Fraud detection system and method

Abstract

The present invention relates to a fraud detection system (10) for a recycling machine (12), the system comprising: a detector (14) adapted to detect at least one container (18) entering the recycling machine; a compactor load sensor (28) adapted to measure the load of a compactor (24) of the recycling machine during operation, wherein the compactor is adapted to compact the entering container downstream of the detector; and a compactor load sensor monitoring device (26) configured to determine, based on the load measured by the compactor load sensor, whether the detected at least one container is compacted as expected.

Description

Technical Field

[0001] This invention relates to a fraud detection system and method for recycling machines. The invention also relates to a recycling machine including such a fraud detection system, and to a computer program product. Background Technology

[0002] Examples of currency fraud using the Recycling System (RVS) are known. The primary cause of this fraud is that someone removes eligible used beverage containers (UBCs) from the system before they are compacted and devalued, or replaces the UBCs with worthless items before they are compacted. Summary of the Invention

[0003] The purpose of this invention is to provide anti-fraud systems and methods that may prevent the aforementioned fraudulent activities.

[0004] According to a first aspect of the invention, a fraud detection system for a recycling machine is provided, the system comprising: a detector adapted to detect at least one container entering the recycling machine; a compactor load sensor adapted to measure the load of a compactor of the recycling machine during operation, wherein the compactor is adapted to compact containers entering downstream of the detector; and a compactor load sensor monitoring device configured to determine, based on the load measured by the compactor load sensor, whether the detected at least one container is compacted as expected.

[0005] This invention is based on the understanding that the load measured by the compactor load sensor can be used to determine whether one or more detected containers are being compacted as expected. If they are not being compacted as expected, the recycling machine may be susceptible to fraud attempts. This system can be interpreted as a device.

[0006] It should be noted that EP2447020A1 discloses a control system, which includes an evaluation device configured to determine the compaction process from the measured phase difference. However, what EP2447020A1 does not disclose is the use of the determined compaction process to determine whether any detected containers have also been compacted.

[0007] To determine whether an individual container is being compacted as expected, a compactor load sensor monitoring device can be configured to compare the measured load of the compactor with the expected load of the individual container (the compactor). For example, a mismatch between the measured load and the expected load might be due to the fact that the container was removed from the recycling machine after being detected by the detector but before it was compacted, or that the detected container is different from the container or item being compacted. For example, the expected load could be a threshold, compactor load signature, cumulative load, etc. Furthermore, this expected load can be expected to occur within a predetermined time range after the container is detected by the detector. For example, the start and / or duration of said time range can depend on the design of the recycling machine.

[0008] The compactor load sensor monitoring device can be configured to determine the total number of compacted containers in a recycling machine session. The compactor load sensor monitoring device can be further configured to calculate changes in the fraud factor based on the number of compacted containers and the number of containers detected by the detector for the session. For example, if the number of containers detected by the detector is greater than the number of compacted containers, the fraud factor can be increased. In this way, the system can detect fraud not only for individual containers but also for a period of time. If the fraud factor exceeds a threshold, a trigger signal can be issued. For example, the trigger signal can trigger an alarm or shutdown of the recycling machine, or prohibit the recording of repayment amounts. Alternatively or supplementarily, the derivative of the fraud factor can be monitored for rapid fraud detection. Furthermore, the compactor load sensor monitoring device can be configured to significantly increase the fraud factor for each container not compacted as expected, and slightly decrease the fraud factor for each container compacted as expected. That is, for each container not compacted as expected, several containers must be appropriately compacted to avoid increasing the fraud factor during the session. This also means that the fraud factor may remain balanced even if not all containers are included in the calculation. This can compensate for any misreads caused by the compactor load sensor.

[0009] To determine whether multiple detected containers are being compacted as expected, a compactor load sensor monitoring device can be configured to accumulate the load for a session at the recycling machine, as measured by the compactor load sensor. Furthermore, the detector can detect containers entering the recycling machine, and based on the predetermined or known compactor loads of various containers, the expected accumulated load can be determined. The compactor load sensor monitoring device can then compare the accumulated load with the expected accumulated load. If the loads do not match, a fraudulent attempt may have occurred.

