GB2308004A - Coin recognition apparatus - Google Patents

Abstract

In a coin recognition apparatus a coin 7 entering from one of two directions 1A or 1B comes to rest against accept and/or reject gate members 2A and 2B and is stopped from moving whilst it is being measured preferably by a single sensor. According to the result of the measurement either the accept or reject gate opens to allow passage of the coin. Preferably the sensor 3 is aligned to overlap coins of any diameter with its major axis lying along the locus of the centres of different sized coins or at least aligned to gain maximum coverage of the range of different coin sizes. The sensor may be electromagnetic and take measurements at a variety of frequencies. Further tests which may determine optical characteristics of the coin may subsequently be applied to a coin which is assessed as acceptable.

Description

TITLE Method and Apparatus for Coin Recognition This invention relates to a method and apparatus for coin recognition. Constructions of apparatus are known having a basic arrangement as shown in Figure 1 and wherein a coin is entered in to the mechanism and is brought under dynamic control to roll smoothly down a ramp 1 through one or more sensors 2 and whilst the coin is still rolling a decision is made as to the validity of the coin. If the coin is "good" the accept gate 3 is opened and the coin is taken into the machine and comes out of the Accept Exit 4, otherwise the gate remains closed and the "invalid" coin comes out of the Reject Exit 5 to be returned to the consumer.In the Accept path there may be more sensors to ensure that the coin actually passes through the exit and on into the rest of the machine or into the cash box to prevent fraud.

The sensors are typically electromagnetic, capacitive or optical, Figure 2 illustrates a typical sequence of electromagnetic sensors on the ramp, the sensor shapes and operating frequencies being optimised to detect primarily one characteristic of a coin (e.g.

diameter, thickness or material), although secondary characteristics are increasingly being used to gain extra levels of security in the coin recognition.

One advantage in this method is that, apart from some arrangements used for bringing the coin under dynamic control the progress of the coin is not impeded and the risk of a coin jam is reduced. Additionally, the mechanism can operate correctly at a high throughput rate of several coins per second, and can have opening sections to clear any jams or debris. The tight dynamic timing also make fraud more difficult to effect.

This method of the known art does however lead to a number of disadvantages as follows: (i) Complex timing. The timing of each measurement is critical as the coin is moving through each sensor and there is a very short decision time before the Accept Gate has to be opened which is very critical in a compact mechanism. Each sensor only has time for the capture of one set of data. Also the speeds of coins through the mechanism vary greatly, and the design has to cater correctly for the fastest and the slowest.

(ii) Coin Dynamics - As the measured coin is moving through the sensor the position relative to the sensors on the ramp is critical and if a coin is bouncing or fluttering in any way, measurement performance is degraded.

(iii) Mounting Angle - the mechanism must be mounted to within a tight tolerance otherwise coin speed or dynamics will be outside the mechanisms capture range.

(iv) Cost - since all sensors are separate, and generally operating at the same time as a coin can overlap more than one sensor, parallel circuits are required for each, and a complex multi-tasking function has to be performed.

One of the objects of this invention is to provide a method and apparatus for coin recognition for the "lower security" sector. High security is needed in unattended sites such as street pay telephones, transport ticket machines and in gaming applications where fraudulent credit can be recycled into genuine coins due to a guaranteed percentage payout (70%). High throughput also is only needed in applications where large numbers of coins are expected to be put in quickly, again transport ticket machines and gaming machines primarily.

In the very large number of remaining applications a lower throughput (1 coin per second) and a "standard" rather than "enhanced" level of security is more than acceptable.

In accordance with this invention and for use with a coin-feed apparatus there is provided a method of coin recognition comprising the steps of: a) introducing a coin into the apparatus to travel along an entry ramp, b) bringing the coin to a stop at an accept or reject gate at the end of the ramp, c) determining one or more characteristics of the coin whilst stationary and comparing the value of said characteristic with a stored test value, d) effecting an accept or reject decision based on the comparison, e) operating the accept or reject gate to release the coin according to the decision made.

Preferably a single sensor is used to determine the characteristics of the coin to be tested.

Advantageously the sensor comprises an electromagnetic means operated at a plurality of discrete frequencies either sequentially or simultaneously, stepwise or steplessly.

This invention also provides a coin recognition apparatus comprising a coin entry ramp, stop means at an end of the ramp, an accept gate guarding the entrance to a coin acceptance channel, a reject gate guarding the entrance to a coin reject channel, a sensor to determine one or more characteristics of the coin whilst stationary at the stop means, a comparison means to determine acceptability of said characteristic against a test value held in a store and release means associated with the said gates to selectively operate one or other of the gates in accordance with the comparison.

