Money item acceptor with memory facility for rejected money items

Field of the invention This invention relates to an acceptor for money items such as coins and banknotes and has particular but not exclusive application to a multi-denomination acceptor.

Background

Coin and banknote acceptors are well known. One example of a coin acceptor is described in our GB-A-2 169 429. The acceptor includes a coin rundown path along which coins pass through a coin sensing station at which sensor coils perform a series of inductive tests on the coins in order to develop coin parameter signals which are indicative of the material and metallic content of the coin under test. The coin parameter signals are digitised and compared with stored coin data by means of a microcontroller to determine the acceptability or otherwise of the test coin. If the coin is found to be acceptable, the microcontroller operates an accept gate so that the coin is directed to an accept path. Otherwise, the accept gate remains inoperative and the coin is directed to a reject path.

In banknote acceptors, sensors detect characteristics of the banknote. For example, optical detectors can be used to detect the geometrical size of the banknote, its spectral response to a light source in transmission or reflection, or the presence of magnetic printing ink can be detected with an appropriate sensor. The parameter signals thus developed are digitised and compared with stored values in a similar way to the previously described prior art coin acceptor. The acceptability of the banknote is determined on the basis of the results of the comparison. Also, the denomination of the banknote can be determined so that credit corresponding to the monetary value of the banknote can be allocated.

Money item acceptors can currently achieve approximately 90% acceptance of true banknotes but it would desirable to increase this percentage acceptance rate. The stored reference data against which money items under test are compared,

needs to be highly accurate and this is crucial to obtaining a high acceptance rate for true money items.

Summary of the invention According to the invention, data corresponding to money items rejected by the acceptor are stored for subsequent analysis in order to improve upon the stored reference data and thereby improve the acceptance rate for true money items.

According to the invention, there is provided a money item acceptor for authenticating a money item under test, comprising: a detector to provide parameter signals corresponding to the money item under test, a processor to make a comparison of parameter data corresponding to at least some of the parameter signals from the detector, with stored reference data corresponding to an acceptable money item, to allow the money item under test to be accepted or rejected in dependence on the outcome of the comparison, and a memory to store data corresponding to the parameter data for the money item under test when rejected.

The stored data may be transferred to a remote location for analysis, for example in a batch corresponding to a plurality of rejected money items. The transferred data may be a function of the difference between the parameter data for the or each rejected money item and the stored reference data.

The transferred data may be encrypted to provide security.

The acceptor may also be configured to store data corresponding to the parameter data for money items under test when accepted for example when the parameter data falls within upper and lower acceptance margins corresponding to the tails of a probability distribution of values of the parameter signals for an acceptable money item.

The money item acceptor may comprise a banknote acceptor or a coin acceptor.

The invention also includes a method of authenticating a money item under test, comprising: detecting a money item under test to provide parameter signals corresponding to the money item, comparing parameter data corresponding to at least some of the parameter signals with stored reference data corresponding to an acceptable money item, to allow the money item under test to be accepted or rejected in dependence on the outcome of the comparison, and storing data corresponding to the parameter data for the money item under test when rejected.

Brief description of the drawings

In order that the invention may be more fully understood an embodiment thereof will now be described by way of example with reference to the accompanying drawings in which:

Figure 1 is a schematic plan view of a banknote acceptor device according to the invention;

Figure 2 is a schematic sectional view of the device shown in Figure 1 ;

Figure 3 is a schematic block diagram of the device shown in Figures 1 and 2;

Figure 4 is a schematic illustration of a banknote sampling frame Fl for a banknote entering the acceptor at a skewed angle, and reference frame F2 for stored banknote reference data,

Figure 5 is a flow diagram of processing steps performed by the microcontroller of the acceptor, for accepting or rejecting a banknote,

Figure 6 is a flow diagram for a routine in which difference data corresponding to the difference between stored reference data and corresponding data from a rejected banknote is stored, and

Figure 7 is a flow diagram for encrypting and downloading the difference data to a remote location for analysis,

Figure 8 is a schematic illustration of a coin acceptor,

Figure 9 is diagram of the circuits of the coin acceptor shown in Figure 8, Figure 10 is a graph of probability distributions for a true coin and a fraudulent coin of e.g. a foreign coin set, with similar characteristics to the true coin, and

- A -

Figure 11 is a flow diagram of processing steps performed by the microcontroller of the coin acceptor, for accepting and rejecting a coin under test, and for storing data for rejected and nearly rejected coins.

