Consumption device and method for authenticating a consumption component

Abstract

Consumption device (900), comprising: A sensor (901) configured to acquire sensor data describing the physical behavior of an authentication chip (902) of a consumption component (903); A sender that is configured to send a request for authentication information to the consumption component (903); A receiver that is configured to receive authentication information from the consumption component; wherein the sender and receiver are configured to send and receive the request and authentication information via at least one first communication channel; and wherein the sensor (901) is configured to acquire sensor data representing information transmitted by the consumption component via a communication channel other than the at least one first communication channel; An authentication circuit (904) is configured to determine whether the authentication information corresponds to a predefined reference authentication information; and to implement a machine learning model (905) that is trained to classify consumption components into originals and imitations based on sensor data describing the physical behavior of authentication chips (902) of the consumption components; to feed the acquired sensor data to the machine learning model (905); to determine whether the machine learning model (905) classifies the consumption component (903) as an original; and, depending on whether both the authentication information corresponds to a predefined reference authentication information and whether the machine learning model (905) classifies the consumption component (903) as an original, to authorize the use of the consumption component (903) by the consumption device (900).

Description

[0001] Examples of implementation generally relate to consumer devices and methods for authenticating a consumer component.

[0002] Counterfeiting is a significant issue in consumer markets for consumer goods. There is a high risk that counterfeit companies will create clones of consumer components that behave exactly like the original components.

[0003] The production of clones of consumer components leads to a significant loss of revenue for the manufacturers of the original components, typically more than 20%. Examples of markets where massive damage is caused by counterfeiters include consumables and replacement products such as printer cartridges, camera batteries, e-cigarette boxes, etc.

[0004] Authentication chips are typically used to protect consumable components. Their function is usually to verify that a consumable component (e.g., a printer cartridge) can be identified as genuine to the host device (e.g., the printer). In this example, the printer would only accept and use the cartridge for printing after it has been authenticated, or conversely, warn the user that it is a non-genuine cartridge. Furthermore, the printer manufacturer may void the warranty if non-genuine parts are used.

[0005] However, a counterfeiter can also imitate (clone) such an authentication chip. Therefore, measures are desirable to prevent the use of counterfeit consumer goods with counterfeit authentication chips in genuine consumer devices.

[0006] Publication US 2019 / 0280885A1 describes implementations of physically unclonable functions (PUFs) for cryptographic and authentication purposes. In particular, it describes implementations of machine learning engines (MLESs) in conjunction with PUFs that generate multi-state outputs.

[0007] Document US 2019 / 0369174A1 describes a system for use in authentication processes. This system comprises a PUF (Powered Unshielded Fiber), a substrate, a variety of magnetized particles randomly distributed within the substrate, and a PUF reader constructed using multiple discrete magnetometer chips with magnetic field sensors arranged on a printed circuit board. The PUF reader measures magnetic field data at multiple locations in close proximity to the magnetized particles. The measured magnetic field data can be compared with previously recorded data to assess authenticity.

[0008] Publication US 8,782,396 B2 describes PUFs for authenticating a variety of electronic devices, including FPGAs, RFIDs, and ASICs. In some implementations, challenge-response pairs corresponding to individual PUFs can be registered and used to determine authentication data that can be managed in a database. Later, when a target object is to be authenticated with a PUF, a series of challenges is applied to each PUF device to authenticate it and thus distinguish it from others.

[0009] Document US 9,619,663 B2 describes a replaceable printer component that includes a first storage device and a communication link. The first storage device is configured to store a first secret. The communication link is configured so that the first storage device communicates with a printer controller when the replaceable printer component is installed in a printing system. The printing system includes a second storage device that stores a second secret. The printer controller is configured to determine the authenticity of the replaceable printer component based on the first and second secrets.

[0010] According to one embodiment, a consumer device is provided, comprising a sensor configured to acquire sensor data describing the physical behavior of an authentication chip of a consumer component, a transmitter configured to send a request for authentication information to the consumer component, a receiver configured to receive authentication information from the consumer component, wherein the transmitter and the receiver are configured to send and receive the request and the authentication information via at least one first communication channel, respectively, and wherein the sensor is configured to acquire sensor data representing information transmitted by the consumer component via a communication channel other than the at least one first communication channel, and an authentication circuit configured toto determine whether the authentication information corresponds to a predefined reference authentication information, and to implement a machine learning model that is trained to classify consumption components into originals and imitations based on sensor data describing the physical behavior of the consumer components' authentication chips, and to feed the acquired sensor data to the machine learning model, to determine whether the machine learning model classifies the consumption component as an original, and, depending on whether the machine learning model classifies the consumption component as an original, to authorize the use of the consumption component by the consuming device.

