Card reading assembly and self-service terminal equipped with the same as well as method for monitoring the same

Abstract

A card reading assembly for a self-service terminal includes a storing compartment for a card containing data to be read. The self-service terminal includes at least one sensor and an evaluation device connected hereto and the card reading assembly is protected against manipulation attempts by arranging at least one sensor in the card reading assembly and at least one linearly extending sensor arrangement that is attached in the storing compartment. The evaluation device checks at least one spatial dimension of the card via the sensor arrangement. Preferably, the sensor system is a sensor arrangement including a plurality of linearly extending sensor elements that extend in a first direction or a second direction in relation to the card retracted into the storing compartment. It can be effectively determined whether a retracted card is a genuine card of if a manipulation is present that targets the inside of the card reading assembly.

Claims

1. A card reading assembly for a self-service terminal, comprising a housing into which a card is inserted to be read, comprising sensors and/or actuators connected to an evaluation device and comprising a plurality of piezo-electric transducers, wherein the plurality of piezo-electric transducers are arranged in or on the housing such that the integrity of the card reading assembly, in particular the housing, is testable, the evaluation device configured to control a first of the plurality of piezo-electric transducers to excite at least a portion of the card reading assembly to vibrate, a second of the plurality of piezo-electric transducers configured to sense vibration and generate an output signal in response to the vibration of the at least a portion of the card reading assembly, the output signal corresponding to a sensed frequency of the vibration, the evaluation device also configured to receive the output signal from the second of the plurality of piezo-electric transducers, the evaluation device also configured to compare the received output signal with at least one predetermined frequency, and the evaluation device also configured to issue a warning signal in response to a deviation between the frequency represented by the output signal and the at least one predetermined frequency, the warning signal indicating a lack of integrity of the card reading assembly.

2. The card reading assembly according to claim 1, wherein a storing compartment is arranged in the housing in which the inserted card can be positioned in order to read out data stored on the card, the content of the storing compartment influencing the frequency represented by the output signal.

3. The card reading assembly according to claim 2, wherein the evaluation device manages reference data for different states of the card reading assembly, in particular whether a card is located in the storing compartment or whether there is no card in the storing compartment, wherein the reference data are managed by a classifier.

4. The card reading assembly according to claim 1, wherein the evaluation device is configured to receive the output signal in response to at least one event.

5. The card reading assembly according to claim 4, wherein the at least one event is one or more of the following: expiration of a period of time, detection by a sensor that a card has been inserted.

6. The card reading assembly according to claim 1, wherein the second of the plurality of piezo-electric transducers is operated at the natural frequency of the components and/or component groups or the housing.

7. The card reading assembly according to claim 1, wherein the plurality of piezo-electric transducers further comprise a third piezo-electric transducer, the third piezo-electric transducer positioned on a side of the housing serving as a sensor.

8. A card reading assembly for a self-service terminal, comprising a housing into which a card is retractable to be read, comprising sensors and/or actuators connected to an evaluation device and comprising a plurality of mechatronic transducers, wherein the mechatronic transducers are arranged in or on the housing such that the integrity of the card reading assembly, in particular the housing, is testable, the evaluation device being arranged to receive from the mechatronic transducers a signal which is excited by a part of the mechatronic transducers and is detected by a part of the mechatronic transducers in order to compare it with reference data, and to issue a warning signal in case of a defined deviation, indicating a lack of integrity of the card reading assembly, and wherein the evaluation device activates the mechatronic transducers to carry out a sweep from 100 Hz to 5 kHz, which starts at 100 Hz and ends at 5 kHz.

9. The card reading assembly according to claim 8, wherein the sweep for the analysis is divided into frequency ranges or is excited only in partial frequencies, whereby different methods for pattern recognition are applicable to each frequency range.

10. The card reading assembly according to claim 9, wherein a feature extraction is applicable to the frequency ranges, in order to then apply classifiers determining a condition to the extracted features.

11. A self-service terminal, in particular an automatic teller machine, with a card reading assembly according to claim 1.

