposed to employ such magnetic characteristics and the resulting “noise” signals as a basis for identifying and verifying individual magnetic stripe documents. In some instances, the general technique has been compared to the science of human fingerprinting, that is, involving the use of a somewhat unique physical characteristic for purposes of identification. However, although various verification techniques and structures, based on the noise or characteristic magnetic component of a sensed magstripe signal have been proposed, they have not been widely accepted.
One prior proposal involved using a separate non-data magnetic stripe placed on a document specifically for purposes of identifying the document. Other proposals have involved various treatments of the magnetic stripe, usually involving the signals that are recorded and sensed from magnetic stripes.
Typically, the magstripes on conventional magstripe cards are divided into parallel tracks that are digitally recorded and sensed. Many of the prior proposals for magstripe document verification have involved the record of one track of the magnetic stripe in the form of digital data represented by magnetic transitions along the length of the magnetic stripe. Portions of analog signals sensed from the magnetic track transitions, or other portions of the signal have been proposed for use in characterizing the card or document. For example, on a time base, peaks in the sensed analog signal (representing magnetic transitions) have been proposed for use in verifying the magnetic stripe. Other proposals have suggested using other portions of the sensed analog signal from one of the magnetic tracks. For example, the relatively flat signal portions representative of magnetized regions of the stripe (located between magnetic transitions) have been proposed for use.
Various techniques also have been proposed for attaining more consistent results in sensing the magnetic characteristic signals. For example, techniques have included: averaging the results of multiple sensings, oversampling (increased sampling as by a factor of one hundred) and using a phase locked loop to control sampling. It also has been opined that for identification and verification purposes, sizable variations are quite acceptable between individual sensings of magstripe characteristics.
In the final analysis, as indicated above, in spite of the numerous structures and techniques proposed for magstripe fingerprinting, the concept has not gained widespread acceptance. That is, using the so called “remanent noise” in a signal sensed from magstripes to verify individual documents has not come into widespread use. Rather, although numerous proposals have been suggested for such operation, the technique simply has not gained significant commercial or industrial acceptance.
Typically, magnetic stripe fraud involves variously copying the recorded data (including verification data) in either a digital or analog form and using it to verify a transaction. Thus, a considerable need exists for a reliable system of combating fraud related to magstripe documents, which fraud may involve various forms of copying, duplicating or misdirecting data.
In general, the system of the present invention is based on recognizing the various problems of using the repeatable characteristics of a magstripe for purposes of verification, for example, in broad commercial and industrial applications. In that regard, it has been determined that many problems arise in the arenas of broad industrial and commercial use which are not apparent in the laboratory.
One major consideration relating to the extensive use of magnetic characteristics for card verification involves the number of cards used in a system. For example, a commercial credit-card system may be required to accommodate many thousands of individual cards operating randomly in combination with hundreds, if not thousands, of individual processing units. Furthermore, in an extensive system, reliability becomes exceedingly important, particularly in the realms of financial and security transactions, as are involved with financial cards.
In general, the system of the present invention involves improvements that enable magnetic characteristic verification techniques to operate successfully and reliably in the realms of widespread commercial and industrial applications. Specifically, the processes and systems of the present invention are directed to developing reliable magnetic stripe verification in a large scale system.
As disclosed herein, repeatable magnetic characteristic signals (resulting from specific characteristic features of the magstripe) may be selected from defined areas of a digitally-recorded magnetic stripe. The defined areas may be located between magnetic data transitions. Such defined areas of the stripe may be magnetized to a level of saturation and accordingly produce relatively flat (stable) sections in the sensed analog signal. Although these signal sections are relatively flat, they manifest a low level signal (noise) representative of the repeatable magnetic characteristics or features of the stripe. Thus, such signal sections may be used to reliably characterize each magnetic stripe in a manner somewhat analogous to a fingerprint. As disclosed herein, further reliability is attained by testing signal amplitude and symmetry characteristics as well as challenging digital copies by excessive similarity tests involving recently sensed characteristics.