[0011] According to another embodiment, a method for authenticating a consumption component is provided in accordance with the consumption device described above.

[0012] The figures do not represent the actual proportions but are intended to illustrate the principles of the various embodiments. Several embodiments are described below with reference to the following figures. Fig. Figure 1 shows a consumer goods arrangement. Fig. Figure 2 shows a flowchart for authenticating a consumer goods authentication circuit with a host authentication circuit. Fig. Figure 3 shows a consumer goods arrangement according to one embodiment. Fig. Figure 4 shows a flowchart for authenticating a consumer goods authentication circuit with a host authentication circuit according to one embodiment. Fig. Figure 5 shows an arrangement with a consumer goods authentication circuit and a consuming device with an on-chip analog / digital converter. Fig. Figure 6 shows an arrangement with a consumer goods authentication circuit and a consuming device with a chip-external analog / digital converter. Fig. Figure 7 shows an arrangement with a consumable authentication circuit and a consuming device, wherein the control chip of the consumable has an application processor that executes classification software. Fig. Figure 8 shows an arrangement with a consumer goods authentication circuit and a consuming device, in which the consuming device additionally has a microcontroller that executes classification software. Fig. Figure 9 shows a consumption device according to one embodiment. Fig. Figure 10 shows a flowchart illustrating a procedure for authenticating a consumption component according to one embodiment.

[0013] The following detailed description refers to the accompanying figures, which show details and exemplary embodiments. These exemplary embodiments are described in sufficient detail to enable a person skilled in the art to carry out the invention. Other embodiments are also possible, and the exemplary embodiments can be modified in structural, logical, and electrical terms without deviating from the subject matter of the invention. The various exemplary embodiments are not necessarily mutually exclusive; rather, different embodiments can be combined to create new embodiments. Within the scope of this description, the terms "connected," "connected," and "coupled" are used to describe both direct and indirect connections, direct or indirect connections, and direct or indirect couplings.

[0014] There are several strategies to prevent a professional counterfeiter from manufacturing counterfeit products (such as consumables). One strategy is to connect / embed an authentication circuit (e.g., an authentication ASIC) with / into the removable medium (i.e., the consumable such as a printer cartridge) and to perform one-way or mutual authentication from the host side (i.e., from the consuming device, such as a printer). This is discussed below with reference to Fig. 1 and Fig. 2 described. Fig. Figure 1 shows a consumer arrangement 100, consisting of a consumer 101 (also referred to as a consumer component device) and a consuming (or using) device 102 (also referred to as a consumer device).

[0015] The consumable 101 is a device that provides (and, for example, stores) a resource that is consumed when the consuming device 102 is in operation.

[0016] For example, a consumable contains a (physical) material that is consumed, such as a printer cartridge, a vaporizer cartridge for insecticides or insect repellents, an e-cigarette refill cartridge, or a medical substance (e.g., medicine) for a medical device in a suitable container.

[0017] Examples of pairs of consumable good 101 and consuming device 102 are: • Printer cartridge - Printer • Refill cartridge - inhaler • Refill cartridge - insect repellent container

[0018] The consuming device 102 can also be a vehicle or a camera (e.g., the consuming component is then a battery).

[0019] The consumable item 101 is physically connected to the consuming device 102, for example by being plugged in or installed. The consumable item 101 is typically interchangeably (in particular, detachably) connected to the consuming device 102.

[0020] The manufacturer of a consuming device 102 typically wishes that only consumables 101 manufactured by him (or a licensee) may be used with the consuming device 102.

[0021] Therefore, it may be provided that the consuming device 102 has a host authentication circuit 103, against which a consumable authentication circuit 104 must authenticate. For example, the consuming device 102 has a control unit 105 that only permits (authorizes) the operation of the consuming device 102 (also referred to as the host) with the consumable 101 (also referred to as the consumable) if the consumable 101 has successfully authenticated itself to the host authentication circuit 103 of the consuming device 102 by means of a consumable authentication circuit 104.

[0022] Fig. Figure 2 shows a flowchart 200 for authenticating a consumer goods authentication circuit 201 with a host authentication circuit 202.

[0023] The consumer goods authentication circuit 201 and the host authentication circuit 202 correspond, for example, to the consumer goods authentication circuit 104 and the host authentication circuit 103, respectively. Fig. 1.

[0024] To authenticate the consumable authentication circuit 201 with the host authentication circuit 202, the host authentication circuit 202 sends a challenge message 204 to the consumable authentication circuit 201 in 203. In 205, the consumable authentication circuit 201 generates a response, which it transmits to the host authentication circuit 202 in 207 as a response message 206. In 208, the host authentication circuit 202 verifies that the response is correct, for example, that it matches the challenge, and authenticates the consumable authentication circuit 201 if so.

