[0001] This invention relates to a method and an apparatus for validating coins.
[0002] It is known that a substantial number of counterfeit coins have electrical and magnetic
properties which resemble those of genuine coins, so that coin validators erroneously
indicate them to be genuine, but have different mechanical properties. In particular,
many counterfeit coins are softer than the genuine coins which they otherwise resemble.
It has been proposed to distinguish between such counterfeit coins and genuine coins
by detecting the vibrations caused after an impact between the tested item and an
element of the coin validator. Piezoelectric elements have been used for sensing the
vibration. See for example GB-A-2 236 609 and EP-A-543 212.
[0003] It would be desirable to provide a more reliable method for testing for such counterfeits,
thereby providing better discrimination, and an apparatus which employs such a method.
[0004] Various aspects of the invention are set out in the accompanying claims.
[0005] According to another aspect, a coin validator determines the shape of at least the
initial part of a vibration caused by an impact of the coin being tested, the coin
validator then taking the determined shape into account in producing a signal indicating
whether or not the tested coin is genuine. The validator can be arranged to indicate
that the coin is not genuine unless the shape is determined to be appropriate. Alternatively,
the closeness of the determined shape to an appropriate shape can be used as one of
a number of factors taken into account in determining whether the coin is genuine,
and possibly the coin denomination.
[0006] Preferably, the vibration produced by the coin impact is sensed by a piezoelectric
element, and the output signal is processed to determine the shape of the initial
part of the vibration. It has been found that this shape is characteristic of material
properties of the coin being tested, and particularly the hardness. Depending upon
the structure of the coin validator and the manner in which the impact is produced
and sensed, the later part of the signal may be less representative of the coin properties
and therefore it is preferable to disregard this part of the signal. For example,
in particular embodiments, it has been found that although the initial part of the
vibration contains information indicative of the coin hardness, the later parts of
the vibration are dominated by the mechanical characteristics of the validator. Preferably,
the shape of the vibration waveform is determined on the basis of the vibration during
the first millisecond after the impact, and more preferably during the first quarter
of a millisecond.
[0007] The data used to determine the shape of the vibration waveform can be derived in
a number of ways. In the preferred embodiment, the waveform amplitude is sampled at
predetermined intervals. Preferably, the sampling commences when the amplitude reaches
a predetermined threshold level.
[0008] In an altemative embodiment, the waveform is monitored to determine the times at
which predetermined amplitudes are reached. This, however, is less preferable because
it has been found that the range of amplitudes varies substantially from impact to
impact, so that allowing for the largest range of amplitudes would result in a loss
of resolution.
[0009] In a further alternative embodiment, the vibration waveform is processed to determine
the time and amplitude at which certain events occur, for example the times and amplitudes
of the peaks and troughs in the vibration waveform. However, this could also suffer
from dynamic range problems, and requires additional processing.
[0010] Various techniques can be used to determine the shape of the vibration waveform using
the derived data samples. In the preferred embodiment, the samples are weighted and
summed, and preferably applied to a non-linear function. This could be performed a
number of times, with the outputs of the non-linear functions also being combined
in a weighted manner. To derive the weighting factors, a neural network can be trained
in a per se known manner, e.g. using back propagation.
[0011] While neural networks provide a rapid method of generating an algorithm to process
the data, algorithms could obviously be developed by other methods to provide discrimination
between numerical representations of the waveforms. Analysis would lead to an understanding
of the relationships between the waveform and the known form of the coin giving rise
to the signal. The waveform data could be analysed to discover deeper interrelationships.
Non linearities might be accommodated by use of power laws, logarithms, trigonometrical
or other functions. Regression techniques could be employed, for example, with polynomials
to develop a model which ultimately discriminates the waveforms. These approaches
would work, but use of a neural network is preferred because it leads to a fast and
sufficiently effective result which is simple to incorporate in a product.
