Classification of objects in the vicinity of an NFC reader device

Abstract

A method for determining a type of an object arranged in a radio frequency, RF, field emitted by an NFC reader device, in particular by an NFC-capable reader device, is disclosed. The object type is from a predetermined group of object types. The group of object types comprises at least one of an NFC tag, a metallic object and no load for any object. The method involves analyzing an oscillation behavior in the NFC reader device after the RF field emitted by the reader has been switched off using an M-degree decomposition scheme that decomposes an attenuated signal trace into M superposed components. The method relies on a training data set being accessible. Therefore, a training data set for a plurality of training objects from a training group of object types, in particular for NFC-capable reader devices, is additionally disclosed, in particular in a database.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the classification of objects in the vicinity of a near field communication (NFC) reader device, in particular to a method for determining the type of an object arranged in a radio frequency (RF) field emitted by a NFC reader device, and a NFC reader device capable of performing this method. The present disclosure further relates to a method of gathering a training dataset characterizing signals generated by bringing a training object from a training population of object types into the RF field of a NFC reader device, and a NFC reader device capable of performing this method. The present disclosure further relates to a machine readable non-transitory storage medium storing a computer program product or a computer program product comprising instructions which, when executed, are configured to control or carry out the method. BACKGROUND

[0002] Near field communication (herein: NFC) is an international transmission standard which is based on radio frequency identification (RFID) technology for the contactless exchange of data by means of electromagnetic induction by means of loosely coupled antenna coils over a short distance of a few centimeters and a normal data transmission rate of up to 848 kilobits per second. To date, this technology has been used primarily in the field of small-value payments, i.e. small (monetary) value contactless payments, and also in the field of toll and car access systems using, for example, special key tokens or smartphones with NFC functionality.

[0003] Figure 1 A conventional NFC arrangement is shown which comprises a polling NFC device which operates as a NFC reader device 100 and a listening NFC device which can also be referred to as a NFC tag 150, for example a smart card, a smartphone or a tablet computer with NFC functionality.

[0004] In the normal operating mode of the NFC arrangement shown in Figure 1 In the normal operating mode of the NFC arrangement shown in

[0005] To provide this functionality, the NFC reader device 100 includes an NFC device 120, which in turn includes a receiver 122, a processing and evaluation unit (P&E unit) 124, and a transmitter 126; an antenna interface 128; and an antenna 130, typically implemented as an antenna coil. The antenna interface 128 and the receiver 122 may be referred to as the front end of the device 100. When the NFC reader device 100 is in transmit mode, a signal stream travels as signal 134 from the P&E unit 124 to the transmitter 1126, further as signal 135 from the transmitter 126 to the antenna interface 128, further as signal 136 from the antenna interface 128 to the antenna 1130, and then enters the air from the antenna 1130 as a transmitted RF signal (not shown) associated with a transmitted RF field 140. When the NFC reader device 100 is in receive mode, the signal stream begins to travel from the RF signal (not shown) received by the antenna 130, travels from the antenna 130 to the antenna interface 128 as the received RF input signal 131, further travels from the antenna interface 128 to the receiver 122 as the signal 1132, and further travels from the receiver 122 to the P&E unit 124 as the received signal 133 after front-end processing.

[0006] like Figure 1 As shown, the NFC arrangement including polling device 100 and listening device 150 is "symmetrical" in the sense that polling device 100 is configured to transmit data to and be able to receive data from listening NFC device 150 via RF field 140, and also in the sense that listening NFC device 150 is configured to transmit data to and be able to receive data from polling NFC device 100 via RF field 140.

[0007] To achieve symmetry in the NFC arrangement including polling device 100 and listening device 150, the listening NFC device, i.e., the NFC tag 150, can be implemented functionally similar to polling device 100, and therefore may include an NFC device, which in turn includes a receiver 162, a P&E unit 164, and a transmitter 166; an antenna interface 168; and an antenna 170, which is typically implemented as an antenna coil. When the NFC tag 150 is in receiving mode, the signal flow begins from the RF signal (not shown) received by the antenna 170, travels from the antenna 170 to the antenna interface 168 as the received RF input signal 171, further travels from the antenna interface 168 to the receiver 162 as signal 172, and further travels from the receiver 162 to the P&E unit 1164 as the received signal 173 after front-end processing. When the NFC tag 150 is in transmit mode, the signal stream travels from the P&E unit 164 to the transmitter 166 as signal 174, further travels from the transmitter 166 to the antenna interface 168 as signal 175, further travels from the antenna interface 168 to the antenna 170 as signal 176, and then enters the air from the antenna 170 as the transmitted RF signal (not shown) associated with the transmitted RF field (not shown).