Regarding the basic characteristic determination, as disclosed herein, after sensing a magstripe to produce a magnetic analog signal (typically containing digital data) a target number of samples are taken from several relatively flat sections of the analog signal using a frequency lock loop. Individual samples in each group are converted to a digital form, then selectively processed to obtain a predetermined subset of samples. The predetermined subset of digitized samples is then further processed by offset selection to accomplish a sub-subset of digitized samples that become a component of a characteristic designation used for identification. That is, several sub-subsets of digitized samples are collected to form part of an identifier for the magstripe-bearing card.
With the magnetic characteristic resolved to a digital representation, dynamic range information is extracted from the digitized samples to supplement the magnetic characteristic in the representative identifier. In the disclosed embodiment, the dynamic range information is based on amplitude symmetry or waveform characteristics as disclosed in detail below. Generally, by including dynamic range information in the identifier for a magnetic stripe, certain forms of the copied identifier can be distinguished as disclosed in detail below.
Other methods of copying may present the identifier in a digital form, for example, in the ultimate form for the test comparison. As disclosed herein, such copying is challenged by comparing the presented identifier with recently sensed identifiers which are the likely source of copies. If the degree of similarity is excessive, copying is suggested. Otherwise, the freshly sensed identifier is tested by correlation with reference identifiers to verify the card or other document.
In the drawings, which constitute a part of this specification, exemplary embodiments exhibiting various objectives and features hereof are set forth, specifically:
As is required, detailed illustrative embodiments of the present invention are disclosed herein. However, documents, physical magnetic systems, data recording formats, and operating structures in accordance with the present invention may take a wide variety of different forms, some of which may be quite distinct from those of the disclosed embodiments. Consequently, the specific structural and functional details disclosed herein are merely representative; yet in that regard, they are deemed to afford the best embodiments for purposes of disclosure and to provide a basis for the claims herein which define the scope of the present invention.
Functionally, the sensing unit 14 produces at least one analog signal AS from a recorded track that carries both the recorded digital data DD of the stripe 16 and representations of the repeatable magnetic characteristic MC. In accordance herewith, the analog signal AS is processed to provide three digital representations, i.e. digital card data (e.g. account number, etc.) magnetic characteristic data and waveform or range characteristic data. The card data is used to process the card, and possibly to enable or reflect a transaction. The characteristic data (magnetic and waveform) serves to verify the card, as against, direct and indirect copies, as well as misdirected data. Furthermore, in accordance herewith, representations of the characteristic data are compared with prior presentations for excessively high degrees of similarity or exactness, which may suggest copying. Note that herein, similar designations are used for data and representative signals.
Returning to the structure of
From the circuit 18, the digital data signal DD is provided to a control and transaction system 28 for further processing as described below. In various applications, the data signal DD may be variously employed, as for example, to support a credit transaction or drive an automatic teller machine.
As described in greater detail below, in the extraction circuit 20, the characteristic analog signal AS is sampled during defined sampling intervals SI, the samples being digitized to accomplish groups or sets of digitized samples DS. Selected samples from the sets representing the magnetic characteristic, are developed by forming circuits 22 and combined with range data signals (circuits 25) in the circuits 22 to provide a sensed test word STW.
Representative signals for the “sensed” test word, termed STW, are applied to a correlation circuit 24. To function, the correlation circuit 24 also receives at least one “reference” test word RTW from a source unit 26. The reference source unit 26 may take a variety of different forms. For example, a sensed test word STW could be formed as described herein then recorded digitally on the card 10 for direct provision by the unit 26 as a reference test word STW. However, in the currently preferred embodiment, the reference unit 26 accesses a data bank (not shown) to provide several reference test words RTW for correlation with a freshly sensed test word STW as described below.
In addition to the correlation of reference and sensed test words by the correlation unit 24, a preliminary test of the words is performed by a bit-by-bit coincidence detector 27. Basically, the detector 27 tests the sensed test word STW against the reference test words RTD for degrees of exact similarity which may suggest fraud. If the test is successful (no excessive similarity), then independently, the correlation unit 24 correlates the test words. If that test is successful, the central and transaction system 28 is signaled.
Note that the control and transaction system 28 may incorporate offsite components and furthermore, may incorporate or overlap apparatus involving the source unit 26. For example, the transaction system 28 may be integral with the reference unit 26 comprising a structure for addressing a data bank containing data on a card user's account number. As a consequence, the data bank provides at least one reference test word for comparison tests. Also note that such data can be combined with other identifiers, e.g. PIN, PAN, etc. to combat fraud otherwise by various comparisons.