[0010] The detector can be selected from the group consisting of barcode readers, security tag readers, shape sensors, and material sensors, or combinations thereof (e.g., the system may include both barcode readers and security tag readers). The detector can be arranged together with the identification chamber of the recycling machine.

[0011] The system may further include a conveyor monitoring sensor positioned downstream of the detector and upstream of the compactor load sensor. The conveyor monitoring sensor may be positioned together with the conveyor or sorter unit of the recycling machine.

[0012] A repayment signal is issued only when each of the detector, transport monitoring sensor, and compactor load sensor has indicated that the container has been properly handled. In this way, a container must pass through at least three "checkpoints" before it is "approved," that is, before repayment is paid for the container.

[0013] A compactor load sensor monitoring device can be configured to analyze compactor load curves measured by a compactor load sensor to classify at least one container. For example, each container can be classified according to its type (plastic bottle, aluminum can, glass bottle, etc.) and / or according to its orientation (bottom-first, lateral, arbitrary, etc.). Any unclassified container may increase the aforementioned fraud factor. Furthermore, the compactor load sensor monitoring device can be configured to record compactor load curves for various containers, wherein the system further includes machine learning software configured to train the system based on the recorded compactor load curves. Specifically, the machine learning software can replace old compactor load curves with newly recorded curves. In this way, the system can adapt to changes in the compactor load curves due to compactor wear and breakage.

[0014] The detector can be adapted to send further information about at least one detected container to the compactor load sensor monitoring device. For example, the further information may include dimensions, weight, material type, etc. This can improve the system's ability to identify the object being compacted.

[0015] To measure the compactor load using the sliding measurement method, the compactor load sensor may include a speed sensor in the compactor's power transmission system. This type of sensor is relatively simple. In other embodiments, the sensor may be selected from the group consisting of: torque converters, load cells mounted in the compactor's power transmission system, frequency converters, power meters, and sliding sensors.

[0016] The system may further include a second compactor and a second compactor load sensor adapted to measure the load on the second compactor during operation. The system may include additional compactors and compactor load sensors.

[0017] The compactor load sensor monitoring device can be configured to determine the expected compaction time window of at least one container based on the moment when the container is detected by the detector.

[0018] The compactor load sensor monitoring device can be configured to construct a compaction event, which includes an expected compaction time window for at least one vessel and corresponding load data representing the load of the compactor as measured by the compactor load sensor.

[0019] The system may further include communication equipment suitable for transmitting the constructed compaction event to a remote device.

[0020] If the measured load within the expected compaction time window does not exceed a predetermined value, the compactor load sensor monitoring device (configured to determine whether at least one detected container is compacted as expected) can be configured to determine that the container is not compacted as expected. This is a relatively "easy" method for detecting fraud, and can be particularly useful if a database of container empty loads is not available.

[0021] The compactor load sensor monitoring device can be configured to normalize the measured load of at least one container and calculate a moving average based on the normalized measured load of the at least one container and the normalized measured load of one or more previous containers. By normalizing the measured load, it is not necessary to know the expected load of containers of different sizes and / or types.

[0022] If the calculated moving average matches a predetermined expected average, the compactor load sensor monitoring device (configured to determine whether at least one detected container is compacted as expected) can be configured to determine that the container is compacted as expected. In this way, it is possible to detect whether a container between the detector and the compactor has been replaced by another item.

[0023] The compactor load sensor monitoring device can be configured to normalize by dividing the measured load of at least one container by the empty load of at least one container.

[0024] According to a second aspect of the invention, a recycling machine is provided that includes the fraud detection system according to the first aspect. This aspect may exhibit the same or similar features and technical effects as the foregoing aspects.