[0008] Each of the NFC devices in polling NFC device 100 and listening NFC device 150 may be a portable device, such as a smartphone or tablet computer with NFC functionality as explained above, while polling NFC device 100 may be a stationary device and operate as an NFC reader device 100, such as a card reader (wherein the expression "reader" refers to the ability of a card reader to read portable device-specific data from listening device or NFC tag 150).

[0009] NFC communication is typically initiated by polling the NFC device 100, for example, by an NFC reader device 100. Upon successful reception, the listening NFC device, i.e., the NFC tag 150, responds with a corresponding answer. The NFC tag 150 responds using active or passive load modulation techniques. When coupled via an RF field, such as RF field 140, devices 100 and 150 can move within a few centimeters of proximity, causing arbitrary coupling between the antennas 130 and 170 of devices 100 and 150. This, in turn, produces different received RF input signals 131, varying in amplitude and phase depending on the distance and orientation of the NFC tag 150 relative to the NFC reader device 100.

[0010] NFC devices, such as Figure 1 The NFC reader device 100 shown operates as a reader, requiring periodic polling of NFC tags placed nearby, such as... Figure 1The polling mechanism causes a non-negligible current consumption in the reader, which becomes a problem in case of battery-powered devices. Therefore, battery-powered NFC reader devices usually provide a so-called low power card detection (LPCD) feature, which avoids the need to periodically poll NFC tags. If the LPCD is more energy-efficient than the polling mechanism, the total current consumption can be reduced. State-of-the-art LPCD implementations are usually highly sensitive to metal objects entering the RF field transmitted by the reader, i.e. can be detected as NFC tags.

[0011] US 9, 124,302 B2 discloses a system and method for carrier frequency variation for device detection in near field communication. A parameter is measured while a transmitter applies a signal with a certain frequency to the matching network of a device. The frequency is varied and a peak of the measured parameter is found (corresponding to the resonance of the whole system). If the peak exceeds a certain threshold, this is interpreted as device detection.

[0012] US 6,476,708 B1 discloses a method for detecting RFID devices by an RF reader unit operating in a power-reduced state. The method involves detecting RFID tags based on decay time measurement values.

[0013] US 10,511,347 B2 discloses a method for device detection in a contactless communication system. The method involves tracking changes in the output of an oscillator when it oscillates at the boundaries of its stability. The problem of confused metal objects and NFC tags is solved by doing periodic polling.

[0014] US 9,819,401 B2 discloses a highly selective low power card detector for near field communication (NFC) and method of operation, involving detecting a reflective impedance by using an oscillator excitation system and comparing the amount of oscillation of the oscillator to a threshold value. SUMMARY

[0015] It is a general object of the present disclosure to provide a method of increasing the efficiency of a low power card detection (LPCD) method. This can generally be achieved according to embodiments of the present disclosure by determining the type of object that has entered the RF field transmitted by a NFC reader device or NFC-capable reader device, and in particular, distinguishing NFC tags from other non-relevant objects in the RF field emitted by the NFC reader device or NFC-capable reader device.

[0016] In particular, it is an object of the present disclosure to provide a method of distinguishing NFC tags from other non-relevant objects.

[0017] According to an embodiment example of the present disclosure, this is achieved by analyzing the oscillation behavior of the NFC arrangement in particular in a NFC reader device after the RF field emitted by the reader has been switched off.

[0018] These objects are achieved by the subject matter having the features according to the independent patent claims. In particular, this object is achieved by a method for determining a type of an object arranged in a radio frequency, RF, field emitted by a near field communication, NFC, reader device according to the attached independent claim 1, and by a NFC reader device capable of performing said method according to the attached independent claim 14. The method for discriminating relies on an accessible training data set. Therefore, a method for acquiring a training data set for a plurality of training objects for a NFC reader device according to the attached independent claim 11 is also provided, as well as a NFC reader device capable of performing said acquisition method according to the attached independent claim 14. Furthermore, a machine readable non-transitory storage medium, or computer program product, storing a computer program product comprising instructions configured to control or perform a method for discriminating and / or an acquisition method according to the attached independent claim 15 is provided. Further embodiment examples of the present disclosure are described in the dependent claims.