The control and transaction system 28 also may embody various other operating equipments or systems, e.g. automatic teller machines, entry or access controls, card verification units, and so on. Thus, the digital data signal DD is processed and variously utilized by the control and transaction system 28.
As suggested, the stripe 16 (
Next in the track 29 format of
In relation to the digital data recorded on the stripe 16, for example, in the track 29 (
In accordance herewith, it is important to understand that magnetic characteristic data MC is sensed from the magnetized regions 42 (
Below the track fragment 39 (
For the magnetic characteristic MC, as mentioned above, digital samples of the variations 48 are formed into digital representations. For the dynamic range information DR, waveform characteristics of the analog signal AS are reduced to digital representations. For example, as disclosed below, the dynamic range information might for example take the form of relative peak amplitudes or line offset ratios in the analog signal AS. Specifically, referring to
A specific test word TW thus characterizes each magnetic stripe card and is stored for reference comparisons as indicated above. However, in accordance herewith, it has been recognized that a prior test word STW might be copied in an effort to support a fraudulent transaction. Generally, such a copied word will result in a closer correlation than if the test word STW were freshly generated. That is, by digitally copying a test word, an exact coincidence may occur, which event would be very unlikely in sensing and developing a fresh test STW. Consequently, in accordance herewith, a record of reference test words RTW (e.g. original and recently used words) is stored for bit-by-bit correlation with words asserted as proper sensed test words to support transactions. The occurrence of a very close bit-by-bit comparison suggests fraudulent digital copying of a test word. For example, a dissimilarity of only four or less bits between a sensed test word STW and an asserted reference test word RTW may suggest fraud.
In accordance herewith, asserted test words STW first are compared with reference test words RTW seeking excessive similarities. In the event of a near identity, the transaction is denied based on the suggestion of digital copying. Essentially, a drop off list of the last several sensed test words STW may be stored as reference test words RTW for comparison. If the bit-by-bit comparison indicates sufficient dissimilarity, the system then proceeds with a correlation of a reference test word RTW with the sensed test word STW, asserted as being freshly sensed.
Recognizing that a variety of structures and processes may be employed to develop test words with magnetic characteristic data MC and dynamic range data DR, techniques of the disclosed system will now be considered. Referring to
During the sampling interval SI (
During a defined sample interval SI (peak-to-peak
In the illustrative system, the frequency lock loop 66 has a target of defining one hundred twenty-eight samples during each sample interval SI (
The A/D converter 68 (
In accordance herewith, a predetermined number of digitized samples DS (subset) is to be consistently used for each sampling interval SI. To accomplish the predetermined number of samples DS for each sampling interval SI, a re-sizing operation is performed to produce a subset (e.g. “64”) from the varying total number of samples (e.g. “128
In the disclosed embodiment, the samples for representing the magnetic characteristic MC are taken from the last eight interspersed portions 56 (
The forming circuit 22 (
The control unit 80 (
Essentially, sets (e.g. “8”) of the digitized samples DS (
It is to be noted that certain elements of the forming circuit as shown in
In view of the above descriptions, a complete understanding of the disclosed system can now best be accomplished by pursuing an explanation of the process and providing further component descriptions where pertinent. Accordingly, assume the presence of a magnetic stripe card 10 (
The extraction circuits 20 first digitize the analog signal AS (
This operation is performed because even though the sampling intervals are controlled by a frequency lock loop, in an operating environment, the number of samples taken during a sample interval may vary, e.g. potentially including either a few additional samples or a few less samples. Re-sizing to obtain a specific number of samples (e.g. “64”) has been determined to be somewhat critical in certain applications. The processor 90 (
If the target number of one hundred twenty-eight was accomplished, the process simply proceeds to a junction 110 (
Assume the deviation from the target number is one hundred twenty-five samples. In that event, the process proceeds to a step as illustrated by a block 114 which is illustrated graphically in
To add samples, select the underlined:
To remove samples a similar alternate selection is employed. For example, to obtain a subset of eight samples from a set of seventeen the pattern might be as follows:
To delete samples, select the underlined:
Thus, according to the process, alternate samples are taken with additions to accomplish a subset of exactly sixty-four digitized samples DS. With such a predetermined subset, the operation proceeds to the junction 110, again, treated subsequently.