[0025] For example, the host authentication circuit 202 inserts a random number into the challenge message 204. The consumer authentication circuit 201 must use this number to correctly generate a response in order to be authenticated by the host authentication circuit 202. For example, the authentication might be based on Elliptic Curve Cryptography (ECC). The response could also depend on a cryptographic key negotiated between the consumer authentication circuit 201 and the host authentication circuit 202.

[0026] With an authentication circuit 104, e.g. in the form of an authentication chip, on a consumable good 101, it can ideally be ensured that a consuming device 102 only uses original consumable component devices.

[0027] However, there is a risk that a manufacturer of a non-original consumer good (i.e., the imitator) will equip the non-original consumer good with a non-original authentication chip (i.e., a clone of the authentication circuit).

[0028] To make it more difficult for imitators, the authentication circuit can be manufactured using special anti-cloning technologies that are difficult to reverse engineer. The main obstacle for an imitator in creating a clone is this reverse engineering step. However, once the imitator has successfully reverse-engineered the circuit and, for example, extracted secret keys from the authentication circuit, they can produce a functional clone using a standard chip design flow. They can even produce the clone with cost optimization in mind, potentially using different technologies, design libraries, and so on.

[0029] The following are examples of implementations that provide a second line of defense for the manufacturer and enable a consuming device to distinguish a counterfeit consumer component with a functional clone of an authentication chip from an original consumer component with an original authentication chip.

[0030] It is assumed that the clone fulfills the original functional specification, so that the consuming device cannot distinguish the clone of the authentication chip using cryptographic protocols or functional tests.

[0031] According to various implementation examples, it is exploited that a clone typically behaves differently from an original authentication chip with regard to other properties and metrics. Examples include the following: • The clone is a different size • The clone has a different layout (placement of memory, analog blocks, routing of electrical lines) • The clone uses a different technology (e.g., it is manufactured using a different technology) • The clone has a different number of metal layers • The clone has a slightly different average power consumption. • The clone has a different dynamic power consumption. • The clone has a slightly different timing behavior (e.g., when switching on or when changing to or from sleep mode) • The clone has a different electromagnetic radiation.

[0032] A clone can, in principle, be identified based on such differences. However, detecting the aforementioned differences often requires laboratory equipment and manual intervention, such as opening the device or component, taking photographs or X-rays, measuring the size, etc.

[0033] However, electrical and electromagnetic side channels offer the possibility of implementing a cost-effective sensor on the host side (e.g., the consuming device 202) that can capture sensor data showing anomalies, i.e., deviations from the typical behavior of the original authentication circuit. Detecting these anomalies based on the sensor data is not trivial, however, due to random variations in the original components (i.e., the original authentication chips) and surrounding noise, such as noise in the consuming device or the consuming component.

[0034] Therefore, according to various embodiments, the host is supplemented not only by a sensor but also by a reliable classification circuit, as described in Fig. 3 is shown.

[0035] Fig. Figure 3 shows a consumer goods arrangement 300, consisting of a consumer good 301 and a consuming device 302.

[0036] As with reference to Fig. As described in Figure 1, the consuming device 302 has a host authentication circuit 303, against which a consumable authentication circuit 304 must authenticate. The consuming device 302 has a control unit 305 that only allows the operation of the consuming device 302 (also referred to as the host) with the consumable 301 (also referred to as the consumable) if the consumable 301 has successfully authenticated itself to the host authentication circuit 303 of the consuming device 302 by means of a consumable authentication circuit 304.

[0037] In contrast to the consumer goods arrangement 100 of Fig. In the consuming device 302, a sensor 306 is also included, which acquires sensor data describing the physical behavior of the consumable authentication circuit 201, such as a current sensor or an electromagnetic side-channel sensor. The consuming device 302 also includes a classification circuit 307, which, based on the acquired sensor data, can distinguish a cloned consumable authentication circuit from an original consumable authentication circuit and communicates the result of this decision (i.e., the classification of the consumable authentication circuit) to the host authentication circuit 303. The host authentication circuit 303 only allows the use of the consumable 301 if the classification circuit 307 classifies the consumable authentication circuit 304 as an original consumable authentication circuit 304 based on the sensor data.

[0038] The sensor 306 and the classification circuit 307 thus create a second line of defense against the counterfeiting of the consumable component, which then comes into play when the first line of defense, the protection against counterfeiting through authentication by the consumable authentication circuit 304 (as with reference to Fig. 2) is breached by cloning the 304 consumable authentication circuit. The second line of defense can extend the consumable's lifespan with regard to its security. It also allows for cost savings in the 304 consumable authentication circuit, as a cheaper consumable authentication circuit can be used, which, for example, has few countermeasures against reverse engineering or is based on an older technology that offers limited security.