[0012] An arrangement embodying the invention will now be described by way of example with
reference to the accompanying drawings, in which:
Figure 1 schematically shows a coin validator in accordance with the invention;
Figure 2 is a diagram illustrating the sampling of a vibration waveform; and
Figure 3 is a diagram illustrating the manner in which the data samples representing
the waveform are processed.
[0013] Referring to Figure 1, the validator 2 comprises a test structure 4. This structure
comprises a deck (not shown) and a lid 6 which is hingedly mounted to the deck such
that the inner sides of the deck and lid are in proximity to and face each other.
Figure 1 shows the test structure 4 as though viewed from the outer side of the lid.
The inner side of the lid is moulded so as to form, with the deck, a narrow passageway
for coins to travel edge first in the direction of arrows A.
[0014] The moulded inner surface of the lid 6 includes a ramp 8 along which the coins roll
as they are being tested. At the upper end of the ramp 8 is an energy-absorbing element
10 positioned so that coins received for testing fall on to it. The element 10 is
preferably made of material which is harder than any of the coins intended to be tested,
and serves to remove a large amount of kinetic energy from the coin as the coin hits
the element. The energy-absorbing element may be structured and mounted as shown in
EP-A-466 791.
[0015] A wall 12 is formed on the outer side of the lid 6, and a piezoelectric element 14
is mounted on this wall. The vibrations caused by the coin impacting the energy-absorbing
element 10 pass through the test structure 4, and in particular are carried by the
lid 6 and the wall 12 to the piezoelectric element 14, which generates a signal on
an output line 16.
[0016] As the coin rolls down the ramp 8, it passes between three coils 18 mounted on the
lid, and a corresponding set of coils (not shown) of similar configuration and position
mounted on the deck, forming three pairs of coils, the coils of each pair facing each
other across the coin passageway. The coin is subjected to electromagnetic testing
using these coils 18.
[0017] The coils 18 are connected via lines 20 to an interface circuit 22, which also receives
signals from the piezoelectric element 14 via output line 16.
[0018] The interface circuit 22 comprises oscillators for driving the electromagnetic coils
18, circuits for appropriately filtering and shaping the signals from lines 16 and
20 and a multiplexing circuit for delivering any one of the signals from the piezoelectric
element 14 and the pairs of coils 18 to an analog-to-digital converter 24.
[0019] A control circuit 26, including a microprocessor, has an output line 28 connected
to the analog-to-digital converter 24, and is able to send pulses over the output
line 28 in order to cause the analog-to-digital converter 24 to take a sample of its
input signal and provide the corresponding digital output value on a data bus 30.
[0020] In this way, the control circuit 26 can obtain digital samples from the test structure
4, and in particular from the piezoelectric element 14 and the coils 18, and can process
these digital values in order to determine whether a received test item is a genuine
coin or not. If the coin is not determined to be genuine, an accept/reject gate 32
will remain closed, so that the coin will be sent along the direction B to a reject
path. However, if the coin is determined to be genuine, the control circuit 26 supplies
an accept pulse on line 34 which causes the gate 32 to open so that the accepted coin
will fall in the direction of arrow C to a coin separator (not shown), which separates
coins of different denominations into different paths and directs them to respective
coin stores (not shown).
[0021] In this embodiment, a single analog-to-digital converter 24 is used in a time-sharing
manner for processing the signals from the coils 18 and from the piezoelectric element
14. However, a plurality of converters could be provided if desired.
[0022] The output from the piezoelectric element 14 is processed in the manner described
below in order to determine whether the received test item is of relatively soft material,
indicating that it is a counterfeit. If so, the control circuits 26 will reject the
test item irrespective of the signals provided by the coils 18.