[0019] According to a first aspect of the present disclosure, a method for determining a type of an object arranged in a radio frequency, RF, field emitted by a near field communication, NFC, reader device, in particular by a NFC-capable reader device, is provided, wherein the object type is from a predetermined group of object types. The NFC reader device has an RF antenna for emitting the RF field and for receiving an RF input signal, and a received signal processing chain for front-end processing the received RF input signal and for providing a front-end processed received signal x corresponding to the received RF input signal as an output of the received signal processing chain. The method has the following steps:

[0020] a) for an object being in the RF field emitted from the NFC reader device, measuring and recording an attenuated signal trace of the front-end processed received signal generated by the object (150, 180, 190) in the RF field after the emitted RF field has been switched off;

[0021] b) estimating an M-degree decomposition of the recorded attenuated signal trace generated by the object into M superimposed components of a predetermined decomposition scheme according to the predetermined decomposition scheme, and in turn, for each superimposed component, determining and in particular storing an associated set of estimated characteristic parameters of the superimposed component, wherein each of the M superimposed components is defined by a predetermined superposition function which in turn is determined by an associated set of characteristic parameters;

[0022] c) based on at least one of the determined estimated characteristic parameters of the M superimposed components according to the decomposition scheme, attempting to classify at least one of the characteristic parameters with respect to a training data set of training characteristic parameters of the M superimposed components; and, in particular,

[0023] d) if the classification of the M estimated characteristic parameters is successful, determining the object type the decay signal traces recorded in step a) are directed to based on the successful classification.

[0024] The operation of estimating a decomposition into M oscillating components and attempting a classification based on the stored M estimated characteristic parameters of the M oscillating components according to the decomposition scheme not only enables a determination of the presence or the past presence of an object in the emitted RF field but even enables a determination of the type of the object present or past present in the emitted RF field.

[0025] It should be noted that whether different object types can be distinguished, in other words whether the classification can be successful, depends on the quality of the training data set, the selection of the decomposition scheme and the distance and relative orientation of the object to the RF antenna, etc.

[0026] According to a second aspect of the present disclosure, a method of acquiring, in particular in a database, a training data set for an NFC reader device, in particular for an NFC-capable reader device, for a plurality of training objects from a training population of object types is provided, wherein the training data comprises a plurality of recorded training decay signal traces for each of at least one training object selected from a type of the training population of object types. The NFC reader device has an RF antenna for emitting an RF field and for receiving an RF input signal and a received signal processing chain for front-end processing the received RF input signal and for providing a front-end processed received signal x corresponding to the received RF input signal as an output of the received signal processing chain. The method has the following steps:

[0027] 1) pre-determining a training population of object types;

[0028] 2) pre-determining an M-degree decomposition scheme of decomposing a decay signal trace into M superimposed components, and storing the decomposition scheme and an indication of the M predetermined superposition functions, in particular in a database accessible to the P&E unit, wherein each of the M superimposed components is defined by a predetermined superposition function, which in turn is determined by an associated set of characteristic parameters;

[0029] 3) for each predetermined object type:

[0030] 3.1) providing a plurality of different training objects of the predetermined object type,

[0031] 3.2) bringing each of the provided plurality of different training objects sequentially into the RF field emitted by the NFC reader device,

[0032] 3.3) for each training object being in the RF field emitted by the NFC reader device, performing the following steps:

[0033] 3.3.1) after switching off the RF field of the NFC reader device, measuring and recording in a database, in particular, a training attenuation signal trace of the front-end processed received signal produced by the training object in the RF field,

[0034] 3.3.2) estimating an M-degree decomposition of the recorded training attenuation signal trace into M superposed components according to a predetermined decomposition scheme, and in turn, for each superposed component, determining an associated set of determined characteristic training parameters of the superposed component of the estimated decomposition of the training attenuation signal trace produced by the training object in the RF field and in particular storing in a database.

[0035] According to a third aspect of the present disclosure, there is provided an NFC reader device, in particular, an NFC-capable reader device, configured to perform the method according to the first aspect of the present disclosure and / or configured to perform the method according to the second aspect of the present disclosure.

[0036] According to a fourth aspect of the present disclosure, there is provided a machine- readable non-transitory storage medium storing a computer program product, or a computer program product comprising instructions which, when executed on, for example, a processor, microprocessor, or computer-controlled data processing system, perform the method according to the first aspect of the present disclosure and / or the method according to the second aspect of the present disclosure.