[0039] The classification circuit 307 can be at least partially executed in software, i.e., the host authentication circuit 303 can have a programmable processor that executes a corresponding program for classification.

[0040] The sensor 306, the classification circuit 307, the host authentication circuit 303 and / or the controller 305 can be provided together on one chip (e.g. a host controller).

[0041] The classification circuit 307 can also be considered part of the host authentication circuit 303, which is designed in the form of an integrated circuit (i.e., as one chip) or several integrated circuits (i.e., multiple chips), depending on the application.

[0042] To reliably distinguish cloned authentication circuits from original authentication circuits, various embodiments provide that the classification circuit 307 implements a machine learning model 308 trained to differentiate cloned authentication circuits from original authentication circuits based on sensor data. An example of this is a neural network, but other machine learning models are also suitable for classification, such as a polyhedral classifier.

[0043] The machine learning model 308 can be trained using training data from a variety of training datasets, each containing sensor data and an associated correct classification (ground truth), according to a machine learning procedure (i.e., a training procedure). This can be done, for example, before the delivery of the consuming device, and the classification circuit is then configured accordingly, but it can also be done subsequently by uploading appropriate program instructions.

[0044] For example, the consuming device 302 is equipped with a sensor 306 (i.e., a measuring device) for the power consumption of the consumer goods authentication circuit 304. In practice, for example, a measured quantity is acquired and recorded over time, e.g., by means of a voltage sensor, current sensor, or a sensor for measuring an electromagnetic field.

[0045] The manufacturer (of the original consumable component and / or the consuming device) characterizes the original consumable authentication chips with respect to dynamic variations of the measured quantity (in this example, the "power trace") in a model building phase for the machine learning model 308. Noise is also taken into account for the machine learning model 308.

[0046] The information on the measured quantity for all available clones is used to maximize the discrimination capability of the machine learning model 308 (by training the machine learning model 308 to correctly distinguish, i.e., classify, clones and originals). The measurements (on clones and originals to generate the training data) during the model setup phase can be performed in a laboratory environment using high-resolution sensors instead of the consuming device's sensor 306, which, for example, has limited measurement capabilities. One result of the measurements is, for example, a dynamic performance model for the consumer goods authentication circuit, which can be embedded in the classification circuit and / or used for its configuration. For example, it can be used to generate training data for the machine learning model 308.

[0047] During operation in the field, the consuming device dynamically records the measured value. The resulting profile (power trace) is classified by the classification circuit. This can be seen as a comparison with the dynamic power model for the consumable authentication circuit.

[0048] If the classification circuit detects a clone, appropriate measures are taken. For example, if it detects that the consumption component authentication circuit 304 is a clone, the host authentication circuit 303 prevents the use of the consumption component 301.

[0049] Fig. Figure 4 shows a flowchart 400 for authenticating a consumer goods authentication circuit 401 with a host authentication circuit 402.

[0050] The Consumer Authentication Circuit 401 and the Host Authentication Circuit 402 are described in 403 to 408 as follows: Fig. 2 described an authentication procedure (using Challenge 404 and Response 406).

[0051] During at least part of the execution of the authentication procedure, a sensor 415, corresponding to sensor 306, performs a measurement 409 of the power consumption (or radiation etc.) of the consumer goods authentication circuit.

[0052] The measurement result is fed to a classifier 414 in 410, according to the classification circuit 307.

[0053] In 411, the classifier 414 performs a classification of the consumable authentication circuit 401 (into the classes "original" and "clone") and transmits the classification result in 412 to the host authentication circuit 402, which in 413 decides whether the consumable is permitted. It allows the consumable to be used with the consuming device if the response to the challenge is correct and the classifier 414 has classified the consumable authentication circuit as original.

[0054] Further measures (or other actions) may also be taken if a clone is detected. The type of measure depends heavily on the consumable and its functionality. For example, in the case of a counterfeit printer cartridge, the host may inform the customer that their warranty is void. In the case of a refill cartridge for an inhaler (vaporizer), the inhaler may cease to function due to a potential health risk to the user.

[0055] An update functionality may be provided so that an improved classification model 308 can be loaded onto the consumable 302, for example when information about new clones is available or when environmental conditions have changed (such as through a printer update).

[0056] The consuming device 302 can stimulate the consumable authentication circuit 304 with unexpected patterns (i.e., inputs, e.g., challenges) that a cloner cannot anticipate and measure the behavior of the consumable authentication circuit 304 when processing these patterns. In this way, the probability of distinguishing between the original and the clone can be increased. Stimulus changes and correspondingly modified classification models can be loaded onto the consuming device 302 (e.g., by updating as described above) to hinder the clone from learning to produce clones that are classified as originals.