If the initial set of digitized samples DS constituted one hundred twenty-six samples, the process (
As indicated above, the actual number of samples in a set also may exceed the target number of samples. In that event, the process proceeds to a query block 120 (
Similar deletions occur in the event of two or three excess samples as represented by the blocks 124 and 126. In such events, samples are dropped in addition to the alternate selection. Thus, with a subset of precisely sixty-four samples DS, the process proceeds from the junction 110.
With each subset of precisely sixty four digitized samples, the next phase of the process is performed by the processor 92 (
The operation of the offset processor 92 may be relatively simple as represented by a block 128 in
With the accomplishment of each sub-subset (twenty-four samples DS) the process proceeds to a query block 130 (
To form the digitized dynamic range data as a digital word, digitized samples DS stored in storage 88 (
In the disclosed embodiment, the waveform characteristic actually is a standard parameter, specifically identified as “ISO/IEC Standard 7811-6, Table 1—parameter Ui6”. The indicated parameter has other uses in relation to magnetic tape; however, in the illustrative embodiment it serves as a specified waveform characteristic. Referring to
To consider an overall operation, assume the card 10 (
With the card recorded and the reference test word RTW stored, the card is issued to a holder. In
In due course, the card is presented, for example, to support a transaction as indicated by the block 156. Accordingly, the card is processed by structure in accordance with
The card 10 is sensed for the analog signal AS which is to provide the digital data DD (identification) and the sensed test word (block 158).
Using the digital data DD, the control and transaction system 28 (
The sensed test word STW and the reference test words are supplied to the detector 27 comparison. The detector 27 executes a bit-by-bit comparison between each referenced test word RTW and sensed test word STW seeking coincidence. Structurally, the detector 27 may simply involve a stepping register for moving the referenced test words RTW through stages for repeated comparison with the sensed test word STW to indicate coincidence between individual binary bits. Each lack of a bit coincidence is tallied. The processed step is illustrated by block 162 in
As explained above, total coincidence between one of the reference test words RTW and the sensed test word STW is a very strong indication of digital copying. Essentially, although the nature of the “fingerprint” system involves identification by a degree of coincidence, identity of coincidence is exceedingly unlikely. Accordingly, the detector 27 is set to count bit-by-bit deviations and provide a represented total to the correlation unit 24. Only in the event of less than substantial coincidence between the sensed test word STW and each of the referenced words RTD does the unit 24 proceed. Such deviation, indicates the lack of digital copying from a recently sensed test word. On the contrary, a coincidence or near coincidence (low number of deviations) prompts the detector 27 to signal the system 28 indicating a rejection of the transaction. These operations are represented in
The continuing test is executed by the correlation unit 24 (
With the occurrence of an approved transaction, the sensed test word STW, as provided, is stored (
In that regard, the sensed word is passed from the correlation unit 24 through the transaction system 28 to the source unit 28 for storage. Thus, a limited list or group of reference test words RTW are provided for subsequent comparison. In that regard, generally it has been determined expedient to limit the number of active reference test words to a predetermined number. Consequently, with the storage of a fresh test word, a test may be performed limiting the active group to the predetermined number (
As a final step in the process, a card 10 may be released as indicated by block 186 in
In executing the process, it is noteworthy that the implementing components may be variously located. Typically, locations will involve a central station serving a multitude of remote terminal stations, e.g. merchant locations. In accordance with one embodiment, implementing circuits and so on to formulate the sensed test word STW and the digital data DD are located at card terminals, with the balance of the system located at a central terminal. However, several other variations may be practical for individual systems.
In view of the above explanation of an exemplary system, it will be appreciated that other embodiments of the present invention may be employed in different applications to verify, authenticate, identify or confirm documents, particularly magnetic stripe-bearing cards. While exemplary operations have been stated herein and certain detail structures have been disclosed, the appropriate scope hereof is deemed to be in accordance with the claims as set forth below.