[0023] Referring to Figure 2, this shows an exemplary vibration waveform produced by the
piezoelectric element 14 on output line 16 following the impact of the test item with
the energy-absorbing element 10. The control circuit 26 repeatedly checks the output
of the analog-to-digital converter until the amplitude of the signal on line 16 reaches
a predetermined threshold T (of, for example, 200 millivolts). (If desired, a hardware
comparator can be provided for this purpose, the comparator providing a signal to
the control circuit 26 when the threshold is reached.) A timer is then started. Subsequently,
the timer causes the control circuit 26 to take five samples X1 to X5 of the output
signal at 35 microsecond intervals.
[0024] Referring to Figure 3, these data samples are then processed as illustrated schematically
here. A first process, schematically illustrated by the neuron 300, takes all five
values and multiplies each one by a respective predetermined weight and then sums
them with a bias value B1. The sum is then applied to a non-linear function, for example
a sigmoid function or a hyperbolic tangent function, to provide an output value P1.
[0025] A second process illustrated by neuron 302 performs a similar operation, except using
different weights and a different bias value B2, to produce an output value P2.
[0026] A third process, illustrated by a summing junction 304, multiplies each of the output
values P1 and P2 by a respective weight and adds these to a bias value B3 to produce
an output value O.
[0027] This output value is dependent on the shape of the initial part of the waveform shown
in Figure 2, which in turn is influenced by the hardness of the test item. The weights
and the bias values are so chosen that the control circuit 26 can determine whether
the test item is relatively soft, indicating a counterfeit coin, by determining whether
or not the output value O exceeds a predetermined threshold. Accordingly, the output
value O is compared with this threshold in order to produce a yes/no output.
[0028] In an alternative embodiment, the output value O is compared to a range of values,
and the processor determines that the test item is a counterfeit in dependence upon
whether or not the value lies within the range. This would for example be useful if
there are counterfeit coins which have a hardness greater than that of genuine coins.
[0029] Because the piezoelectric element 14 is situated upstream of the coils 18, the processing
of the output signal from the element can occur before the output signals from the
coils need to be processed. If the output of the piezoelectric element indicates that
a counterfeit coin has been received, the processing of the output signals from the
coils can be omitted.
[0030] In an alternative embodiment, the signal from the piezoelectric element 14 is used
(preferably together with the signals from the coils 18) to determine the denomination
of a genuine coin. Thus, the validator can be arranged to store acceptance criteria
for each of the denominations it is intended to validate. For each denomination, there
may be stored criteria determining the type of signals expected to be received from
the coils when testing a coin of that denomination. In addition, in accordance with
this embodiment, there can be acceptance criteria for the value O, which criteria
would vary according to denomination. Thus, the value O could be compared with a plurality
of ranges, each associated with a different denomination. Also or alternatively, if
desired, in order to determine whether the tested item corresponds to any of a number
of different coin denominations, it is possible to use different sets of weights and
bias values, each set being used to determine whether the item corresponds to a respective
denomination.
[0031] Instead of using the output value O to indicate whether or not a particular coin
has been tested, it could be used as one of a number of discriminants which are considered
in combination to evaluate the tested item, for example using the techniques of EP-A-0
496 754.
[0032] The weights and the bias values used in the processing illustrated in Figure 3 can
be derived using an iterative training process. Conventional neural network techniques,
such as back propagation, can be used. Samples of genuine and counterfeit coins would
be repeatedly tested, while the weights and bias values are modified to enhance the
discrimination between them, and, if desired, between coins of different denominations.
This operation can be performed after assembly of the coin validator using a training
procedure on each individual validator. Preferably, however, the training procedure
uses data from a plurality of reference validators, whereby the derived weights and
biases will not be validator-specific, so that common values can thereafter be used
in production of new validators and it is not necessary to determine individual weights
and biases for each production validator. The output values O may however be different
for different validators, in which case individual calibration of the validators may
be performed by insertion of genuine coins to derive suitable acceptance criteria
for the values O.