[0057] Measurements (e.g., power trace measurements) can be performed at different stages of communication or processing: • Startup phase: The focus here is on dynamic differences in current during wake-up, e.g., of RAM (Random Access Memory), non-volatile memory, power supply, and analog components. • During authentication (as with reference to Fig. (described in section 4): The focus here is on dynamic differences in the execution of cryptographic algorithms. • During memory write cycles • During memory read cycles

[0058] The classification circuit 307 (which, for example, may be configured to control the sensor 306 regarding when and / or how it measures) can change its behavior and switch between accessing measurements in different phases. This, in turn, can hinder imitators from learning to produce clones that are classified as originals.

[0059] For example, the consumer device 301 evaluates sensor data from phase A during operation (i.e., performs classification and authentication). However, it may be designed to switch to classification based on sensor data from one of several other phases B (C, D,...), i.e., sensor data measured in a different operating phase. For this purpose, it may already have classification models for the other phases available, which it does not initially use. The switching can occur via various mechanisms, e.g.: a) Triggered by, for example, a firmware update from the manufacturer (e.g., in response to a hack) b) Via a time-based control, e.g. after x operating hours, a switch is made to a different classifier (e.g. a different classification model) (which a potential imitator could not have observed before).

[0060] Furthermore, it may be provided that the consumption device 302 not only evaluates the sensor data (analog measurement signals) in normal operating mode at any phase (such as in the start-up phase, during a crypto operation, etc., see the list above) and performs the classification with the ML model built for this time window, but also controls the consumption item 301 in a "false" (i.e., not occurring in normal operation) sequence and performs the classification based on the response of the consumption item in this situation.

[0061] As mentioned above, it may be planned that the classification model 308 will be updated, for example via a firmware update. This can take the following forms: - another previously unused sensor is activated - a different time phase is evaluated (i.e., sensor data from a different phase is used as the basis for classification) - a modified, e.g. more accurate, model or a different ML classification algorithm is used.

[0062] For the implementation of sensor 306 and / or the conversion of an analog sensor voltage or analog sensor current that reflects a physical behavior of the consumer product authentication circuit, an analog-to-digital converter is provided, for example, on the consuming device 302. This can be located on the same chip or externally to the other components 303, 305, 306, and 307. The following sections refer to Fig. 5 to Fig. Eight different implementation variants are described.

[0063] Fig. Figure 5 shows an arrangement 500 with a consumable authentication circuit 501 (slave) and a consuming device 502 (host), where the consuming device includes a control chip 503 and a power supply 504. The chip implements, for example, the classifier 307, the host authentication circuit 303, and the controller 305.

[0064] An analog-to-digital converter 505 is integrated on the chip 503. The analog-to-digital converter 505 detects, for example, the power consumption of an authentication circuit of the consumer good 501 (from electrical energy supplied by the power supply) and transmits this measurement to the control chip 503.

[0065] Fig. Figure 6 shows an arrangement 600 with a consumable authentication circuit 601 (slave) and a consuming device 602 (host), where the consuming device includes a control chip 603 and a power supply 604. The chip implements, for example, the classifier 307, the host authentication circuit 303, and the controller 305.

[0066] An analog-to-digital converter 605 is provided externally to the chip 603. The analog-to-digital converter 605 detects, for example, the power consumption of an authentication circuit of the consumable 601 (from electrical energy supplied by the power supply) and transmits this measurement to the control chip 603.

[0067] Fig. Figure 7 shows an arrangement 700 with a consumable authentication circuit 701 (slave) and a consuming device 702 (host) similar to Fig. 6, in which the control chip 703 of the consumable 702 has an application processor 704 that executes classification software (and thus implements the classification circuit 307).

[0068] Fig. Figure 8 shows an arrangement 800 with a consumable authentication circuit 801 (slave) and a consuming device 802 (host) similar to Fig. 6, in which the consuming device 802, however, in addition to the control chip 803, has a microcontroller 804 which is connected to the control chip 803 and executes classification software 806 (and thus implements the classification circuit 307). Accordingly, the analog-to-digital converter 805 supplies the power consumption it detects to the microcontroller 804, which transmits the classification result to the control chip 803. The microcontroller 804 is, for example, a safety controller.

[0069] In another variant, the security controller on the host side has an embedded analog-to-digital converter. This increases security because an attacker cannot intercept, modify, or replay the communication between the analog-to-digital converter and the security controller.

[0070] In summary, according to various embodiments, a consumption device is provided, as described in Fig. 9 is shown.

[0071] Fig. Figure 9 shows a consumption device (i.e., a consuming device) 900 according to one embodiment.

[0072] The consumption device 900 has a sensor 901 which is configured to acquire sensor data describing a physical behavior of an authentication chip 902 of a consumption component 903 (i.e. a consumable).