Detailed description of embodiments of the invention

The invention has application to both money item acceptors in the form of banknote acceptors and also coin acceptors and examples of them will now be described in more detail. The term "money item" as used herein includes both banknotes and coins.

Banknote acceptor

A banknote acceptor is shown in Figures 1 and 2. As used herein the term "banknote" means a promissory note especially from a central bank or other governmental organisation payable to the bearer on demand for use as money, also known as "paper money" and in the USA as "currency" or a "bill", and term also includes other sheet objects with an attributable monetary value, such as tokens and vouchers.

Referring to Figures 1 and 2, a banknote acceptor 1 receives a banknote 2 through an inlet 3 wider than the banknote, such that the banknote passes along a path 4 shown in dotted outline to an outlet 5 through a sensing station S where image data corresponding to the banknote is captured to determine its authenticity.

A solenoid operated gate 6 is disposed at the outlet 5 to direct acceptable banknotes along an acceptance path shown by arrow 7, or to rotate to a position shown in dotted outline to direct unacceptable banknotes along reject path 8 shown in dotted outline. Alternatively, an unacceptable banknote can be rejected by reversing it back through the inlet 3, as described in more detail below.

As shown in Figure 2, the path 4 for the banknote is defined between a main body 9 having a platen 10, overlaid by a spaced upper part 11. As shown in Figure 1, the platen 10 is formed with upstanding regions 12, 13 that define side edges of

the path 4. The banknote 2 is driven along the path 4 by means of a belt and pulley arrangement 14 and a roller 15 driven by electric motors (not shown).

The banknote 2 under test can be illuminated with optical radiation in three different ways at the sensing station S to test its reflective properties on each side and also its transmissive properties. To test reflection from the upper surface of the banknote 2, a light source 16-1 extends transversely across the platen 10 and directs optical radiation downwardly in a flat beam across the entire width of the platen 10. One example of the source 16-1 is an array of surface mounted LEDs arranged in closely packed rows, to emit optical radiation of different wavelengths, in a light box covered by a diffusing sheet to provide spatially uniform illumination over a broad optical band. The optical radiation may be visible or non-visible radiation such as ultra violet or infra-red. A light emitting polymer sheet or other light sources can be used as an alternative to the light box. The optical radiation from source 16-1 is reflected by the banknote 2 towards planar mirror 17-1, which directs the reflected radiation towards a sensor 18-1. The sensor 18-1 in this example comprises a TAOS device with a row of 120 pixel CCD sensors. In use, only a portion of the row of pixels is used in order to accommodate variances in alignment that occur during manufacture, and for example only a successive run 102 of the 120 pixels may be utilised for signal processing when detecting banknotes.

A lens arrangement comprising converging lens 19-1 and associated stop 20-1 directs light from the mirror 17-1 onto the sensor 18-1. The lens arrangement may be telecentric although other lens configurations can be used. An advantage of a telecentric arrangement is that it provides an image of fixed size regardless of variation in distance of the banknote 2 from the lens 20-1 in the region of the sensing station S. The image focus quality will change slightly with variations in distance to the banknote, but the image will not change in size. The use of a small aperture for the stop 20-1 increases the depth of field and so makes focus errors of less significance. The lens system can be configured so that despite movement of the banknote relative to imaging system and assembly errors in the building of

the apparatus, the image size will always cover the same number of pixels on the CCD sensor array 18-1.

In order to test the transmission properties of the banknote, a second light source 16-2 extends across the width of the platen 10 and directs optical radiation downwardly through a transparent window 21 towards mirror 17-2 where it is reflected through lens 19-2 with an associated stop 20-2, to a second CCD sensor array 18-2.