Description

DESCRIPTION OF THE FIGURES

(1) In the following, the present invention is described in accordance with embodiments and the attached figures which show the following schematic representations:

(2) FIG. 1a shows a cross-sectional view of an installation of the card reading assembly;

(3) FIG. 1b shows a three dimensional view of the card reading assembly to be installed within a self-service terminal;

(4) FIG. 2 shows a schematic view of an arrangement of sensor patches to verify the dimensions (length, width, height) of a card;

(5) FIGS. 3a-c show logical connections between the steps of the method; a=inserting the card, b=retracting the card, c=checking/verifying the housing integrity;

(6) FIG. 4 shows a sweep divided into 6 frequency ranges, which are evaluated differently;

(7) FIG. 5a, 5b show clusters and their separations,

(8) FIG. 6a-c show weighted feature spaces, from a sensor, a differential signal that is not manipulated and is manipulated,

(9) FIG. 7 Classificator representations in different ranges shown in FIG. 4.

DETAILED DESCRIPTION REFERENCING THE FIGURES

(10) FIGS. 1a and 1b show a schematic view of the card reading assembly 20 comprising a storing compartment 13 for a card to be read. The storing compartment 13 also comprises the card reader device or card reading elements as such, that for instance comprise a contact area/pad for reading card chips and a reading head/pick-up to read magnetic strips. The card 11 or 11 to be read is supplied to the storing compartment 13 via the inserting slot by conventional means to be optimally positioned with respect to the card reading elements for reading. For this purpose conventional guiding and supply elements can be used.

(11) In the present invention card reading assembly refers to device as a whole (cf. FIG. 1b) thus comprising the housing 1, a base plate 2, a card reader device 3, in some cases a so called IDKG-add-on 5, additional sensors 6, in particular light sensor(s) or sensor arrangement(s), and optionally a camera 10, and card-transportation means. Depending on the actual version it is also possible that the device comprises less components. The term card reader device refers to the device 3 that is used for the actual reading of the card. The housing 1 circumferences the card reader device 3 in connection with the base plate 2 completely. Preferably, the transducer elements 14a, b (mechatronic transducers) are mounted at/in the housing 1; but basically a mounting at all other single components is possible, too. For this purpose it is useful to consider a superposition of the modal stretching (functions of strain) in the frequency ranges to be considered. By doing so significant and therefore suitable positions can be visualized and a positioning can be done. The positioning of the transducer can be done after analysing the housing. In the present design, a piezo element 14a is arranged on top of the housing as actuator and two 14b on the sides as sensors.

(12) The sole openings of the housing are represented by the opening area for insertion of the card (IDKG-slot unit/module 5) comprising the detection including the sensors 6 and by the opening for retraction of cards being monitored by the light barrier 7.

(13) As is shown in particular in FIG. 1b, the card reading assembly 20 comprises a retraction compartment 8 in its rear area that is intended for storing/withholding cards 11 which the self-service terminal, due to have not met specific conditions, cannot give back to the user. The compartment 8, which is referred to as retraction compartment is located at the end of the supply/transport chain, meaning even behind the storing compartment 13 in which the specific card is read. After reading or attempting to read the card 11, said card is transported further to the retraction compartment 8.

(14) The card reading assembly 20 is equipped with a sensor system (of FIG. 2) that is mounted to a sensor carrier (of FIG. 1A) and can exactly detect and check the spatial dimensions (length, width and optionally height) of the inserted card 11. Optionally a material determination via discrete spectroscopy in the IR-range can be performed by means of the sensor system.

(15) The sensor system is arranged such that at least one dimension can be captured/detected that is preferably the width b or the length l or optionally the height h of the card. The sensor system 6B measures the width b of the card but can also be used to measure the length l of the card, e.g. by a temporally triggered capturing by the sensor 6B, wherein the length of the card is determined via the intake velocity/intake time. Moreover, single sensors can be used for each dimension. Said sensors can particularly be sensor arrangements such as opto-electric sensor arrays or strips of the type TSL208R that are fabricated by the company TAOS and comprises 512 photodiodes linearly arranged in a distance of 125 m. Herewith a very precise measurement can be achieved. Furthermore, an additional sensor 6C can be arranged within the card reader device or the storing compartment 13 to measure or check the height of the card (in z-direction). Depending on the specific housing it can be sufficient to measure only one or two dimensions that are preferably the length and/or the width.

(16) By means of the integrated sensor systems 6A, 6B and/or 6C (optional) as well as by means of the light barrier 7 in combination with connection with the signal to retract coming from the card reader device 3 the slots of the housing can be secured. Additionally an installed camera 10 (of FIG. 1b) can be used. The functional connections are explained according to the FIGS. 4a-c.