[0025] According to a third aspect of the invention, a fraud detection method in a recycling machine is provided, the method comprising: detecting at least one container entering the recycling machine; measuring the load of a compactor attempting to compact the entered container; and determining, based on the load measured by a compactor load sensor, whether the detected at least one container has been compacted as expected. This aspect may exhibit the same or similar features and technical effects as the foregoing aspects, and vice versa.

[0026] According to a fourth aspect of the invention, a computer program product is provided, comprising code that, when run on a computer device, performs the following steps: determining, via a compactor located downstream of a detector, whether at least one container detected by a detector of a recycling machine is compacted as expected, based on the load of the compactor measured by a compactor load sensor. This aspect may exhibit the same or similar features and technical effects as the foregoing aspects, and vice versa. The computer device may, for example, be the aforementioned compactor load sensor monitoring device.

[0027] According to another aspect, a fraud detection system for a recycling machine is provided, comprising: a detector adapted to detect containers entering the recycling machine; a sensor for measuring a characteristic of a portion of the recycling machine while the container is being processed; and means configured to compare the measured characteristic with predetermined characteristic values ​​based on the type of container. This aspect is based on the fundamental principle of creating characteristic curves associated with a specific type of container being processed by the machine using the operating characteristics of components of the recycling machine, and on the fundamental principle of detecting potential fraud and / or processing errors using any mismatch between the expected characteristic curve and the actually measured operating characteristic. The portion of the recycling machine being measured may be a compactor adapted to compact containers entering downstream of the detector, and the characteristic may be the load of the compactor during the operation of compacting the container. The system may include a compactor load sensor adapted to measure the load of the compactor during operation; and a compactor load sensor monitoring means configured to compare the load measured by the compactor load sensor with the expected load of the compactor, wherein the expected load is based on the detected container. Attached Figure Description

[0028] These and other aspects of the invention will now be described in more detail with reference to the accompanying drawings, which illustrate the presently preferred embodiments of the invention.

[0029] Figure 1 A recycling machine including a fraud detection system according to an embodiment of the present invention is illustrated schematically.

[0030] Figure 2 The compactor load curve of the aluminum can is shown.

[0031] Figure 3This is a flowchart of a fraud detection method for individual containers according to an embodiment of the present invention.

[0032] Figure 4 This is a flowchart of a session-based fraud detection method according to an embodiment of the present invention.

[0033] Figure 5 This is a flowchart of a session-based fraud detection method according to another embodiment of the present invention.

[0034] Figure 6 A recycling machine including a fraud detection system is illustrated schematically according to another embodiment of the present invention.

[0035] Figures 7a to 7c One or more embodiments of the present invention are shown. Detailed Implementation

[0036] Figure 1 A fraud detection system 10 is shown integrated into the recycling machine 12. The recycling machine 12 may have a front-end unit and a rear-end unit, or it may be an integrated machine in which all functions are integrated into one unit. The overall function of the recycling machine 12 may be to automatically collect, sort, and process used beverage containers for recycling or reuse.

[0037] The fraud detection system 10 includes a detector 14. The detector 14 can be arranged together with the identification chamber 16 of the recycling machine 12. The detector 14 is adapted to detect containers 18 entering the recycling machine 12. The detector 14 can be a conventional barcode and / or security tag reader, or a shape or material sensor. The detector can be further adapted to issue information about the container 18, such as size, weight, material type, expected compactor load, etc.

[0038] System 10 may further include a transport monitoring sensor 20. The transport monitoring sensor 20 is arranged downstream of detector 14. The transport monitoring sensor 20 may be arranged together with the conveyor or sorter unit 22 of the recycling machine 12. The conveyor or sorter unit 22 is typically adapted to transport the container 18 to the compactor 24 of the recycling machine 12. The compactor 24 is designed to compact the container 18. Figure 1 In the illustrated embodiment, the recycling machine 12 has two destinations 26a and 26b, for example, one for (compacted) plastic bottles and one for (compacted) aluminum cans. A conveyor or sorter unit 22 ensures that the containers 18 are delivered to their correct destinations. Figure 1 In the example, container 18 is sent to destination 26b.