[0037] In an embodiment of the method according to the first aspect, the group of object types comprises at least one of NFC tags, metal objects, and no load for any object. In a preferred embodiment, the group of object types comprises at least one of NFC tags, objects made of copper, objects made of aluminum, objects made of steel, and no load for any object; that is, for metal objects, it can be distinguished whether the metal is one of copper, aluminum, and steel.

[0038] In an embodiment of the method according to the first aspect, the NFC reader device can have a processing and evaluation, P&E, unit for processing and evaluating the front-end processed received signal, wherein in step a) the attenuation signal trace is recorded and stored in the P&E unit, and wherein steps b), c), and d) are performed in the P&E unit.

[0039] In an embodiment of the method according to the first aspect, the training data set can be stored in a database accessible to the P&E unit and can have been obtained during a training phase for the NFC reader device using a plurality of training objects from each object type in a predetermined group of object types.

[0040] In an embodiment of the method according to the first aspect, the measuring and recording of the decay signal trace in step a) can comprise recording the time dependence of the decay signal trace during a time interval starting at or after the time of the switch-off of the RF field, in particular after a predetermined time delay after the time of the switch-off of the RF field. Here, the decay signal values are complex values.

[0041] In an embodiment of the method according to the first aspect, the estimating of the M- degree decomposition according to the predetermined decomposition scheme in step b) comprises fitting the M-degree decomposition to the recorded decay signal trace by varying the M estimated characteristic parameters of the estimated decomposition.

[0042] In a particular embodiment of the method according to the first aspect, the measuring and recording of the decay signal trace in step a) comprises recording the time dependence of the decay signal trace (250, 260; 612, 622) in the form of a time series of signal values x[n], where n denotes a discrete time index, where x[n] is a complex value and has:

[0043] - a real part real(x[n]) and an imaginary part imag(x[n]), and / or

[0044] - an absolute value |x[n]| and an angle ∠x[n].

[0045] In an embodiment of this particular embodiment of the method, the estimating of the M- degree decomposition of the recorded decay signal trace in step b) can comprise modeling the time series of signal values x[n] as a sum of M superimposed components x m [n] plus a term d[n] representing white noise according to:

[0046] where n e [0, N - 1], (1)

[0047] where N is the number of measured signal samples, n is the discrete time index corresponding to the measured signal samples, m is an index denoting the superimposed components x m [n], where m = [1, M], M is the number of superimposed components x m [n] in the decomposition, and n, N, m and M are integers.

[0048] In this embodiment, the M superimposed components x m[n] represents the decomposition of the decay signal trace into a set of superposition functions. This set can be selected from the group consisting of: a set of oscillatory functions, in particular a set of weighted oscillatory functions; a set of sinusoidal functions, in particular a set of weighted sinusoidal functions; a set of impulse functions, in particular a set of weighted impulse functions; and a set of step functions, in particular a set of weighted step functions.

[0049] In another embodiment of the above particular embodiment of the method, the M-degree decomposition of the recorded decay signal trace in step b) can comprise modeling the time series of signal values x[n] as M oscillatory components wherein the sum of terms d[n] represents white noise:

[0050] wherein n e [0, N - 1], (2)

[0051] wherein N is the number of measured signal samples, n is the discrete time index corresponding to the measured signal samples, m is an index denoting a weighted oscillatory component wherein m = [1, M], M is the number of weighted oscillatory components admitted in the decomposition, and n, N, m and M are integers. Here, each superposition component can be represented as a weighted oscillatory component wherein z m is a complex-valued oscillatory parameter according to a predetermined decomposition scheme, denotes an oscillatory function associated with z m and is the nth value of the oscillatory function corresponding to the discrete time index n, and a m is a complex-valued weight parameter.

[0052] In this further embodiment, the oscillatory component m as a function of time is derived by

[0053]

[0054] In this further embodiment, the complex-valued oscillatory parameter z m can be mapped to the oscillatory function at the discrete time index n by

[0055]

[0056] Further in this further embodiment, the linear frequency f m of the oscillatory component m and the argument ∠z m are connected as follows:

[0057]

[0058] wherein f s denotes the sampling frequency.