(17) First of all it is referred to the FIG. 3a that shows the verification of the inserted card 11, wherein said verification/check is executed with the opto-electric sensor arrangement. In FIG. 4a there are functional blocks A1-A12 that represent the following:

(18) A1: The opto-electric sensor elements provide/generate measurement signals for a width b, a length l and optionally for the height of the card 11.

(19) A2: The evaluation device/electronics 4 checks/verifies the measured data/values comparing said values with standardized values of normalized banking cards.

(20) A3: If the measured values match/correlate to the standardized values the banking card is supposed to be a normal one.

(21) A4: Exciting via the piezo-electric sensor arrangement field 6D is preferably not done during operation of the card reader device.

(22) A5: However, monitoring of the card reader devices is executed, in particular of the card reader device signals and/or energy consumption of the card reader device.

(23) A6: If the measured data, as determined in A2, do not correlate to the standardized values, this indicates that a manipulation attempt has occurred.

(24) A7: Shutting down the card reader device, and retracting the manipulated card if possible.

(25) A8: The software control of the self-service terminal, which can be a PC, provides a warning signal.

(26) A9: An excitation can be executed at determined times of operation to verify the integrity of the housing.

(27) A10: An optional camera surveillance (cf 10 in FIG. 1a) can generate signals (images, video and/or audio).

(28) A11: The camera-signals are sent to the evaluation device 19 or to the computer in order to document the manipulation attempt and to store images of suspicious individuals for a subsequent identification.

(29) A12: Optional step wherein it is indicated/signalled that block/step A9 is executed if this is allowed by the card reader device data/signals.

(30) FIG. 3b is about monitoring the retract compartment via the sensor system or light barrier 7 (cf FIG. 1a) installed therein. In FIG. 3b there are functional blocks B1-B12 that display the following:

(31) B1: The opto-electric sensor system or light barrier 7 at the retract slot creates signals, if a card 11, a fake card or another object is transported through this slot or if an alien object is attempted to be inserted through the compartment 8 from behind into the card reader device 3.

(32) B2: The evaluation device compares the result to the status of the card reader device, meaning that the result is okay if there is a retract situation. A11 other results are considered to be manipulation attempts.

(33) B3: Depending on the signals and measuring values it is determined that a normal card has been transported/supplied trough the retract slot 7 or that a normal retract process has happened.

(34) B6: If the transport of an abnormal card trough the retract slot 7 or the absence of a normal retract procedure has been determined in block/step B2, this indicates that there is a manipulation attempt.

(35) B7: The card reader device is shut down/switched off.

(36) FIG. 3c refers to a verification of the integrity of the card reading unit. The functional principle shown in blocks/steps CI-CVII however refers to a material-check of the self-service terminal housing to determine if it has been manipulated. FIG. 4c refers to the verification of the housing (cf 1 in FIG. 1b):

(37) CI: The evaluation device 4 triggers the verification/check of the housing by exciting piezo-electric actuators that are mounted at the housing to vibrate and by evaluating the measured values coming from same wise mounted sensor arrangements. The actuators can be integrated within the sensor arrangements (comparable to 6D in FIG. 1b) or can be single piezo-electric elements of a certain field/area that are controlled to vibrate.

(38) CII: First of all the piezo-electric actuators are excited at known frequencies by a sweep.

(39) CIII: The sensors capture the signal pattern.

(40) CIV: The evaluation device evaluates via the described method.

(41) CV: If the integrity of the housing is verified, the cycle starts from CI.

(42) CVI: The integrity of the housing is not present, then the reader is switched off

(43) CVII: The card reader device is switched off; if necessary, even the entire self-service terminal

(44) In the following the verification of the card material via the piezo-electric or optical sensor arrangement 6D (cf FIG. 2) that is installed in the card reader device is described in detail. This solution can also be embodied/executed as an independent solution, but is described as a part of the disclosed method in the present description according to FIG. 2 and FIGS. 5-9:

(45) To verify the integrity of the housing 1 of the card reading assembly, the card material and/or the storing compartment for the card 11, the measurement signals coming from the sensor arrangements 6D are pre-processed in the evaluation device 4.