The reflective properties of the underside of the banknote are tested using a third optical source 16-3 that directs optical radiation into region of the window 21, to be reflected by the banknote 2 towards mirror 17-2 and then to sensor 18-2 lens and stop arrangement 19-2, 20-2.

The banknote thus can be analysed in terms of its optically reflective properties on both sides, and also in terms of its transmissive properties. Appropriate data can be gathered by selective use of the light sources 16-1, 2, 3, so as to provide sampling data to processing circuitry 21 shown in Figures 2 and 3. The banknote can be accepted or rejected in the manner described hereinafter, by using the gate 6 to direct acceptable banknotes along accept path 7 and rejected banknotes along reject path 8. Alternatively, the belt and pulley arrangement 15 can be driven in reverse to reject the banknote 2 through the inlet 3 after it has been fed in its entirety from the inlet 3 through the sensing station S.

When the banknote 1 is initially inserted into the inlet 2, the drive belt and pulley arrangement 14 progressively moves the banknote through the sensing station S so that successive rows of pixel data are developed by the detectors 18 over the entire surface region of the banknote. In this example, it is assumed that the rows of pixelated data are derived from the use of optical source 16-1 and associated CCD detector 18-1 although the ensuing description applies equally well to data developed at sensor 18-2 in response to optical radiation from light sources 16-2 or 16-3.

The sensor array 18-1 comprises a CMOS chip. The individual pixels of the array are closely spaced on the chip 18 and the lens 19-1 ensures that each pixel is responsive to a respective sampling location disposed along the line A-A', across path 4, as illustrated by dotted lines 22 in Figure 1.

Processing circuitry 23 for controlling operation of the device may be mounted in the main body 9. The processing circuitry 23 is shown in block diagrammatic form in Figure 3 and comprises a microcontroller 24 that receives digital samples from the pixelated photo sensors in chip 18. It will be understood that digital samples can be received from either of the chips 18-1, 2 and only one is shown to simplify the explanation. The data samples are compared with corresponding samples for acceptable banknotes stored in memory 25. As explained in more detail later, the successive rows of data samples may be pre-processed and stored in the memory 25 so that an image of the face of the banknote under test can be displayed to the user for authentication purposes.

Operation of the belt and pulley arrangement 14 shown in Figure 2 is controlled by the micro controller 24 through a driver circuit 26. The gate 6 is driven by driver circuit 27 so that acceptable banknotes are allowed to pass along path 7 and non-acceptable banknotes are passed along path 8 as illustrated in Figure 2. Alternatively, rejected banknotes can be reversed out of the inlet 3 by the microcontroller 24 instructing the driver 27 to reverse the belt and pulley arrangement 14 and roller 15.

The light sources 16-1, 2, 3 (shown collectively in Fig, 3 as light source 16) are operated individually under the control of the microcontroller 24 through a driver circuit 28.

Referring to Figure 2, the banknote acceptor includes a display panel 29 that includes a first display device 30 that displays an image 31 of the banknote under test derived from the pixelated data stored in memory 25 as a result of the

banknote 2 passing through the sensing station S. The display panel 29 also includes a second display device 32, which may comprise one or more seven segment display units which display the denomination the banknote under test, which as will be explained hereinafter, is determined by comparing the pixelated data corresponding to the banknote with corresponding reference data held in memory 25 for acceptable banknotes of known, different denominations.

Display panel 29 also includes an actuator 33 in the form of an accept button that is depressible by the user to indicate acceptance of the banknote on the basis of the banknote image 31 and the corresponding denomination displayed on the display unit 32. A reject button 34 is provided on panel 29 to allow the user to reject the banknote under test. The display panel 29 further includes a button 35 which allows the user to retrieve images of previously accepted banknotes. Thus, by successively actuating the button 35, successive images of previously accepted banknotes are displayed.

As shown in Figure 3, the displays 30, 32 are coupled to the micro controller 24 together with the accept, reject and retrieve buttons 33-35, to enable the micro controller 24 to control and coordinate the data displayed on the displays 30, 32, and acceptance or rejection of the banknote under test.