Further and preferred features of the method and apparatus according to this invention are now described in conjunction with several preferred embodiments and illustrated with reference to the drawings.

The basis of this invention is to stop a coin from moving whilst it is being measured, and to use a single sensor to make the measurement. These aspects are described in the next two sections. A subsequent section describes the use of the method as the post gate sensor to augment the security of a low cost application.

Figure 3 shows an overall mechanical arrangement for the coin acceptor. The coin 7 can enter from one of two directions 1A or 1B and comes to rest in the V formed by the two gate members 2A and 2B. The sensor 3 is aligned to overlap coins of any diameter with its major axis (if non circular) at nominally 450 to the ramp 4 which is the locus of the centres of different sized coins, as illustrated in Figure 4, or at least is aligned to gain maximum coverage of the range of different coin sizes.

In Figure 3, one a series of measurements have been made one or other of the gate members 2A or 2B is opened to let the coin go out of either the accept exit 5 or the reject exit 6. Clearly with different control logic for the gate the accept and reject exits can equally be swapped.

Figure 5 shows a section through the coin ramp of Figure 3. This shows the surface on which the coin rests at an angle of approximately 150 to the vertical to ideally ensure that the coin at rest is always at the same distance from the sensor or sensors 3A and 3B. The ramp angle in Figure 3 needs to be at least 200 to the horizontal to ensure that when gate 2A is opened a faceted coin (e.g. UK 50 pence) will start rolling under gravity. Ideally this angle can be much greater for faster operating speed, the greater tolerance to the.

overall mechanism mounting.

The sensors coils used can be either one or more ferrite pot cores (circular or shaped), mounted on either side of the coin to be measured as shown in Figure 6, or an enclosing coil without ferrite core as shown in Figure 7. (Note that the inner dimensions of such a sensor would need a longer dimension of J2 times the maximum diameter coin to enable it to pass through.) The gating function may be operated by for example two solenoids (one operating gate 2A and the other operating gate 2B), or by a three position mechanical drive using for example a motor with both gates coupled.

The three positions being: (i) Both gates closed.

(ii) Gate 2A open, Gage 2B closed.

(iii) Gate 2B open, Gate 2A closed.

To extract the maximum information from a single sensor, this invention operates the sensor at multiple frequencies, sensing the change in inductance in the coil caused by the coin at these multiple frequencies. At frequencies between 100 KHz and 1MHz there are different inductance changes depending on the particular mechanical construction of the coin, viz its diameter, thickness, bulk material, surface material and embossing. This gives the basis of the means of discriminating between different types of coins and slugs or foreign coins.

Figure 9 shows a block diagram of an example measurement system, which uses an oscillator which includes the sensor coil 2. The frequency at which the oscillator resonates is controlled by the microprocessor.

At the selected normal resonant frequency the frequency and amplitude of the oscillator is measured with and without the coin, this being repeated for multiple frequencies, gathering a profile of the inductance with and without the coin versus frequency.

Figures 10 and 11 show examples of the frequency and amplitude values that might occur for two types of coin in an arrangement using 6 separate oscillator steps.

Figure 12 shows an example of how switched capacitors can be used to provide the oscillator steps in a Colpitts transistor oscillator, with the switches controlled by the microprocessor. The switches themselves are bipolar or field effect transistors or alternatively relays.

To determine that a coin has entered the measurement system, the oscillator is operated at a single frequency periodically to look for a change in frequency. This need only be carried out infrequently to conserve power, and could be carried out more often when a coin was most likely to be inserted. When such a frequency change is detected, and the measurement is stable (indicating that the coin is no longer bouncing) the measurement sequence is commenced.

The measurement consists of measuring frequency and amplitude of the oscillator at a number of oscillator steps n as described above. This gives an array of measurements F1 - Fn for the frequencies, and A1 - An for the amplitudes. A further array of "idle" measurements when a coin is not present are periodically taken between coins, and updated as FI1 - Fin and AI1 - Al n respectively.

Differences or ratios are taken between the measurements (differences typically for frequencies, and ratios for amplitudes, though not exclusively) to compensate for temperature and voltage variations to the circuitry, and normal production spreads. These are calculated as a further array of results FR1 - FRn and AR1 and AR1 and ARn: Difference: F - FI or A - Al Ratio: (Fx - FIX)/FIx or (Ax - AIX)/AIx The mechanism would have programming for each coin type to be accepted the array of results expected for an ideal or mean coin CF1 - CFn and CA1 - CAn.The array of results FR1 - FRn and AR1 - ARn would be compared against each other stored ideals to determine which, if any, of the programmed coins the measurements related to. This decision being made by either checking the results on a one by one basis against a range of acceptable values for each programmed parameter for each coin (as, for example, +ve or -ve acceptable values, or within a certain percentage of the measurement).