[0073] The consumption device 900 further comprises an authentication circuit 904, which is configured to implement a machine learning model 905 trained to classify consumption components as originals and imitations based on sensor data describing the physical behavior of authentication chips of the consumption components. The authentication circuit 904 is further configured to feed the acquired sensor data to the machine learning model and, depending on whether the machine learning model classifies the consumption component as an original, to authorize (i.e., permit) the consumption device's use of the consumption component.

[0074] In other words, according to various embodiments, a machine learning model is trained to distinguish genuine authentication circuits (typically in the form of integrated circuits, i.e., chips) from imitation authentication circuits, i.e., to classify a given authentication circuit. The model receives sensor data as input, which characterizes the behavior of the given authentication circuit. This is therefore not the data transmitted by the authentication circuit as part of an authentication procedure, but rather data that characterizes additional information provided by the authentication circuit through means other than the authentication communication.The sensor data thus contains information that is transmitted from the consumer authentication chip to the authentication circuit of the consuming device via one or more side channels (in addition to the communication channel between the consumer authentication chip and the authentication circuit of the consuming device).

[0075] The components of the consumer device, in particular the authentication circuit, can be implemented by one or more circuits. In one embodiment, a "circuit" is understood to be any unit that implements logic and that may be hardware, software, firmware, or a combination thereof. Thus, in one embodiment, a "circuit" can be a hard-wired logic circuit or a programmable logic circuit, such as a programmable processor, e.g., a microprocessor. A "circuit" can also be understood to be a processor that executes software, e.g., any type of computer program, such as a computer program in programming code for a virtual machine. In one embodiment, a "circuit" can be understood to be any type of implementation of the functions described herein.

[0076] According to one embodiment, a method is provided as described in Fig. 10 is shown. Fig. Figure 10 shows a flowchart 100 illustrating a method for authenticating a consumption component in a consumption device according to one embodiment.

[0077] In 1001, sensor data is recorded that describes the physical behavior of an authentication chip of a consumer component.

[0078] In 1002, the acquired sensor data are fed into a machine learning model that is trained to classify consumer components into originals and imitations based on sensor data that describe the physical behavior of authentication chips of the consumer components.

[0079] In 1003, the use of the consumption component by the consumption device is authorized (i.e., permitted) depending on whether the machine learning model classifies the consumption component as original.

[0080] The following are various examples of implementation.

[0081] Exemplary embodiment 1 is a consumer device according to claim 1.

[0082] Exemplary embodiment 2 is a consumption device according to exemplary embodiment 1, wherein the sensor is configured to acquire sensor data that describes the physical behavior of the authentication chip of the consumption component when generating authentication information requested by the consumption component.

[0083] Exemplary embodiment 3 is a consumption device according to exemplary embodiment 1 or 2, wherein the sensor is configured to acquire sensor data that describes the physical behavior of the authentication chip of the consumption component during memory accesses.

[0084] Exemplary embodiment 4 is a consumer device according to one of the exemplary embodiments 1 to 3, wherein the sensor data are side channel measurement data.

[0085] Exemplary embodiment 5 is a consumer device according to one of the exemplary embodiments 1 to 4, wherein the sensor is configured to detect a voltage of the authentication chip, a current absorbed by the authentication chip, an electromagnetic field emitted by the authentication chip and / or other side-channel data, such as temperature, electromagnetic radiation and light.

[0086] Exemplary embodiment 6 is a consumption device according to one of the exemplary embodiments 1 to 5, wherein the machine learning model is a neural network.

[0087] Exemplary embodiment 7 is a consumer device according to one of the exemplary embodiments 1 to 6, wherein the consumer device is a printer, an evaporator, a battery-operated electrical device or a vehicle.

[0088] Exemplary embodiment 8 is a consumer device according to one of the exemplary embodiments 1 to 7, wherein the sensor comprises at least one sensor from a group consisting of an optical sensor, a magnetic sensor, an inductive sensor, a capacitive sensor, a temperature sensor, a current sensor, a voltage sensor, an acceleration sensor, a vibration or oscillation sensor and a chemical sensor.

[0089] Exemplary embodiment 9 is a consumer device according to one of the exemplary embodiments 1 to 8, comprising a power supply which is configured to supply the consumer component with energy via a power supply line, wherein the physical behavior is a power consumption behavior of the consumer component via the power supply line.

[0090] Exemplary embodiment 10 is a consumer device according to exemplary embodiment 9, wherein the sensor for detecting the power consumption of the consumer component is coupled to the power supply line via the power supply line.

[0091] Exemplary embodiment 11 is a consumer device according to one of the exemplary embodiments 1 to 10, wherein the sensor is configured to detect an analog measurement signal and comprising an analog / digital converter which is configured to convert the analog measurement signal into digital sensor data and supply it to the authentication circuit.