[0059] In this further embodiment, each weighted oscillation component may be represented by a subset of characteristic parameters. This subset can comprise the absolute value of the weight parameter, i.e. |a m ; the angle of the weight parameter, i.e. ∠a m ; the absolute value of the oscillation parameter, i.e. |z m ; the angle of the oscillation parameter, i.e. ∠z m . Here, each subset of characteristic parameters can be summarized as a vector θ m :

[0060]

[0061] where · T denotes the transposed vector, and

[0062] where the parameter matrix Θ is defined from the vectors θ m as:

[0063] Θ = [θ1 θ2... θ M ], (7)

[0064] such that the parameter matrix Θ contains the complete set of characteristic parameters for all M superimposed components.

[0065] Further in this further embodiment, the absolute value of the respective weight parameter can correspond to a scaling of the oscillation component, and the angle of the respective weight parameter can correspond to a phase shift of the oscillation component. Yet further in this further preferred embodiment, the absolute value of the oscillation parameter can correspond to a damping of the oscillation function, and the angle of the oscillation parameter can correspond to a frequency of the oscillation function.

[0066] Further in this further embodiment, the attempt to classify the M characteristic parameters with respect to the training data set can comprise performing a classification of a subset of characteristic parameters comprising at least one of: the absolute value of the weight parameter, the angle of the weight parameter. Further, the subset of characteristic parameters comprises, with respect to the corresponding training set of characteristic parameters included in the training data set, also the absolute value of the oscillation parameter and the angle of the oscillation parameter, for M = 3, i.e. for m = 1, 2 and 3, preferably for M = 2, i.e. for m = 1 and 2, and more preferably for M = 1, i.e. for m = 1.

[0067] In an embodiment of the method according to the first aspect, the estimating of the M-degree decomposition of the recorded decay signal trace and determining of the associated set of estimated characteristic parameters in step b) comprises fitting the decomposition of M degrees to the recorded decay signal trace by varying the characteristic parameters of the associated set of estimated characteristic parameters of the superimposed components of the estimated decomposition, wherein the determining of the associated set of characteristic parameters involves a mathematical method selected from the group of Prony’s method, the function bundle method and the (total) least squares estimator;

[0068] In an embodiment of the method according to the first aspect, the attempting to classify in step c) can involve a mathematical method selected from the group consisting of: Bayesian classification, Support Vector Machines, and Artificial Neural Networks.

[0069] In an embodiment of the method according to the second aspect of the disclosure, the object type training group comprises at least one of: NFC tags, metal objects, and no load for any object,

[0070] In an embodiment of the method according to the second aspect of the disclosure, the NFC reader device can have a processing and evaluation, P&E, unit for processing and evaluating the received signal subjected to the front-end processing.

[0071] In an embodiment of the method according to the second aspect, the NFC reader device has a database accessible to the P&E unit.

[0072] In an embodiment of the method according to the second aspect, in step 2) the decomposition scheme and the indication of the M predetermined superposition functions are stored in the database accessible to the P&E unit, wherein in step 3.3.1) the training decay signal trace is stored in the database, and wherein in step 3.3.2) the characteristic training parameters are stored in the database.

[0073] In an embodiment of the method according to the second aspect, the measuring and recording in the database in step 3.3.1) can comprise recording the time dependency of the training decay signal trace in the form of a time series of training signal values x[n], where n denotes a discrete time index, where x[n] is a complex value and has:

[0074] - a real part real(x[n]) and an imaginary part imag(x[n]), and / or

[0075] - an absolute value |x[n]| and an angle ∠x[n].

[0076] In an implementation of this particular embodiment of the method according to the second aspect, the estimating of the M-degree decomposition of the recorded decay signal trace in step b) can comprise modeling the time series of signal values x[n] as a sum of M superposition components x m [n] plus a term d[n] representing white noise according to:

[0077] where n e [0, N - 1], (1')

[0078] where N is the number of measured signal samples, n is a discrete time index corresponding to the measured signal samples, and m is an index denoting a superposition component x m [n], where m = [1, M], M is the number of superposition components x mThe number of [n], where n, N, m, and M are integers.

[0079] In this implementation scheme, there are M superimposed components x m [n] can represent the decomposition of the decaying signal trace into a set of superimposed functions. This set can be freely selected from the following clusters: oscillatory function set, specifically, weighted oscillatory function set; sine function set, specifically, weighted sine function set; impulse function set, specifically, weighted impulse function set; and step function set, specifically, weighted step function set.

[0080] In another embodiment of the specific embodiment described above according to the second aspect of the method, estimating the predetermined M-degree decomposition in step 3.3.2) may include: modeling x[n] as the sum of M superimposed components plus a term d[n] representing white noise according to the following formula.