Banknote acceptance and rejection

In use, when the banknote 2 is inserted into the inlet 3 shown in Figures 1 and 2, it passes through the sensing station S and successive rows r of pixelated image data are captured by a selected sensor array 18 and fed to the microcontroller 24. The pixelated signals thus comprise parameter signals that characterise optical characteristics of the banknote under test. This banknote parameter data is processed and compared with reference data corresponding to acceptable banknotes held in memory 25. If the banknote is found to be acceptable as a result of the comparison with the reference data, denomination data corresponding to the banknote denomination is displayed on the display 32 and an

image of the banknote derived from the captured, pixelated, data is displayed on the display 30.

The user is thereby given an opportunity to review the banknote image 31 and its detected denomination indicated on display 32. If the displayed data is acceptable to the user, the accept button 33 is operated, in which case the micro controller 24 instructs the gate driver 26 shown in Figure 3 to move the gate 6 to an acceptance position, and the belt driver 27 is operated to move the banknote along the accepted path 7. However, if the user does not agree with the displayed image or denomination on displays 30, 32, the reject button 34 can be operated, which correspondingly causes the banknote under test to be rejected, either by passing it to reject path 8 by appropriate operation of gate 5, or by reversing the note back out through the inlet 3.

This process will now be described in more detail with reference to Figure 5. At step S5.1, successive rows of image data are detected from the banknote under test at the sensing station S. Referring to Figures 1 and 4, the width of the path between the inlet 3 and outlet 5, is wider than some of the denominations of banknote to be tested, because different denominations of banknotes have different widths. As a result, the banknotes may not pass along the path 4 parallel to the side edges 12, 13 of the path through the acceptor 1. This is illustrated schematically in Figure 4 in which banknote 2 is shown in solid outline is illustrated passing along the path in a direction of arrow 4, parallel to the side edges 11, 12 of the acceptor path. However, the banknote 2' is illustrated at a skewed angle θ to the direction of the path. Also, the banknote may not necessarily pass along the longitudinal centreline of the path depicted by arrow 4. Instead, it may be shifted to one side or the other depending on the manner in which the note 2 is inserted into the inlet by the user. Thus, for a banknote 2' depicted in dotted outline, successive rows of pixelated data^ developed by the detector 18-1 may be skewed relative to the side edges of the banknote. The pixelated data is thus developed in a banknote sampling frame Fl that is skewed relative to a reference frame F2 shown in Figure 4. The stored reference data for

acceptable banknotes held in memory 25 are held in the reference frame F2 whereas the captured pixelated data is in the banknote sampling frame Fl, which may vary from banknote to banknote under test.

Referring to Figure 5, at step S5.2, the micro controller 24 utilises a de-skewing algorithm to convert the sampled data for the banknote under test from the banknote sampling frame Fl into the reference frame F2. This can be achieved in a number of different ways, for example by determining the edges of the banknote in the banknote sampling frame so as to compute the skewing angle θ and then transforming the data according to the angle θ. The microprocessor 24 may also compute the lengths of the side edges of the banknotes for comparison with the banknote reference data in order to assist in determining the banknote denomination.

At step 5.3, the resulting de-skewed data is compared with the stored reference data for banknotes of different denominations. This may involve a comparison of the length of side edges of the banknote and a comparison of regions bearing key visual features with corresponding stored reference data in the memory 25. If a true banknote is detected at step S 5.4, then at step S5.5 the de-skewed image data developed at step S5.2 is displayed as image 31 on the display device 30. Also, data corresponding to the denomination of the banknote determined at step S5.3 is displayed on the display device 32, at step S5.6.

Also, at step S5.7, the image data that provides the banknote image display 31 is stored in memory 25 together with the data denoting the denomination displayed on display 32.

If the banknote is not accepted at step S5.4, the micro controller 24 causes it to be rejected at step S5.8, by appropriately instructing the gate driver 26 and belt driver 27, either to pass the banknote along reject path 8 or to reverse it back through the inlet 3.