[0092] Exemplary embodiment 12 is a method for authenticating a consumption component in a consumption device according to claim 12.

[0093] Exemplary embodiment 13 is a method according to exemplary embodiment 12, further comprising training the machine learning model, including acquiring, for each consumption component of a plurality of consumption components, each classified as an original consumption component or as a mimicked consumption component, sensor data describing a physical behavior of an authentication chip of the consumption component, and training the machine learning model using the sensor data acquired for the plurality of consumption components and the classification of the plurality of consumption components as training data.

[0094] Exemplary embodiment 14 is a method according to exemplary embodiment 13, wherein the training of the machine learning model is carried out by adjusting parameters of the machine learning model in such a way that the ability of the machine learning model to predict, based on the respective acquired sensor data, whether a consumption component of the multitude of consumption components is classified as an original consumption component or a mimicked consumption component is increased.

[0095] Embodiment 15 is a method according to embodiments 12 to 14, wherein the sensor data are side channel measurement data.

[0096] Exemplary embodiment 16 is a method according to one of the exemplary embodiments 12 to 15, comprising detecting a voltage of the authentication chip, a current absorbed by the authentication chip, an electromagnetic field emitted by the authentication chip and / or other side-channel data, such as temperature, electromagnetic radiation and light.

[0097] Exemplary embodiment 17 is a method according to one of the exemplary embodiments 12 to 16, wherein the machine learning model is a neural network.

[0098] Embodiment 18 is a method according to one of embodiments 12 to 17, wherein the consumption device is a printer, an evaporator, a battery-operated electrical device or a vehicle.

[0099] Exemplary embodiment 19 is a method according to one of the exemplary embodiments 12 to 18, comprising acquiring the sensor data by means of a sensor which has at least one sensor from a group consisting of an optical sensor, a magnetic sensor, an inductive sensor, a capacitive sensor, a temperature sensor, a current sensor, a voltage sensor, an accelerometer, a vibration or oscillation sensor and a chemical sensor.

[0100] Embodiment 20 is a method according to one of embodiments 12 to 19, comprising supplying the consumption component with energy via a power supply line, wherein the physical behavior is a power consumption behavior of the consumption component via the power supply line.

[0101] Embodiment 21 is a method according to one of embodiments 12 to 20, comprising acquiring an analog measurement signal, converting the analog measurement signal into digital sensor data and supplying the digital sensor data to the authentication circuit.

[0102] Exemplary embodiment 22 is a computer-readable storage medium comprising model parameter data that specify a trained machine learning model which is trained to classify consumer components into originals and imitations based on sensor data that describe the physical behavior of authentication chips of the consumer components.

[0103] It should be noted that features described in connection with one embodiment can also be used in combination with all other embodiments, provided they do not contradict the other embodiment. In particular, all embodiments described for the consumption device and the method for authenticating a consumption component can be used in combination with the computer-readable storage medium.

[0104] Although the invention has been shown and described primarily with reference to specific embodiments, those familiar with the field should understand that numerous modifications regarding its design and details can be made without departing from the essence and scope of the invention as defined by the following claims. The scope of the invention is therefore determined by the appended claims, and it is intended that all modifications falling within the literal or equivalent scope of the claims are included. Reference symbol list 100 Consumer Goods Arrangement 101 Consumer goods 102 consuming device 103 Host authentication circuit 104 Consumer Goods Authentication Circuit 105 Control unit 200 Flowchart 201 Consumer Goods Authentication Circuit 202 Host authentication circuit 203 Step 204 Challenge message Step 205 206 Response message 207, 208 process steps 300 consumer goods arrangement 301 Consumer goods 302 consuming device 303 Host Authentication Circuit 304 Consumer Goods Authentication Circuit 305 Control unit 306 Sensor 307 Classification circuit 400 Flowchart 401 Consumer Goods Authentication Circuit 402 Host Authentication Circuit Step 403 404 Challenge message Step 405 406 Response message 407-413 Procedure steps 414 classifiers 415 Sensor 500 arrangement 501 Consumer Goods Authentication Circuit 502 consuming device 503 control chip 504 Power supply 505 Analog / Digital Converter 600 arrangement 601 Consumer Goods Authentication Circuit 602 consuming device 603 Control chip 604 Power supply 605 Analog / Digital Converter 700 arrangement 701 Consumer Goods Authentication Circuit 702 consuming device 703 control chip 704 Application Processor 800 arrangement 801 Consumer Goods Authentication Circuit 802 consuming device 803 control chip 804 Microcontroller 805 Analog / Digital Converter 806 Classification software 900 consumer device 901 Sensor 902 authentication chip 903 Consumption component 904 Authentication circuit 905 Machine Learning Model 1000 Flowchart 1001 process steps