Upper and lower "Limit" values can also be used to define the range of acceptable values for a parameter, and in many cases the distribution is assumed to be a normal statistical distribution and the limits are derived from the mean + multiples of the standard deviation (+ standard deviations as shown in Figure 14.) In many cases however the distribution of measurements is asymmetrical and a pair of 3 standard deviation limits may not give the optimum performance, as shown in Figure 15. This figure also shows limits calculated from a further aspect of this invention which treats the distributions above and below the mean value separately and derives standard deviations for each half to generate limits better matching the distribution. Figure 16 shows the steps involved in this process.Firstly the data is split in half about the mean and then an artificial set of "mirrored" data is added to each part to produce two symmetrical distributions for which the standard deviations are calculated.

An alternative approach to the decision mechanism is to calculate a normalised difference (taking the modulus to give only positive differences) of each element in turn of the results array from the "ideal" array. The normalisation would be by multiplication with a further constant array NF1 - NFn and NA1 - NAn.

The normalised difference would be calculated as: For frequencies: FNDx = I (FRx - CFx) I N x For amplitudes: ANDx = I (ARx - CAx) I x NAx If any normalised difference were above a given preset number it would be determined that it was NOT that particular coin.For the remaining stored coin "ideals" the total difference would be calculated as: Total Difference = E FND x AND x x=1,n x=1,n The type of coin inserted being determined as the one with the lowest total difference The "ideals"/normalisation constants or limits are either factory programmed or "learnt" from a sample number of the coin to be programmed, and can also be made to change adaptively in the field, should the coin population change with time, due to the effects of ageing on the circuitry, or due to natural wear.

In practice, using an oscillator with six steps, not all frequency and amplitude measurements are used for discrimination, more typically four frequency measurements and two amplitude measurements provide sufficient performance. Also a mixture of the decision algorithms may be used.

A potentially powerful application of the technology described in the preceding sections relates to its use as a second level of measurement. Figure 13 outlines such an arrangement. This figure is an extension of the diagram shown in Figure 1, with a location for a stationary coin 9 after the main Accept Gate 3. An "enclosing" type of sensor 6 is shown in this example to measure the stationary coin, and gates 2 and 3 (7) and (8) are used firstly to locate the stationary coin, and secondly to either permit that coin to be accepted or rejected respectively, when they are moved. In practice gates 2 and 3 would generally be coupled and operated in a three way operation as described in section 3.

On coin entry the presence thereof is detected by the first sensors 2 and a preliminary decision taken as to its validity. If the coin is thought to be valid the first accept gate 3 is opened and the coin passes into position 9. As soon as the coin is stable in position 9 a series of measurements are made with the post gate sensor 6 and a further decision as to the coin's validity is made. If the coin is determined to be valid, gate 2 (7) is opened and the coin drops out of the Accept Exit 4. If the coin is found to be invalid gate 3 (8) is opened and the coin travels down an additional ramp to the reject exit 5.

The operation of this arrangement can be varied by only applying the second sensor to those coins that are for example high value, or are only marginally valid" from the first sensors. (Marginally valid coins would be those whose measurements are near the edges of the first sensor's measurement distributions, or have a Total Difference" from the first sensor beyond a certain level indicating that the probability of the measurements indicating a genuine coin is lower than acceptable).

This has the advantage when secondary measurements are not used in that the sensor 6 reverts to a traditional post gate sensor" and is used only to detect the correct passage of the coin, and the second accept gate 7 can be opened at the same time as the first gate 3 for faster throughput of coins. Alternatively, the secondary measurement can be applied only in situations when higher fraud protection is required, at the discretion of the machine operator.

With such an arrangement using two sensors the.

second sensor may use different algorithms to the first sensor and depending upon the results of the first sensor.

Since the coin is held stationary whilst it is being measured, the location could also be used as a single coin store or "escrow" without any additional components.

This is advantageous in applications such as pay telephones where a "pre-pay" function is required. In this case a coin is inserted into the mechanism and is checked to be valid before a phone call can be made, and the coin is "accepted" into the cash box when a connection is made or when the first metering pulse is received. If no connection is made the coin is returned to the consumer. Clearly further coin stores can be cascaded on the accept route for multiple coin capacity.