However, for an acceptable banknote, the user that presented the banknote under test to the acceptor 1, is given an opportunity to review the outcome of the acceptance process before agreeing to acceptance of the banknote. Thus, the user can review the image 31 of the acceptable banknote together with its detected denomination as displayed on display 32 and decide whether it corresponds to the user's perception of the banknote and its denomination. Thus, in the example shown in Figure 2, if the user considers that the banknote inserted through inlet 3 was a US$10 banknote, this would correspond to the displayed image 31 and the detected value shown on display 32, in which case, the user would operate the accept button 33. However, if the user believes that a US$50 banknote was presented to the acceptor 1 , the user would not find the displayed image and denomination shown in Figure 2 to be acceptable and would depress the reject button 34.

Referring to Figure 5, if the accept button 33 is actuated, the micro controller 24 operates the gate driver 26 and belt driver 27 to drive the banknote along the accept path 7, to cause acceptance of the banknote as illustrated at step S5.10.

If the reject button 34 is actuated as shown at step S5.l l , the micro controller 24 operates the gate driver 26 and belt driver 27 to cause rejection of the banknote as previously described with reference to step S5.8.

There may be situations where the user wishes to review the images of more than one inserted banknote, for example where more than one banknote is required to provide monetary credit of a purchase value greater than the individual, accepted banknotes. In this situation, the user can operate the retrieve button 35 as shown in Figures 2 and 3 and thereby cause a display of the images and denominations of previously inserted, acceptable banknotes. The micro controller 24 retrieves the image data for successive banknotes in response to successive operations of the retrieve button 35, and displays the data on the display devices 30, 32.

Storage and retrieval of data for rejected banknotes

Data corresponding to each rejected banknote is stored in memory 25 for subsequent analysis. As shown in Figure 3, data corresponding to rejected banknotes can be periodically downloaded to an external processing station illustrated by computer 38 through communication link 39. The processing station 38 may receive data corresponding to rejected banknotes from a number of different banknote acceptors, which may be disposed at remote locations so that a statistical database of such data can be developed. This can be used to improve the stored reference data for acceptable banknotes held in memory 25, which may be updated periodically from the processing station 38 so as to optimise discrimination between true and false banknotes of individual denominations and thereby improve the acceptance rate for true banknotes.

The communication link 39 between the processing station 38 and the banknote acceptor 1 may comprise a local area network for example to communicate data from banknote acceptors in a casino to a central station 38. Also, wide area networks can be used so that data can be received at the station 38 from large numbers of banknote acceptors distributed over a wide geographical area, for example through the internet.

Referring to Figure 5, when a banknote under test is considered false at step S5.4 following the comparison with the stored reference data, the banknote is rejected at step S5.8 and thereafter, at step S5.14, data corresponding to the rejected banknote is stored in memory 25. The process is repeated for banknotes under test that are rejected subsequently, so that a batch of data corresponding to rejected banknotes is built up in memory 25 over time, to be downloaded as a batch to the processing station 38.

Data corresponding to banknotes rejected by operation of reject button 34 may also be stored at step S5.4 and marker data may be included to show that the data relates to rejection caused by operation of button 34. This may assist in the analysis performed at processing station 38.

The image data stored at step S5.14 may comprise all of the de-skewed, pixelated data for the rejected banknote. However, in order to reduce the amount of data to be downloaded, preferably, only data corresponding to pixels which differ significantly from the stored reference data are stored and downloaded. The processing station 38 may have a copy of the stored reference data stored in memory 25 so that by receiving data from the regions that differ from the stored reference data, a full analysis can be performed at processing station 38 for the rejected banknote, by utilising a copy of the stored reference data held in the banknote acceptor 1 and the difference information.

Figure 6 illustrates an example of how pixelated data for the rejected banknote can be processed to identify and store significant differences between the de- skewed pixelated data and the corresponding stored reference data for a true banknote of a particular denomination.