[0033] The processing illustrated in Figure 3 can be varied considerably. The neurons 300
and 302 represent a hidden layer. If desired, there could be additional neurons in
this layer, or one or more additional layers, or the layer can be omitted. The non-linear
functions performed by these neurons can be omitted, or a further non-linear function
can be added to the neuron 304. Instead of combining the weighted samples before applying
the sum to a non-linear function, non-linear functions can be applied to the samples
prior to combining them. Instead of using simple weighting and summing operations,
other techniques can be used for processing and combining the individual values. It
is, however, preferable for there to be at least three values, and preferably five
or more values, representing different amplitude/time points on the waveform, to provide
at least an approximate indication of waveform shape. In the embodiment described
above there are effectively six values, representing respective points in amplitude/time
space, because the samples X1 to X5 all represent amplitudes at particular intervals
after a starting point when the amplitude was at a known value. The starting point
therefore represents a further known point in amplitude/time space. An alternative
embodiment might use only the starting point and two subsequent samples.
[0034] Alternative embodiments, for example one which use asynchronous sampling, may be
such that the samples do not bear a consistent relationship to a known starting point,
and in these embodiments it is desirable for at least three samples to be used.
[0035] Instead of a piezoelectric element, any other form of microphone could be used.
[0036] In the embodiment described above, the mechanical properties of the coin, and particularly
its hardness, are tested by examining the results of a coin impact. However, similar
techniques could be used for examining the output of a sensor responsive to different
characteristics of the coin. For example, an electromagnetic sensor, such as formed
by a pair of the coils 18 referred to above, produces a time varying signal as the
coin passes it. The signal could for example represent the frequency or the amplitude
of the sensor output. (Depending on the nature of the signal, it may not be necessary
for analog-to-digital conversion. For example, a counter may be used to measure frequency,
in which case the output is already digital.)
[0037] When applying the invention to such sensors, the variations in the signal correspond
to variations in the position of the coin, instead of the nature of the vibrations
caused by the coin impact. Nevertheless, these variations are characteristic of the
coin, and it would therefore be possible to validate the coin using a technique similar
to that described above, by analysing samples of the waveform to provide a value indicative
of the waveform's shape. Variations in coin speed would cause the shape to expand
or contract along the time axis, but this could be dealt with either by sensing coin
movement and taking this into account in the processing, or by training the neural
network such that speed variations have little effect on the results.
[0038] Although the embodiment described above processes digital samples using a microprocessor,
this is not essential. For example, an analog sensor output could be fed to sequentially-triggered
sample and hold circuits, so as to derive a plurality of analog samples which are
fed to a hardware neural network.
[0039] The invention has been described in the context of coin validators, but it is to
be noted that the term "coin" is employed to mean any coin (whether valid or counterfeit),
token, slug, washer, or other metallic object or item, and especially any metallic
object or item which could be utilised by an individual in an attempt to operate a
coin-operated device or system. A "valid coin" is considered to be an authentic coin,
token, or the like, and especially an authentic coin of a monetary system or systems
in which or with which a coin-operated device or system is intended to operate and
of a denomination which such coin-operated device or system is intended selectively
to receive and to treat as an item of value.
1. A method of validating coins in which a sensor responds to a coin impact by producing
a time-varying signal having characteristics dependent on those of the impact, the
method including the step of deriving from the signal data representing points in
time/amplitude space, and combining the data in a weighted manner to produce an output
indicative of coin validity.
2. A method of validating coins in which a sensor responds to a coin impact by producing
a time-varying signal having characteristics dependent on those of the impact, the
method including the step of deriving from the signal data indicative of three or
more points in time/amplitude space, and combining the data to produce an output indicative
of coin validity.
3. A method of validating coins in a coin validator including a sensor which produces
a time-varying signal in response to an impact of a coin received by the validator,
the method comprising the step of sampling the amplitude of the signal at predetermined
intervals and combining the sampled data in a predetermined manner to produce an output
indicative of the validity of the coin.
4. A method as claimed in claim 3, including the step of determining when the amplitude
crosses a predetermined threshold, and commencing a sampling process at that time.
5. A method as claimed in claim 2, 3 or 4, in which the data are combined in a weighted
manner.