To prevent coin jams in the sensing area, or more than one coin being accepted at a time (one possibly being "stolen"), a blocking peg as shown schematically in Figure 8 could be used to segregate one coin from the next. The blocking peg could be solenoid or motor operated, or mechanically linked to the presence of a coin in the measurement area. Additionally, the ramp area could be opening, in response to a reject lever, as in state of the art coin acceptors to clear any jams or debris. If an "enclosing sensor is used this cannot be opened but could have a central gap wide enough to ensure overlapped coins do not jam, and the ramp "throat" could be pinched in advance of the sensor to catch jams and enable them to clear satisfactorily.

It is additionally feasible to use the single sensor arrangement at multiple frequencies in some types of low cost application with a moving coin, although the number of frequencies that can be sampled would be limited, and in phase transitions from frequency to frequency would be required.

The preferred coin acceptance arrangement described offers a number of benefits over other methods (i) Lower cost - through single sensor, single oscillator and measurement interface, and simple mechanical construction possibilities, without the need for careful dynamic control of coins.

(ii) Wide angle of operation. The coin ramp could be anything from 200 to nearly 900 from the horizontal, and rotated + 100 at least, compared with + 3 or 40 for arrangements measuring moving coins.

(iii) Simple timing and sequencing requiring less code space (ROM) in the microprocessor.

(iv) Very repeatable measurements.

(v) A powerful but low cost, adaptive secondary measurement option to significantly enhance the performance of existing technologies, particularly when combined with "capacitive" primary sensors.

In a preferred construction the following features alone or in combination characterise the invention, (i) Single coin sensor aligned on a stationary coin for measurement.

(ii) Sensor being either an enclosing sensor or conventional coil/pot core arrangement.

(iii) Non Circular sensor in above arrangement.

(iv) Sensor aligned along major locus of coin centres for a range of diameters.

(v) Variable frequency operation of the sensor.

(vi) Switched oscillator for the variable frequency operation.

(vii) Measurement of inductance with and without coin in this arrangement.

(viii) Measurement of frequency change in oscillator in this arrangement.

(ix) Measurement of amplitude change in oscillator in this arrangement.

(x) Use of amplitude and frequency changes over wide range of basic frequencies of a single sensor as a coin discriminator.

(xi) Decision algorithms as described.

(xii) Programming methods as described, including the methods for handling asymmetric distributions of measurements.

(xiii) The described measurement method being used as a second "post gate" measurement, with and without adaptive methods for its use as a secondary measurement.

(xiv) The described arrangement with a "self-learn" function.

(xiv) The basic principle of using one measurement (or series of measurements), making a preliminary decision based upon its results, and then making a second measurement (or series of measurements) once the coin has passed a first "accept gate" leading to a second decision as to the validity of the coin.

A modification and alternative method of recognising a stationary coin and in accordance with another aspect of this invention is described and illustrated with reference to Figures 17 to 20 of the drawings.

In this arrangement an optical sensor is used instead of, or in combination with, an electromagnetic sensor to recognise a stationary coin.

As shown in Figure 17, such a sensor may be a simple linear or two dimensional optical array detecting the outline dimensions of the coin through the occlusion of light falling on the sensor array by the stationary coin.

Such a sensor can be a complete image sensor, such as a CCD (Charge Coupled Device) camera sensor as used in Cam-corders or security cameras, or a cheaper CMOS image sensor see Figure 18. In this case a complete picture of the coin is taken and digitised into a memory. The control microprocessor then applies standard image recognition techniques to validate the coin. The advantage of this method is that the coin is firstly stationary permitting the lighting to be optimised to give an adequate contrast level in the digital image, and there is a fixed reference point for the coin enabling the major dimensions to be rapidly established and the centre calculated. Once the centre is known the image can be rotated about this point to recognise the embossing design on the coin. Figure 19 gives the flow chart of events and checks as to the validity of the coin in this arrangement, whilst Figure 20 is a block diagram of the measurement system.

Claims (16)

1. A method of coin recognition for use in a coin-feed apparatus, the method comprising the steps of:a) introducing a coin into the apparatus to travel along an entry ramp, b) bringing the coin to a stop at an accept or reject gate at the end of the ramp, c) determining one or more characteristics of the coin whilst stationary and comparing the value of said characteristic with a stored test value, d) effecting an accept or reject decision based on the comparison, e) operating the accept or reject gate to release the coin according to the decision made.

2. A method in accordance with Claim 1, wherein a single sensor is used to determine the characteristics of the coin to be tested.

3. A method in accordance with Claim 2, wherein the sensor comprises an electromagnetic means operated at a plurality of discrete frequencies either sequentially or simultaneously, stepwise or steplessly.