[0081] Where n∈[0, N-1], (2′)

[0082] Where N is the number of training signal samples measured, n is the discrete-time index corresponding to the measured training signal sample, and m is the identifier of the weighted oscillation component. The index is given by , where m = [1, M], M is the number of weighted oscillatory components allowed in the decomposition, and n, N, m, and M are integers. Here, each superimposed component can be represented as a weighted oscillatory component. Where zm is a complex-valued oscillatory function based on a predetermined decomposition scheme. Let a represent the oscillation function associated with zm, and be the nth value of the oscillation function corresponding to the discrete-time index n, and a m It is a complex-valued weight parameter.

[0083] In this alternative implementation, the complex-valued oscillation parameter zm can be mapped to the oscillation function of the discrete-time index n using the following formula:

[0084]

[0085] In this alternative implementation, the time-varying weighted oscillation component can be obtained by the following formula.

[0086]

[0087] In another embodiment, the weighted oscillation components m and ∠z m linear frequency f m Connected as follows:

[0088]

[0089] Where f s Indicate the sampling frequency.

[0090] In this further embodiment, each weighted oscillation component may be represented by a subset of characteristic parameters. This subset can comprise the absolute value of the weight parameter, i.e. |a m |; the angle of the weight parameter, i.e. ∠a m ; the absolute value of the oscillation parameter, i.e. |z m |; the angle of the oscillation parameter, i.e. ∠z m Here, each subset of characteristic parameters can be summarized in a vector θ m :

[0091]

[0092] wherein. T denotes a transposed vector, and

[0093] wherein the parameter matrix Θ is derived from the vector θ m is defined as:

[0094] Θ = [θ1 θ2... θ M ], (7')

[0095] such that the parameter matrix Θ contains the complete set of characteristic parameters for all M superimposed components.

[0096] Further in this further embodiment, the absolute value |a m | of the respective weight parameter can correspond to a scaling of the oscillation component, and the angle ∠a m of the respective weight parameter can correspond to a phase shift of the oscillation component. Yet further in this further preferred embodiment, the absolute value |z m | of the oscillation parameter can correspond to a damping of the oscillation function, and the angle ∠z m of the oscillation parameter can correspond to a frequency of the weighted oscillation component . BRIEF DESCRIPTION OF DRAWINGS

[0097] In the following, exemplary embodiments of the present disclosure are described in detail with reference to the accompanying drawings, in which:

[0098] Figure 1 showing an arrangement of an NFC reader device emitting an RF field and an alternative object type arranged in the emitted RF field.

[0099] Figure 2 comprising Figure 2 (a) and 2(b), Figure 2 (c) and 2(d), and Figure 2 (e) and 2(f), showing respective pairs of diagrams showing the real and imaginary parts of the following respective functions:

[0100] Figure 2 (a) and 2(b) show, respectively, Figure 2 the real and imaginary parts of the first superimposed component of the decomposition of the exemplary damped signal trace shown in

[0101] Figure 2 (c) and 2(d) show, respectively, Figure 2 the real and imaginary parts of the second superimposed component of the decomposition of the exemplary damped signal trace shown in

[0102] Figure 2 (e) and 2(f) show, respectively, the real and imaginary parts of an exemplary damped signal trace, which can be represented as Figure 2 the sum of the first superimposed component shown in Figure 2 (c) and 2(d) and

[0103] Figure 3 comprising Figure 3 (a) and 3(b), Figure 3 (c) and 3(d), and Figure 3 (e) and 3(f), show respective pairs of diagrams showing a representation of a complex-valued oscillation parameter in the complex plane and a representation of an associated oscillation function as a function of a discrete time index, as follows:

[0104] Figure 3 (a) and 3(b) show, respectively, a representation of a first exemplary complex-valued oscillation parameter having a first damping and a first frequency in the complex plane and a representation of an associated first exemplary oscillation function as a function of a discrete time index.

[0105] Figure 3 (c) and 3(d) show, respectively, a representation of a second exemplary complex-valued oscillation parameter having a first damping and a second frequency which is double the first frequency in the complex plane and a representation of an associated second exemplary oscillation function as a function of a discrete time index.

[0106] Figure 3 (e) and 3(f) show, respectively, a representation of a third exemplary complex-valued oscillation parameter having a second damping which is half the first damping and a second frequency which is double the first frequency in the complex plane and a representation of an associated third exemplary oscillation function as a function of a discrete time index.