[0039] System 10 further includes a compactor load sensor 28. The compactor load sensor 28 is adapted to measure the load on the compactor 24 during operation. For example, the load on the compactor 24 can be interpreted herein as power consumption and / or torque during compactor operation. Thus, when the container is compacted, power consumption or torque and load increase. In one embodiment, the compactor load sensor 28 includes a speed sensor arranged in the power transmission system of the compactor 24. For example, the speed sensor can measure the rotational speed per minute (rpm) of the rotor of an electric motor. Using the speed sensor, the slippage of the motor can be measured, whereby the slippage determines the torque of the motor, and thus the load on the compactor 24 can be measured. In other embodiments, for example, the compactor load sensor 28 can be a torque converter, a load element mounted in the electric motor of the compactor 24, a frequency converter, a power meter, or other slip sensor.

[0040] System 10 further includes a compactor load sensor monitoring device 30. The compactor load sensor monitoring device 30 may be a standalone device or it may be integrated with the host computer or control system of the recycling machine 12. The compactor load sensor monitoring device 30 is connected to at least the detector 14 and the compactor load sensor 28 via connections 31a, 31b. Connections 31a, 31b may be wired or wireless. The compactor load sensor monitoring device 30 is typically configured to determine whether at least one detected container 18 is compacted as expected based on the load measured by the compactor load sensor 28.

[0041] For a single container 18, the compactor load sensor monitoring device 30 can compare the load measured by the compactor load sensor 28 with the expected load of the compactor 24 for the container 18. For example, the expected load could be a general threshold, a container-specific threshold, a compactor load marker (see below), cumulative load, etc. Depending on the layout of the recycling machine, this expected load is expected to appear some time after the container has been detected by the detector 14. For example, the expected time of the expected load can be expressed as a range, since the transfer time through the recycling machine 12 may vary slightly between different containers.

[0042] Furthermore, the compactor load sensor monitoring device 30 can be configured to analyze the compactor load curve of the load measured by the compactor load sensor 28. In this way, the compactor load sensor monitoring device 30 can classify the compacted containers 18 based on the analyzed compactor load curve. For example, each container 18 can be classified according to its type (plastic bottle, aluminum can, glass bottle, etc.) and / or according to its orientation (bottom preferred, lateral, any orientation, etc.). Figure 2An example of a compactor load curve or marker for an aluminum can is shown. Furthermore, the compactor load sensor monitoring device 30 can be configured to record compactor load curves for various containers. For example, the recorded curves can be stored in a database 32. The database 32 may also include pre-stored compactor load curves. Additionally, the system 10 may include machine learning software 34 configured to train the system 10 based on the recorded compactor load curves. Specifically, the machine learning software 34 can replace old compactor load curves (in the database 32) with newly recorded curves. In this way, the system 10 can adapt to changes in the compactor load curve due to wear and breakage of the compactor 24 over time.

[0043] The compactor load sensor monitoring device 30 can be further configured to calculate a fraud factor. If the calculated fraud factor exceeds a threshold X, a trigger signal can be issued. For example, the trigger signal can trigger an alarm or shut down the recycling machine 12. Optionally or complementaryly, the derivative of the fraud factor can be used for rapid fraud detection. The fraud factor can initially be set to a value below the threshold X.

[0044] In one implementation, the fraud factor is changed based on container detection. For each container detected by detector 14, a signal that the container (UBC) is en route can be sent to compactor load sensor monitoring device 30. Compactor load sensor monitoring device 30 attempts to find a compactor load curve that matches any compactor load curve stored in database 32 among the loads measured by compactor load sensor 28. For each match and for each found compactor load curve that does not match the curve in database 32 (unclassified containers), compactor load sensor monitoring device 30 increments a counter. The change in the fraud factor can then be calculated based on the counter and multiple containers detected by detector 14 for a session of recycling machine 12 (session = the first to last container entered by the consumer or a subset of these containers):

[0045] The change in the fraud factor = (|(detected container # - compacted container #)|*A) - (compacted container #*B) where (A>B)

[0046] If the detector detects 15 containers and a total of 12 containers are compacted (10 classified and 2 unclassified), the change in the fraud factor is (15-12)A–(12)B=3A-12B.