The process starts at step S6.1 and at step S6.2, selected, de-skewed pixels of the rejected banknote are compared in value with corresponding pixels in the stored reference data in memory 25 for an acceptable banknote on a particular denomination. A variance V _x , _y is computed for each, selected pixel of the de- skewed data, where:

V _x , _y = P _x , _y - P _x , _y ref (1)

where: P _x , _y is the value of a selected de-skewed pixel at location x,y in the reference frame F2, and

P _x , _y ref is the corresponding stored value of the pixel at location x,y in reference frame F2 for an acceptable banknote of a particular denomination.

The pixels selected for comparison may be chosen on the basis of the denomination of a banknote and may for example correspond to regions of a

banknote which tend to vary significantly between true and false banknotes. Alternatively, all of the de-skewed pixels can be compared with their corresponding reference values held in memory 25.

Then, at step S6.3, the modulus of the variance V _x , _y is compared with a threshold value T. The modulus can be expressed as | V _x , _y | and thus the following inequality is tested at step S5.3:

I V _x , _y I <T? (2)

If the value of | V _x , _y | exceeds the threshold T, the value of the variance for the pixel concerned is stored in memory 25 at step S6.4. Otherwise, the process returns to check other selected pixels at step 6.5. When all of the selected pixels have been checked, the process concludes at step S6.6. Thus, in this way, only the difference between the values of selected pixels and corresponding reference values from the rejected banknote are stored in memory 25 for subsequent downloading to the processing station 38.

The data stored at step S6.4 may include an associated marker that indicates the particular data set and the denomination for the reference data set of pixels P _x , _y ref, against which the stored variance values V _x , _y are measured, so that the processing station 38 can associate a locally stored copy of the reference data with the variances, after downloading has been completed.

Figure 7 illustrates the process of downloading the stored data for rejected banknotes to the processing station 38 performed under the control of microcontroller 24. The process starts at step S7.1 and at step S7.2, the micro controller 24 periodically checks to determine whether a download request has been received through communication link 39 (Figure 3) from the processing station 38.

When a download request is received, the data corresponding to rejected banknotes is retrieved by micro controller 24 from memory 25 at step S7.3, and is encrypted for secure transmission to the processing station 38. The encryption may be carried out by any suitable encryption technique known in the art and is intended to encompass scrambling, encoding or otherwise security protecting the data by unauthorised persons. In one example public/private key encryption is used.

Then, at step S7.4, the micro controller acts as a transferring means and causes the encrypted data to be transferred to the processing station 38 by downloading the encrypted data over communication link 39. Then at step S7.5, the downloaded data V _x , _y for the rejected banknotes is cleared from memory 25 and the process ceases at step S7.6.

The data received at the processing station 38 from one or more banknote acceptors can be analysed in order to improve upon the reference data that is used by the micro controller 24 of each banknote acceptor to discriminate between true and false banknotes of a particular denomination. Changes to the reference data can then be effected and uploaded from the processing station 38 or by other means to the memory 25 of the, each or selected ones of banknote acceptors to improve its capability to discriminate between true and false banknotes.

Coin acceptor The invention also has application to coin acceptors and an example will now be described. As used herein the term "coin" includes a coin of a denomination of a currency set and also tokens or other such similar generally disc shaped items of value.

Figure 8 illustrates the general configuration of an acceptor according to the invention for use with coins. The coin acceptor is capable of validating coins of a number of different denominations, including bimet coins, for example the euro

coin set and the UK coin set including the bimet £2.00 coin. The acceptor includes a body 40 with a coin run-down path 41 along which coins under test pass edgewise from an inlet 42 through a coin sensing station S and then fall towards a gate 43. A test is performed on each coin as it passes through the sensing station S. If the outcome of the test indicates the presence of a true coin, the gate 43 is opened so that the coin can pass to an accept path 44, but otherwise the gate remains closed and the coin is deflected to a reject path 45. The coin path through the acceptor for a coin 46 is shown schematically by dotted line 47.

The coin sensing station S includes four coin sensing coil units Sl , S2, S3 and S4, which are energised in order to produce an inductive coupling with the coin. Also, a coil unit PS is provided in the accept path 44, downstream of the gate 43, to act as a credit sensor in order to detect whether a coin that was determined to be acceptable, has in fact passed into the accept path 44.