[0107] Figure 4 a flowchart showing a method for determining a type of an object arranged in an RF field emitted by an NFC reader device according to an embodiment of the first aspect of the present invention.

[0108] Figure 5A flow chart illustrating a method for obtaining a training data set for an NFC reader device for a plurality of training subjects from a group of training subjects according to an embodiment of the second aspect of the application.

[0109] Figure 6 The left-hand side diagram of Fig. 6 illustrates the output of the analog-to-digital converter (ADC) in the I channel of the received signal processing chain, ADC I, and its M = 3 degree decomposition approximation.

[0110] Figure 6 The right-hand side diagram of Fig. 6 illustrates the output of the analog-to-digital converter (ADC) in the Q channel of the received signal processing chain, ADC Q, and its M = 3 degree decomposition approximation.

[0111] Figure 6 The right-hand side diagram of Fig. 6 illustrates the output of the analog-to-digital converter (ADC) in the Q channel of the received signal processing chain, ADC Q, and its M = 3 degree decomposition approximation.

[0112] Fig. 7 comprises Figure 7A 、 7B and 7C, illustrating a respective diagram triplet, respectively illustrating the M = 3 degree decomposition oscillation parameters fitted to the respective plurality of measured and recorded attenuated signal traces using the functional bundle method for a plurality of unloaded RF fields, for a plurality of RF fields loaded with NFC tags, and for a plurality of RF fields loaded with metallic objects.

[0113] Figure 7A Fig. 8 illustrates a diagram triplet, respectively illustrating the first oscillation parameter z1 of the M = 3 superimposed components obtained from fitting an M degree decomposition to a plurality of measured and recorded training attenuated signal traces for a plurality of training unloaded RF fields (left-most diagram), for a plurality of RF fields loaded with training NFC tags (middle diagram), and for a plurality of RF fields loaded with training metallic objects (right-most diagram).

[0114] Figure 7B Fig. 9 illustrates a diagram triplet, respectively illustrating the second oscillation parameter z2 of the M = 3 superimposed components obtained from fitting an M degree decomposition to a plurality of measured and recorded training attenuated signal traces for a plurality of training unloaded RF fields (left-most diagram), for a plurality of RF fields loaded with training NFC tags (middle diagram), and for a plurality of RF fields loaded with training metallic objects (right-most diagram).

[0115] Figure 7CA third oscillation parameter z3 of M=3 superimposed components obtained from the operation of fitting M degrees to a plurality of measured and recorded training attenuation signal traces, respectively for a plurality of training unloaded RF fields (leftmost graph), for a plurality of RF fields loaded with training NFC tags (middle graph), and for a plurality of RF fields loaded with training metal objects (rightmost graph).

[0116] Fig. 8 includes Figure 8A 、 8B and 8C, showing respective pairs of graphs, respectively showing absolute values of M=3 degrees respective decompositions first to third oscillation parameters and angles of said M=3 degrees respective decompositions first to third oscillation parameters, fitted to respective pluralities of measured and recorded attenuation signal traces, for a plurality of RF fields loaded with NFC tags varying with distance of the NFC tags to an antenna of an NFC reader device, and for a plurality of RF fields loaded with metal objects varying with distance of the metal objects to an antenna of an NFC reader device.

[0117] Figure 8A showing respective pairs of graphs, respectively showing absolute values of M=3 degrees respective decompositions first oscillation parameters z1 and angles of M=3 degrees respective decompositions first oscillation parameters z1.

[0118] Figure 8B showing respective pairs of graphs, respectively showing absolute values of M=3 degrees respective decompositions second oscillation parameters z2 and angles of M=3 degrees respective decompositions second oscillation parameters z2.

[0119] Figure 8C showing respective pairs of graphs, respectively showing absolute values of M=3 degrees respective decompositions third oscillation parameters z3 and angles of M=3 degrees respective decompositions third oscillation parameters z3.

[0120] Figure 9 showing graphs in respective complex planes, the graphs including graphs showing respective pluralities of oscillation parameters z1, z2, z3 of respective decompositions of training attenuation signal traces for a training tag and a plurality of oscillation parameters z1 of respective decompositions of training attenuation signal traces for a training metal plate, and graphs showing associated oscillation functions varying with discrete time indices in time domain.