[0047] If the compactor load sensor monitoring device 30 detects that the number of compacted containers is somewhat greater than the number of containers detected by the detector 14, the fraud factor can also be increased by using the absolute value of the difference between the detected containers and the compacted containers. For example, the compactor load sensor monitoring device 30 can find two compactor load marks for the detected containers and thus double the counter.

[0048] In another implementation, container-based classification further alters the fraud factor. Here, the compactor load sensor monitoring device 30 only counts compactor load curves that have been found to match the compactor load curves in the database 32; that is, it counts compacted containers classified as valid containers. The change in the fraud factor can then be calculated according to the following exemplary formula:

[0049] The change in the fraud factor = (|(the detected container's # – the classified container's #)|*A) - (the classified container's #*B) where (A>B)

[0050] Using the above examples of 15 detected containers and 10 classified containers, the change in fraud factor is (15-10)A–(10)B=5A-12B.

[0051] Figure 3This is a flowchart of a fraud detection method for a single container 18. In S1, detector 14 detects container 18 in the identification chamber 16 of recycling machine 12. In S2, system 10 checks whether a barcode is found on container 18 and accepts it. If not, container 18 is returned to the consumer who has submitted container 18 to recycling machine 12. If yes, system 10 further checks whether a security mark is found on container 18 (S3) and accepts it. If not, container 18 is returned to the consumer. If yes, container 18 is conveyed by conveyor or sorter unit 22 to its designated destination 26a or 26b (S4). In S5, system 10 checks whether the conveyor monitoring sensor 20 is triggered as expected, i.e., whether container 18 is correctly conveyed or sorted by conveyor or sorter unit 22. If not, the return of container 18 is not recorded. If so, the compactor load sensor monitoring device 30 determines whether the container 18 has been compacted as expected, for example, by comparing the load measured by the compactor load sensor 28 with the expected load of the container 18 (S6). If yes (the measured load and the expected load match), a repayment signal for the container 18 can be issued. If no (the measured load and the expected load do not match), repayment for the container 18 is not recorded. In general, the container 18 must pass four checkpoints (S2, S3, S5, S6) before it is approved, i.e., before repayment is made to the container 18. In other embodiments, any one of checkpoints S2, S3, and S5 can be omitted.

[0052] Figure 4 This is a flowchart of a session-based fraud detection method according to an embodiment of the present invention. In S10, detector 14 detects containers 18 in the discrimination chamber 16 of recycling machine 12. In S11, compactor load sensor monitoring device 30 determines whether any containers have been compacted. If yes, the counter is incremented accordingly (S12). If no, the counter is not incremented (S13). Then, system 10 checks whether the session has ended (S14). If no, another container 18 is detected in S10, and so on. If yes, compactor load sensor monitoring device 30 calculates a change in fraud factor based on the counter and the number of containers 18 detected by detector 14 (S15). Then, system 10 checks whether the fraud factor exceeds a threshold X (S16). If yes, a trigger signal is issued (S17). If no, the operation of recycling machine 12 continues as normal (S18).

[0053] Figure 5This is a flowchart of a session-based fraud detection method according to another embodiment of the present invention. In S20, the compactor load sensor 28 measures the load of the compactor 24 during the session. The graph to the right of S20 shows the load measured over time. In S21, the compactor load sensor monitoring device 30 accumulates the load measured by the compactor load sensor 28. The graph to the right of S21 shows the load accumulated over time. In S22, the system 10 determines the expected accumulated load for the session. That is, the detector 14 detects which containers enter the recycling machine 12 during the session and determines the expected accumulated load based on the predetermined or learned compactor loads of the various containers. Then, in S23, the accumulated load from S21 is compared with the expected load from S22. If the loads do not match, a suspicious fraud attempt is indicated (S24). If they match, the operation of the recycling machine 12 can proceed as normal.