Claims

[1] Consumption device (900), comprising: A sensor (901) configured to acquire sensor data describing the physical behavior of an authentication chip (902) of a consumption component (903); A sender that is configured to send a request for authentication information to the consumption component (903); A receiver that is configured to receive authentication information from the consumption component; wherein the sender and receiver are configured to send and receive the request and authentication information via at least one first communication channel; and wherein the sensor (901) is configured to acquire sensor data representing information transmitted by the consumption component via a communication channel other than the at least one first communication channel; An authentication circuit (904) is configured to determine whether the authentication information corresponds to a predefined reference authentication information; and to implement a machine learning model (905) that is trained to classify consumption components into originals and imitations based on sensor data describing the physical behavior of authentication chips (902) of the consumption components; to feed the acquired sensor data to the machine learning model (905); to determine whether the machine learning model (905) classifies the consumption component (903) as an original; and, depending on whether both the authentication information corresponds to a predefined reference authentication information and whether the machine learning model (905) classifies the consumption component (903) as an original, to authorize the use of the consumption component (903) by the consumption device (900). [2] Consumption device (900) according to claim 1, wherein the sensor (901) is configured to acquire sensor data describing the physical behavior of the authentication chip (902) of the consumption component (903) when generating authentication information requested by the consumption component (903). [3] Consumption device (900) according to claim 1 or 2, wherein the sensor (901) is configured to acquire sensor data that describes the physical behavior of the authentication chip (902) of the consumption component (903) during memory accesses. [4] Consumption device (900) according to any one of claims 1 to 3, wherein the sensor data are side channel measurement data. [5] Consumer device (900) according to any one of claims 1 to 4, wherein the sensor (901) is configured to detect a voltage of the authentication chip (902), a current absorbed by the authentication chip (902), an electromagnetic field emitted by the authentication chip (902) and / or other side-channel data, such as temperature, electromagnetic radiation and light. [6] Consumption device (900) according to any one of claims 1 to 5, wherein the machine learning model (905) is a neural network. [7] Consumption device (900) according to any one of claims 1 to 6, wherein the consumption device (900) is a printer, an evaporator, a battery-operated electrical device or a vehicle. [8] Consumption device (900) according to any one of claims 1 to 7, wherein the sensor (901) comprises at least one sensor from a group consisting of: an optical sensor; a magnetic sensor; an inductive sensor; a capacitive sensor; a temperature sensor; a current sensor; a voltage sensor; an accelerometer; a vibration or oscillation sensor; and a chemical sensor. [9] Consumption device (900) according to any one of claims 1 to 8, comprising a power supply which is configured to supply the consumption component (903) with energy via a power supply line, wherein the physical behavior is a power consumption behavior of the consumption component via the power supply line. [10] Consumption device (900) according to claim 9, wherein the sensor (901) for detecting the power consumption of the consumption component is coupled to the power supply line via the power supply line. [11] Consumption device (900) according to any one of claims 1 to 10, wherein the sensor (901) is configured to detect an analog measurement signal and comprising an analog / digital converter configured to convert the analog measurement signal into digital sensor data and supply it to the authentication circuit (904). [12] Method for authenticating a consumption component (903) at a consumption device (900), wherein the method comprises: Request and receive authentication information from the consumption component (903); Determine whether the authentication information matches a given reference authentication information; Acquiring sensor data that describe the physical behavior of an authentication chip (902) of a consumption component (903); wherein the authentication information of the consumption component (903) is requested and received via at least one first communication channel; and where the sensor data are acquired via a communication channel other than the at least one first communication channel; Feeding the acquired sensor data to a machine learning model (905) that is trained to classify consumer components into originals and imitations based on sensor data describing the physical behavior of authentication chips (902) of the consumer components; Determine whether the consumption component (903) is classified as an original consumption component by the machine learning model (905); and Authorizing the use of the consumption component (903) by the consumption device (900) depending on whether both the authentication information matches a specified reference authentication information and whether the machine learning model (905) classifies the consumption component (903) as original. [13] Method according to claim 12, further comprising training the machine learning model (905): Acquire, for each consumption component of a plurality of consumption components, each classified as an original consumption component or as a counterfeit consumption component, sensor data describing a physical behavior of an authentication chip (902) of the consumption component; and Training the machine learning model (905) using the sensor data acquired for the multitude of consumption components and the classification of the multitude of consumption components as training data. [14] Method according to claim 13, wherein the training of the machine learning model (905) is carried out by adjusting parameters of the machine learning model (905) such that the ability of the machine learning model (905) to predict, based on the respective acquired sensor data, whether a consumption component of the plurality of consumption components is classified as original consumption components or imitation consumption components is increased.

Images