4. A coin recognition apparatus comprising a coin entry ramp, stop means at an end of the ramp, an accept gate guarding the entrance to a coin acceptance channel, a reject gate guarding the entrance to a coin reject channel, a sensor to determine one or more characteristics of the coin whilst stationary at the stop means, a comparison means to determine acceptability of said characteristic against a test value held in a store and release means associated with the said gates to selectively operate one or other of the gates in accordance with the comparison.

5. Method or apparatus according to any preceding claim, using a single coin sensor aligned on a stationary coin for measurement.

6. Method or apparatus according to any preceding claim, wherein the sensor is either an enclosing sensor or conventional coil/pot core arrangement.

7. Method or apparatus according to any preceding claim, wherein the sensor is aligned along major locus of coin centres for a range of diameters.

8. Method or apparatus according to any preceding claim, using variable frequency operation of the sensor.

9. Method or apparatus according to any preceding claim, wherein a sensor is used to effect one measurement (or series of measurements), making a preliminary decision based upon the result, and then making a second measurement (or series of measurements) once a coin has passed a first accept gate leading to a second decision as to the validity of the coin.

10. Method or apparatus in accordance with any preceding claim, wherein the sensor, or a further sensor, detects optical characteristics of the coin.

11. Method or apparatus in accordance with Claim 10, wherein the sensor detects optical characteristics of either a portion of the face of a coin or the whole face of the coin, the output from the sensor being preferably digitised and stored to be compared with a pre-stored image, to effect validation.

12. A method of coin recognition for use in a coin-feed apparatus substantially as described in the specification and as illustrated by one or more of the figures of drawings.

13. A coin recognition apparatus constructed and arranged to function as disclosed and described herein with reference to one or more of the figures of drawings.

Amendments to the claims have been filed as follows 1. A method of coin recognition for use in a coin-feed apparatus, the method comprising the steps of a) introducing a coin into the apparatus to travel along an entry ramp, b) bringing the coin to a stop at an accept or reject gate at the end of the ramp, c) determining one or more characteristics of the coin whilst stationary using a sensor aligned along the major locus of coin centres for a range of diameters, d) comparing said characteristic with a stored test value, e) effecting an accept or reject decision based on the comparison, f) operating the accept or reject gate to release the coin according to the decision made.

2. A method of coin recognition for use in a coin-feed apparatus, the method comprising the steps of a) introducing a coin into the apparatus to travel along an entry ramp, b) using a sensor to effect one measurement (or a series of measurements), making a preliminary decision as to the validity of the coin, and operating a first accept or reject gate in accordance with the decision, c) bringing the coin to a stop at a second accept or reject gate, d) determining one or more characteristics of the coin whilst stationary and comparing said characteristic with a stored test value e) effecting an accept or reject decision based on the comparison, f) operating the second accept or reject gate to release the coin according to the decision made.

3. Method or apparatus according to Claim 2, wherein the sensor measuring a stationary coin is aligned along the major locus of coin centres for a range of diameters.

4. Method or apparatus according to any preceding claim, wherein a single sensor is used to determine the characteristics of a stationary coin to be tested 5. Method or apparatus according to any preceding claim, wherein the sensor comprises an electromagnetic means operated at a plurality of frequencies either sequentially or simultaneously, stepwise or steplessly.

6. Method or apparatus according to any preceding claim, wherein the sensor is an enclosing sensor.

7. Method or apparatus according to any preceding claim, wherein a mechanical means is used to separate the measured coin from a subsequent coin.

8. Method or apparatus according to any preceding claim, wherein the location used for measurement is additionally used as a temporary escrow position for the coin.

9. A method or apparatus in accordance with any preceding claim, wherein the sensor, or a further sensor, detects optical characteristics of the coin.

10. Method or apparatus in accordance with claim 8, wherein the sensor detects optical characteristics of either a portion of the face of a coin or the whole face of the coin, the output from the sensor being preferably digitised and stored to be compared with a pre-stored image, to effect validation.

11. A method in accordance with Claim 9 or Claim 10 wherein the sensor comprises a Charge Coupled Device.

12. A method of calculating the acceptance limits for a coin measurement parameter which treats the distributions above and below the mean value of a plurality of measurements separately mirroring each partial distribution.

13. A method in accordance with Claim 12 wherein separate statistical calculations are applied to each partial distribution.

14. A method in accordance with Claim 13, wherein the statistical calculations applied are based on a Normal Statistical Distribution.

15. A method of coin recognition for use in a coin-feed apparatus substantially as described in the specification and as illustrated by one or more of the figures or drawings.

16. A coin recognition apparatus constructed and arranged to function as disclosed and described herein with reference to one or more of the figures or drawings.