[0054] Figure 6 A recycling machine 12 including a fraud detection system 10 is shown according to another embodiment of the present invention. Figure 6 Recycling machine 12 and Figure 1 Similar to recycling machines, except Figure 6 The recycling machine 12 includes two detectors 14, 14' and two discrimination chambers 16, 16', as well as two conveyor monitoring sensors 20, 20' and two conveyor or sorter units 22, 22'. The system may further include at least two compactors 241, 24'. 2-n and at least two compactor load sensors 281, 28 2-n A standard compactor load sensor monitoring device 30 can be used to monitor and handle all compactors 24. 1-n Alternatively, the system may include a compactor load sensor monitoring device for each compactor / compactor load sensor.

[0055] According to one or more embodiments of the present invention, the compactor load sensor monitoring device 30 can be configured to determine the expected compaction time window 36 of the container 18 based on the moment when the detector 14 detects the container 18 (see [reference]). Figure 7b In other words, the expected compaction time window 36 is defined relative to the point in time when the container 18 is observed by the detector 14. Depending on the layout of the recycling machine 12, the expected compaction time window 36 may be a time range with a predetermined duration or length, and the expected compaction time window 36 may begin or occur at a predetermined time after the container 18 is detected by the detector 14.

[0056] Here, detector 14 may be adapted together with conveyor or sorter unit 22 (see Figure 7aContainer 18 may be inspected together with the last sorting unit before compactor 24 if there is more than one sorting unit.

[0057] The compactor load sensor monitoring device 30 can further be configured to generate compaction events, as schematically indicated by reference numeral 38. A compaction event 38 may include a projected compaction time window 36 for at least one container 18 and corresponding load data representing the load on the compactor 24 as measured by the compactor load sensor 28. The load data may include at least one of the following: a time 40 during which the measured load exceeds a predetermined value, the peak value of the measured load, and the integral of the load measured at that time. The predetermined value may be the no-load condition of the compactor 28 when no container is being compacted plus an offset. Figure 7b In the middle, the predetermined value is set to 50 on the vertical axis.

[0058] For a container 18 that is compacted individually, compaction event 38' includes the expected compaction time window and load data for that container. For containers 18 that are compacted together in some or more ways, where their expected compaction time windows 36 overlap, compaction event 38' may include the expected compaction time windows and load data for all of those containers.

[0059] The compaction event 38 may include additional information such as at least one of the following: the material of container 18 (ALU, FE, PET, glass, etc.), empty weight (e.g., in grams), and volume (e.g., in milliliters). For example, this additional information may be retrieved in a database based on the barcode of container 18 read by barcode reader 46 of system 10.

[0060] System 10 may further include a communication device 42 adapted to transmit the constructed compaction events 38 to a remote device 44, for example, for data visualization and / or offline analysis. Offline analysis may include analyzing constructed compaction events from each machine 12 over a period of several days / weeks and looking for significant changes in behavior, and / or looking for significant time differences between similar machines. For example, the constructed compaction events 38 may be transmitted to the remote device 44 once a day.

[0061] If the load measured in the corresponding expected compaction time window 36 does not exceed the predetermined value, for example if in the expected compaction time window (such as in Figure 7bIf no load is measured within the expected compaction time window 36”, the compactor load sensor monitoring device 30 may be further configured to determine that the detected container 18 has not been compacted as expected. Conversely, if the measured load exceeds a predetermined value within the expected compaction time window 36’, the compactor load sensor monitoring device 30 may determine that the item (potentially the corresponding container 18) has been compacted. In this way, the compactor load sensor monitoring device 30 can determine the percentage of uncompacted container (“loose item”) for each detected container 18 for each compaction event constructed. In the expected compaction time window 36’, the percentage is 0 / 1 = 0%. In the expected compaction time window 36”” (compaction event 38”), the percentage is 0 / 5 = 0%. A moving average of the uncompacted container share of multiple (e.g., 50) last detected containers 18 can be calculated by the compactor load sensor monitoring device 30. A moving average can be compared to a predetermined maximum value (the amount of uncompacted container). Moving averages can be used to calculate and / or update fraud factors. Moving averages can be exponential moving averages.