The coils are energised at different frequencies by a drive and interface circuit 48 shown schematically in Figure 9. Eddy currents are induced in the coin under test by the coil units. The different inductive couplings between the four coils and the coin characterise the coin substantially uniquely. The drive and interface circuit 48 produces corresponding digital coin parameter data signals xi, X2, X3, X4, as a function of the different inductive couplings between the coin and the coil units Sl , S2, S3 and S4. A corresponding signal is produced for the coil unit PS. The coils S have a small diameter in relation to the diameter of coins under test in order to detect the inductive characteristics of individual chordal regions of the coin.

In order to determine coin authenticity, the coin parameter signals produced by a coin under test are fed to a microcontroller 49 which is coupled to a memory 50. The microcontroller 49 processes the coin parameter signals xi, - X4 derived from the coin under test and compares the outcome with corresponding stored values held in the memory 50. The stored values are held in terms of windows having

upper and lower value limits. Thus, if the processed data falls within the corresponding windows associated with a true coin of a particular denomination, the coin is indicated to be acceptable, but otherwise is rejected. If acceptable, a signal is provided on line 51 to a drive circuit 52 which operates the gate 43 shown in Figure 8 so as to allow the coin to pass to the accept path 44. Otherwise, the gate 43 is not opened and the coin passes to reject path 45.

The microcontroller 49 compares the parameter signals with a number of different sets of operating window data appropriate for coins of different denominations so that the coin acceptor can accept or reject more than one coin of a particular currency set. If the coin is accepted, its passage along the accept path 44 is detected by the post acceptance credit sensor coil unit PS, and the unit 48 passes corresponding data to the microcontroller 49, which in turn provides an output on line 53 that indicates the amount of monetary credit attributed to the accepted coin.

The sensor coil units S each include one or more inductor coils connected in an individual oscillatory circuit and the coil drive and interface circuit 48 includes a multiplexer to scan outputs from the coil units sequentially, so as to provide data to the microcontroller 49. Each circuit typically oscillates at a frequency in a range of 50-150 kHz and the circuit components are selected so that each sensor coil Sl -S4 has a different natural resonant frequency in order to avoid cross- coupling between them.

As the coin passes the sensor coil unit Sl, its impedance is altered by the presence of the coin over a period of ~100 milliseconds. As a result, the amplitude of the oscillations through the coil is modified over the period that the coin passes and also the oscillation frequency is altered. The variation in amplitude and frequency resulting from the modulation produced by the coin is used to produce the coin parameter signals xi, - X4 representative of characteristics of the coin.

Figure 10 illustrates a bell shaped distribution curve 54 of the values of one of the parameters, xi, produced when a number of coins of the same denomination are passed through the acceptor. It can be seen that most of the occurrences of the parameter value xi occur at a peak value x _p and a generally bell shaped distribution occurs around this peak value. The distribution can be determined by passing a number e.g. 100 coins of the same denomination through the acceptor and recording the corresponding values of xi . Alternatively, the data can be derived externally of the coin acceptor. The memory 50 stores data corresponding to a window of acceptable values of the parameter xi for each denomination of coin to be accepted by the coin acceptor. In Figure 10, one of the windows, referred to herein as a normal acceptance window NAW, is shown, extending between upper and lower window limit values wi, W2. The stored data in memory 50 may comprise the upper and lower window limit values wi, W2 themselves or may comprise a mean value and a standard deviation, such that the microcontroller 49 can define the window NAW from the stored data as a predetermined number of standard deviations about the mean.

The graph of Figure 10 can also be considered in a different way. For coins of the true denomination that correspond to the NAW, the most likely value of parameter xi is the peak value x _p and the least likely value occurs at the upper and lower window limits wi, W2. Whilst it is possible for an acceptable value Xf to occur close to one of the window limits wi, the probability distribution shown in Figure 10 makes it clear that it is unlikely that many such values Xf will occur for the true coin concerned. If several values Xf occur, this is more likely to indicate the presence of a fraudulent distribution 55 as shown in dotted outline, with a peak value centred on or around Xf, which partially overlaps the distribution 54 for the true coin. The fraudulent distribution 55 may for example correspond to the characteristics of a low value coin from a different, foreign currency. Thus, the low value denomination of coin from the foreign currency set may fraudulently used to achieve acceptance by the coin acceptor. The distribution 55 can also correspond to slugs and other fraudulent money items.