(46) FIG. 4 shows a sweep from 100 Hz to 5 kHz, which is divided into 6 frequency ranges, which are evaluated differently; the frequency ranges have already been described above.

(47) TABLE-US-00002 Range Frequency range 1 650-2140 Hz 2 2190-2550 Hz 3 2810-3470 Hz 4 3470-4000 Hz 5 4000-4300 Hz 6 4400-4880 Hz

(48) In the first range there are 45 eigenmodes of the housing, in the second range there are 11 eigenmodes. It becomes clear that the behaviour of the housing is not completely predictable and can be described by models. The system is too complex. This means that a separate calibration should be carried out for each installed system. The ranges are therefore broken down according to criteria such as eigenmodes in order to carry out different analysis in the ranges. The ranges are thus essentially determined by looking at the peaks of the signal. The blue range shows a normal course.

(49) In FIG. 4, 4 different signals were superimposed on each other to illustrate the differences. The blue signal represents a proper condition. The red signal shows a trimmed state and the turquoise signal shows a state that is not currently manipulated, but after manipulation. The green signal indicates a proper condition, but the valve of the housing has been opened first. The detailed views clearly show that the states differ from each other. These signals can be used to determine the ranges. The ranges have been selected to include features such as peaks or obvious differences. Range 1 contains many eigenmodes of the system. The range 2 has a peak which could be used well, because the trimmed signal has a much higher amplitude. Range 3 contains a peak. The ranges 4 and 5 each contain a peak with significantly increased amplitude. However, it can also be seen that the amplitude levels show clear differences between the individual signals. Range 6 contains the last eigenmodes that can be recorded with the sweep.

(50) FIG. 5a shows an exemplary cluster distribution and how these are summarized by a distance classifier.

(51) FIG. 5b shows an example of cluster separation using a straight line. Difficult for the separation with the help of a straight line is the fact that the method can be trained more elaborately, since the straight line has to be extracted from the training data. If one has the possibility to train the classifier with training data derived from both conditions, the straight line can be placed between the two clusters. However, if the training data only reflect one condition, it will be difficult to tell which condition is on the right or left of the straight line, since the position of the other condition cluster is unknown.

(52) At feature extraction the presented stochastic means were considered.

(53) The FIGS. 6a-6c show the feature spaces that are used for the algorithm. These feature spaces show the individual variations that are possible by looking at the signals and extracting features. The 6 ranges are distinguished. It can be seen that clusters are formed in different ranges without signal processing. A weighting of the data points is almost unimportant, but can make sense in individual cases. A closer examination of the feature spaces reveals a slight positive effect in the individual clusters.

(54) The sensor 14b, which is arranged on one side, forms in the range 2 and 3 reproducible clusters in case of a skimming manipulation. Range 3 of sensor 14b shows a better possibility for classification with the support vector classifier.

(55) The difference signal (see FIG. 6a), which is obtained by a subtraction between the two sensors 14b, appears to be a suitable evaluation feature, since there is a larger difference between the cluster of the normal state and the cluster in the skimmed state (see FIG. 6b). However, this approach seems to be only partially suitable for classification, please refer to the application below.

(56) Further sensor rows were used to check for other types of manipulation, such as bores, attachments or other changes to the housing. These series have shown that range 2 provides a good opportunity for analysis (see FIG. 6c). In contrast to the skimmer check, the differential signal is evaluated during the surface-mounting check, as otherwise changes on the opposite side of the sensor to be evaluated are not detected. The distance classifier is used for this evaluation.

(57) The combination of different approaches makes this algorithm complex but also very flexible for all kinds of attacks. A calibration function should be used for each classifier. The challenge in calibrating the support vector classifier is to extract a trend line from the feature values, which is subsequently shifted by a fixed value to differentiate the states. The distance classifier is calibrated unambiguously by calculating the mean value and the standard deviation.

(58) In order to classify the clusters within a range in a target-oriented manner, it is necessary to define meaningful classifiers that can be evaluated. In this method, a support vector and a distance classifier are used to solve the problem.

(59) The support vector classifier is used to detect skimmers. It is described by two points in a two-dimensional feature space. Two classifiers of this type are used for reliable classification.

(60) The distance classifier is used to detect other changes. It is completely described by a point and radius.

(61) When storing the classifiers, care must be taken to ensure a small amount of data.