[0062] Understandably, the functionality described in the preceding paragraphs can determine that something has been compacted within the expected compaction time window, but cannot definitively determine that the compacted item is actually the same container as the detected container. To this end, the compactor load sensor monitoring device 30 can be further configured to normalize the measured load of at least one detected container 18 by dividing by the empty load of at least one container 18 as indicated in the additional information above. The normalized measured load can be expressed as hysteresis per gram, where "hysteresis" is the delay in compaction of the compactor 28 and thus represents the load of the compactor 28. The compactor load sensor monitoring device 30 can be further configured to calculate a moving average based on the normalized measured load of at least one container 18 and based on the normalized measured load of one or more previous containers. For example, the moving average can be an exponential moving average, and the one or more previous containers can be 10, 50, or 100 previous containers. If the calculated moving average matches the expected predetermined average (within a predetermined margin), the compactor load sensor monitoring device 30 can be further configured to determine that the container 18 is generally being compacted as expected. If the calculated moving average does not match the expected predetermined average, some containers 18 are not being compacted as expected (removed or replaced), which may indicate a fraud attempt. This can be used to calculate and / or update the fraud factor.

[0063] In operation ( Figure 7cIn step S1, the expected compaction time window 36 for container 18 is determined. In step S2, the load on compactor 24 is measured. In step S3, the aforementioned additional data, including the container's empty load, is retrieved.

[0064] In step S4, a compaction event 38 is constructed based on the expected compaction time window 36, the load, and additional data. In step S5, the constructed compaction event 38 can be sent to a remote device 44.

[0065] In step S6, if the measured load within the expected compaction time window does not exceed a predetermined value, it is determined that the container has not been compacted as expected, thereby determining the proportion of uncompacted detected containers 18. In step S7, a moving average of the proportion of uncompacted containers is calculated.

[0066] In step S8, the measured load of one or more containers 18 is standardized. In step S9, a moving average of the standardized measured load is calculated. In step S10, the calculated moving average is compared with a predetermined expected average.

[0067] Those skilled in the art will recognize that the present invention is by no means limited to the embodiments described above. Rather, various modifications and changes can be made within the scope of the appended claims.

Claims

1. A fraud detection system for a recycling machine, the fraud detection system comprising: A detector adapted to detect at least one container entering the recycling machine; A compactor load sensor, adapted to measure the load of the compactor of the recycling machine during operation, wherein the compactor is adapted to compact the incoming container downstream of the detector; and A compactor load sensor monitoring device is configured to determine, based on the load measured by the compactor load sensor, whether at least one of the detected containers is compacted as expected. The compactor load sensor monitoring device is configured to determine an expected compaction time window for the container based on the moment the detector detects the container, the expected compaction time window starting or occurring at a predetermined time after the detector detects the container. If the measured load within the expected compaction time window does not exceed a predetermined value, the compactor load sensor monitoring device is configured to determine that the container has not been compacted as expected.

2. The fraud detection system according to claim 1, wherein, The compactor load sensor monitoring device is configured to determine whether a separately detected container is compacted as expected by comparing the load measured by the compactor load sensor with the expected load for the detected container.

3. The fraud detection system according to claim 2, wherein, The expected load is expected to appear within a predetermined time range after the container is detected by the detector.

4. The fraud detection system according to any one of the preceding claims, wherein, The compactor load sensor monitoring device is configured to determine the number of compacted containers for a session of the recycling machine, and to calculate a change in the fraud factor based on the number of compacted containers and the number of containers for the session detected by the detector.

5. The fraud detection system according to claim 4, wherein, If the fraud factor exceeds the threshold, a trigger signal is issued.