Thus, if the coin signal xi for a coin under test has a value falling within the tails of the bell curve of the true distribution 54, there is a significant likelihood that the coin is a fraud rather than a true coin. In our WO 2004/063995 there is described a process in which the micro controller 49 compares the value of the coin signal x with the NAW and also determines if the value falls within an upper safety margin USM or a lower safety margin LSM corresponding to the tails of the distribution 54. If the value of x falls within the USM or LSM, the coin is accepted as a true coin, but there is a significant risk that it is a fraud and that the coin acceptor is subject to an attack by a fraudster. In order to minimise this risk, the microcontroller 49 then uses a restricted acceptance window RAW for one or more coins subsequently presented in order to reject coins with a value of x falling within the USM and LSM regions of the true coin distribution 54. The micro controller 49 reverts to use of normal window NAW once the risk of a fraud attack is deemed to be over.

Storage and retrieval of data for rejected and nearly rejected coins According to the invention, coin parameter signals xi-x _n are stored for rejected coins, for subsequent retrieval and analysis generally in the manner described with reference to the banknote acceptor shown in Figures 1 and 2. Additionally, coin parameter signals x corresponding to coins which fall within the regions

USM LSM of the distribution 54 are also stored for analysis, since these coins are of interest due to the fact that they may represent coins of the fraudulent distribution 55. The storage process performed by microcontroller 49 is illustrated in Figure 11.

The process starts at step SH .1 , and, as illustrated at step SI l .2, for each of the coin parameter signals xi-x _n for a coin under test, their values are compared with a corresponding normal acceptance window NAW at step SH .3. If the values of the coin parameter signals corresponds to the respective NAW for a coin of acceptable denomination, the coin is accepted as illustrated at step SI l .4. Thus, the microprocessor instructs the gate driver 52 shown in Figure 9 to open the gate 43 and allow the coin to pass along accept path 44 shown in Figure 8.

However, if the coin parameter signals xi-x _n fall outside of the respective NAWs for the coin denomination, the coin is rejected at step SH .5. In this case, the gate driver 52 is not operated and the coin is deflected by gate 43 to reject path 45 shown in Figure 8.

When the coin is rejected, the corresponding coin parameter signals xi-x _n are stored at step SH .6 in memory 50 shown in Figure 9 for subsequent retrieval by remote station 38 over communication link 39.

For an acceptable coin, the values of xi-x _n are compared with reference data corresponding to USM and LSM. If the values fall within their respective USM or LSM, as tested at step SI l .7, the values of xi-x _n are stored in memory 50, since they correspond to a coin having coin parameter signals towards the outer limits of the distribution 54. They can be considered as nearly rejected coins. Such coins are of interest to the operator because they may constitute a fraud and are worthy of further data analysis.

However, if the test at step SI l .7 is negative i.e. not within USM or LSM, the process terminates at step Sl 1.8.

The processing station 38 may periodically poll the microcontroller 49 in order to download the stored data for rejected and nearly rejected coins from memory 50 in a similar manner to that described with reference to the banknote acceptor. The stored data can then be analysed at the processing station 38 so that updated values of NAW and IW can be computed and then uploaded to the coin acceptor for storage in memory 50, for use with subsequent coins under test.

As previously explained, data from a plurality of coin acceptors may be downloaded to the processing station 38 for use in the analysis and updating of the coin in acceptance windows.

The downloading of data from the or each coin acceptor and also the uploading of the updated window data may utilise encryption as described previously with reference to Figure 7.

From the foregoing, it will be understood that features of the banknote acceptor data storage process and the corresponding process for the coin acceptor can be used interchangeably in the two described examples.