6. A method as claimed in any preceding claim, including the step of applying a non-linear
function to the data.
7. A method as claimed in any preceding claim, including causing the coin to impact an
energy-absorbing element in order to produce the coin impact.
8. A method as claimed in any preceding claim, including the step of taking further measurements
of the coin in order to determine the validity and denomination of the coin.
9. A method as claimed in claim 8, wherein the further measurements are taken after the
coin impact has been produced.
10. A method as claimed in claim 8 or claim 9, wherein said output is also used to determine
coin denomination.
11. A method of validating coins in a coin validator including a sensor which produces
an output having a waveform representing variations in a sensed parameter in response
to the sensing of a coin, the method including the step of deriving data from points
on the waveform such that for each point the value of the parameter and the position
along the waveform at which the value was established are known, and combining the
data in a weighted manner to produce an output indicative of coin validity.
12. A method of validating coins in a coin validator including a sensor which produces
an output having a waveform representing variations in a sensed parameter in response
to the sensing of a coin, the method including the step of deriving data from at least
three points on the waveform such that for each point the value of the parameter and
the position along the waveform at which the value was established are known, and
combining the data to produce an output indicative of coin validity.
13. A method of validating coins in a coin validator including a sensor which produces
an output having a waveform representing variations in a sensed parameter in response
to the sensing of a coin, the method including the step of sampling the waveform at
predetermined intervals to derive sample data representing points on the waveform
such that for each point the value of the parameter and the time at which the value
was established are known, and combining the data to produce an output indicative
of coin validity.
14. A method as claimed in claim 13, including the step of determining when the amplitude
crosses a predetermined threshold, and commencing a sampling process at that time.
15. A method as claimed in claim 12, 13 or 14, in which the data are combined in a weighted
manner.
16. A method as claimed in any one of claims 11 to 15, including the step of applying
a non-linear function to the data.
17. A method as claimed in claim 1, 5, 11 or 15, or any claim dependent on claim 1, 5,
11 or 15, wherein the weights have been derived by an iterative training process involving
the testing of genuine and counterfeit coins, such that the combined data is determined
by the shape of the signal and thus is characteristic of the coin.
18. A method as claimed in claim 17, wherein the weights have been derived by an iterative
training process using a plurality of reference validators so that common weights
can be used for production validators.
19. A coin validator arranged to operate in accordance with a method of any preceding
claim.
20. A method of setting up a production coin validator, the method comprising:
(a) deriving, by sensing a coin using a reference validator, an output waveform representing
variations in a sensed parameter;
(b) deriving data from points on the waveform such that for each point the value of
the parameter and the position along the waveform at which the value was established
are known;
(c) combining the data in a weighted manner;
(d) repeating steps (a) to (c) using genuine and counterfeit coins while adjusting
the weights;
(e) repeating steps (a) to (d) using one or more further reference validators in order
to derive common weights for the reference validators which result in the combined
data discriminating between genuine and counterfeit coins; and
(f) storing data representing the weights in the production validator for use in performing
validation operations.
21. A method as claimed in claim 20, including the further step of individually calibrating
the production validator by using it to sense genuine coins and using the sensor output
to derive a suitable acceptance criterion for the combined data.
22. A method of setting up a coin validator, the method comprising the steps of:
(a) sensing a coin using a sensor of the validator and deriving therefrom an output
waveform representing variations in a sensed parameter;
(b) deriving data from points on the waveform such that for each point the value of
the parameter and the position along the waveform at which the value was established
are known;
(c) combining the data in a weighted manner;
(d) repeating steps (a) to (c) using genuine and counterfeit coins, while adjusting
the weighting in order to enhance the discrimination between genuine and counterfeit
coins; and
(e) storing data representing the weighting in the validator for use in performing
validation operations.
23. A method as claimed in claim 20, 21 or 22, wherein the waveform is a time-varying
signal produced in response to a coin impact.