6. The fraud detection system according to claim 4, wherein, The compactor load sensor monitoring device is configured to increase the fraud factor by a larger amount A for each container that is not compacted as expected, and decrease the fraud factor by a smaller amount B for each container that is compacted as expected.

7. The fraud detection system according to claim 1, wherein, The compactor load sensor monitoring device is configured to determine whether multiple detected containers are compacted as expected by accumulating the load for a session of the recycling machine as measured by the compactor load sensor.

8. The fraud detection system according to claim 1, wherein, The detector is selected from the group consisting of: barcode readers, security tag readers, shape sensors, and material sensors.

9. The fraud detection system of claim 1, further comprising a transport monitoring sensor disposed downstream of the detector and upstream of the compactor load sensor.

10. The fraud detection system according to claim 9, wherein, A repayment signal is issued only when each of the detector, the transport monitoring sensor, and the compactor load sensor has indicated that the container is being handled correctly.

11. The fraud detection system according to claim 1, wherein, The compactor load sensor monitoring device is configured to analyze the compactor load curve of the load measured by the compactor load sensor to classify at least one of the containers.

12. The fraud detection system according to claim 11, wherein, The compactor load sensor monitoring device is configured to record the compactor load curves for various containers, and the fraud detection system further includes machine learning software configured to train the fraud detection system based on the recorded compactor load curves.

13. The fraud detection system according to claim 1, wherein, The detector is adapted to send further information about at least one of the detected containers to the compactor load sensor monitoring device.

14. The fraud detection system according to claim 1, wherein, The compactor load sensor includes a speed sensor in the power transmission system of the compactor.

15. The fraud detection system of claim 1, further comprising a second compactor and a second compactor load sensor adapted to measure the load of the second compactor during operation.

16. The fraud detection system according to claim 1, wherein, The compactor load sensor monitoring device is configured to generate compaction events, the compaction events including at least one of the expected compaction time windows of the container and corresponding load data representing the load of the compactor as measured by the compactor load sensor.

17. The fraud detection system of claim 16, further comprising a communication device adapted to transmit the constructed compaction event to a remote device.

18. The fraud detection system according to claim 1, wherein, The compactor load sensor monitoring device is configured to normalize the measured load of at least one of the containers and to calculate a moving average based on the normalized measured load of at least one of the containers and the normalized measured load of one or more previous containers.

19. The fraud detection system according to claim 18, wherein, If the calculated moving average matches the expected predetermined average, the compactor load sensor monitoring device is configured to determine that the container is compacted as expected.

20. The fraud detection system according to claim 18 or 19, wherein, The compactor load sensor monitoring device is configured to normalize the measured load of at least one of the containers by dividing the measured load of at least one of the containers by the empty load of at least one of the containers.

21. A recycling machine comprising a fraud detection system according to any one of claims 1 to 20.

22. A fraud detection method in a recycling machine, the method comprising the following steps: Detect at least one container entering the recycling machine; The expected compaction time window of the container is determined based on the time when the container is detected, wherein the expected compaction time window begins or occurs at a predetermined time after the container is detected; Measuring the load on the compactor intended to compact the contents of the container; and The determination of whether at least one of the containers is compacted as expected is based on the load measured by the compactor load sensor, wherein if the measured load within the expected compaction time window does not exceed a predetermined value, it is determined that the container is not compacted as expected.

23. A computer-readable non-volatile storage medium coupled to a processor and storing a program executable by the processor, which, when run by the processor, performs the following steps: The expected compaction time window of the container is determined based on the moment when the detector detects that the container has entered the recycling machine, wherein, The expected compaction time window begins or occurs at a predetermined time after the container is detected by the detector; and Based on the load of the compactor measured by the compactor load sensor, it is determined whether the container, as detected by the detector of the recycling machine, is compacted as expected. If the measured load within the expected compaction time window does not exceed a predetermined value, it is determined that the container has not been compacted as expected.

Images