(62) The extracted features include, for example, standard deviation; skew, Kurtosis, mean absolute deviation from the median; median of the absolute deviation. The standard deviation is a measure of the scattering of the values of a random variable around its expected value. The skew is a statistical key figure that describes the type and strength of asymmetry of a probability distribution. It shows whether and to what extent the distribution is inclined to the right (positive skewness) or to the left (negative skewness). Kurtosis is a measure of the steepness or peakedness of a (single peak) probability function, statistical density function or frequency distribution. The curvature is the central moment of 4th order. Distributions with low curvature scatter relatively evenly; for distributions with high curvature scattering results more from extreme but rare events.

(63) The median or mean value is an average value for distributions in statistics. The median of a collection of numeric values is the value that appears in the middle of the list when sorting the values by size. The mean absolute deviation from the median is the scattering around the median. In descriptive statistics and stochastics, dispersion (also known as statistical spread or mean absolute deviation) refers to various measures that describe the range of values of a frequency distribution or probability distribution around a suitable position parameter. The different calculation methods differ in principle in their susceptibility or sensitivity to outliers. The spread of the frequency distribution is called standard error.

(64) In addition to the installation of the opto-electrical sensor technology for testing the card dimensions (see sensor strips 6A and 6B as well as sensor 6C in FIG. 1a) and the piezo-electric sensor field 6D for testing the card material and/or the condition of the storing compartment, housing 1 of the self-service terminal (FIG. 1b) can also be equipped with piezo patches which monitor manipulations on the housing itself. The housing can be made of steel and/or plastic. Together with base plate 2 and IDKG insert 5, it forms a closed housing. The only openings are the card slots for card insertion and card retraction (area 8). The piezo patches are preferably glued on, but can also be injected directly into a plastic part. The sensors are operated by the evaluation electronics or evaluation device 4. The sensors can be operated both actuator and sensor. For this purpose, the evaluation device 4 actuates one of the sensors with a defined pattern and the remaining piezo patches acquire the excitation signal. The electronics adjusts the measured signal to the theoretical signal.

(65) Furthermore, the PC of the self-service terminal (e. g. ATM) is connected to the electronics. The electronics energizes the card reader device and is also (optionally) logically connected to it. The former is used to switch the card reader device on and off, while the latter is used to process any firmware signals from the card reader device, such as a retraction or card feeder. If the signal output of the card reader device is not yet implemented in its firmware, the current consumption of the card reader device can alternatively be measured and thus draw conclusions about the operating mode (card feeder/retract/output/standby) of the reader.

(66) The storing compartment (see FIG. 1b) guides and centres the card 11 to the card reader device 3, which is equipped with the said opto-electrical sensor technology, i.e. sensor and light barrier strip (s) that completely measures the geometric dimensions of the card. With the help of the at least one sensor strip, e. g. 6B in FIG. 2, it is possible to differentiate between a regular, valid card or an invalid object, e. g. a device for inserting a skimmer into the interior of the device. The signals from the sensor strip (s) 6A and/or 6B and the optional sensor 6C are evaluated by evaluation device 4. The same applies to light barrier 7 at the retraction output. Here, however, the assignment of the light barrier is not qualitatively evaluated, but an information fusion with the event or event Retract of the card reader device is created. In addition, the evaluation device can send 4 signals to the PC, which in turn activates the optional surveillance camera 10 via software (e. g. OSG) and checks the integrity of the card reader device slot.

LIST OF REFERENCE SYMBOLS

(67) 20 card reading assembly 1 housing 2 base plate 3 card reader device 4 evaluation electronics/evaluation device 5 IDKG insert 6 additive sensors on sensor carrier for card insertion monitoring 7 light barrier at the draw-in area (retract area) 8 storing compartment for retraction cards/withdrawn area (retract area) for cards to be retained 10 camera (s) (optional) 11 cash card (EC/Master/Visa) introduced 11 cash card (EC/Master/Visa) in the insertion slot 13 storing compartment for card (area of the card reader device) 6A, 6B 6B linearly extending sensor arrays, here sensor bars to check the length l or width b of the card 6C additional sensors to check the height of the card 6D sensor field with piezo electric sensor elements to check the material of the card 14a mechanical transducer/piezo element (actuator) 14b mechanical transducer/piezo element (sensor)