[0121] Figure 10A respective set of complex plane diagrams is shown for a plurality of training NFC tags and for a plurality of training metal objects, a first set showing respective pluralities of oscillation parameters z1, z2, z3 for respective decompositions of training attenuation signal traces for a plurality of training tags arranged at four different distances (far field, 50 mm, 30 mm and 5 mm) in the RF field of an NFC reader device, and a second set showing respective pluralities of oscillation parameters z1, z2, z3 for respective decompositions of training attenuation signal traces for a plurality of training metal objects arranged at four different distances (far field, 50 mm, 30 mm and 5 mm) in the RF field of an NFC reader device.

[0122] For the sake of brevity, features described with respect to a particular figure can not be described again when presented in another figure, equally or similarly. DETAILED DESCRIPTION

[0123] Before describing example embodiments of the present disclosure with reference to the figures, some general aspects of the present invention proposed by the inventor of the present invention should still be explained.

[0124] One of the standards in the Near Field Communication (NFC) standards is given in the following document: ISO / IEC 14443-3: “Cards and security devices for personal identification - Contactless proximity objects - Part 3: Initialization and anticollision”. ISO, Fourth edition, Geneva, Switzerland, 2018.

[0125] In a standard NFC scenario, an NFC reader periodically switches between (i) standby and (ii) poll. During standby, the reader does not transmit. When polling, the reader transmits a command to activate cards located in its vicinity. The specific command is defined by the communication standard used (e.g. REQA or EEQB for ISO / IEC 14443). In addition, the duration of both standby and poll need to be chosen such that the timing requirements of this standard are met. Actual communication between a card and the reader is initiated by placing the card in the radio frequency (RF) field of the reader. Subsequently, the card is energized and responds to the reader poll command (e.g. ATQA or ATQB for ISO / IEC 14443).

[0126] Obviously, in the case of a battery-powered reader device, the current consumption caused by the polling mechanism is not negligible. Therefore, there are so-called low power card detection (LPCD) methods aiming at avoiding the need to periodically poll the card. The main idea behind LPCD is to avoid sending a poll command in case of no card in the reader's RF field. Therefore, before polling, the reader assesses whether a card is in its RF field or not. If a card is detected, the poll sequence is sent, if no card is detected, the reader does not need to poll and switches back to standby again. If the card detection mechanism consumes less current than sending a poll command, the total current consumption is reduced compared to the standard scenario.

[0127] The state-of-the-art LPCD method relies on analyzing the loading and detuning of the antenna by switching the RF field on and off. This operation is usually done as follows:

[0128] 1) After a full polling cycle, the unmodulated RF field is switched on and the LPCD is calibrated by setting the analog chain (including the HF attenuator, the baseband attenuator, the DC offset compensation) to its operating point.

[0129] 2) After calibration, the reader field is periodically switched on and off. Since the operating point of the analog chain is fixed according to the initialization, the change at the input (with respect to the previous measurement) is linked to a change in the loading conditions or detuning of the antenna. If this change exceeds a certain threshold, the polling procedure is triggered. This event shall be referred to as wake-up in the following.

[0130] Generally, the steady-state behavior of the system is analyzed to decide whether it is necessary to poll or not. The main drawback of this procedure is that any object interacting with the RF field of the reader (e.g. any metal block or the like) influences the loading conditions or detuning of the antenna. Therefore, while the state-of-the-art LPCD method can detect whether an object enters or is removed from the field of the reader, it usually cannot distinguish a NFC tag from e.g. a metal object. This means that a method that can distinguish a NFC tag from an irrelevant object can help to further reduce the current consumption of the reader device.

[0131] The average current consumption of the reader can be modeled as follows:

[0132]

[0133] where Istandbyrepresents the current consumption in standby or power-off mode, I standby represents the current consumption during "normal" operation (e.g. polling), t on is the duration of the phase in which the reader is in power-off, and t standby is the duration of the phase in which the reader polls or performs LPCD. on ​

[0134] To minimize the current consumption, the LPCD target is to minimize t on while ensuring reliable card detection. For further analysis of the LPCD mechanism, the duration t on can be further subdivided into two phases: t on，LPCD , in which the LPCD is performed, and t on，poll , in which a polling request is performed after a card is detected.

[0135]

[0136] Obviously, the wake-up due to e.g. a metal object placed in the field will increase t on，poll and the current consumption. A method that helps to send only polling requests in case a card is placed in the reader field will thus reduce the current consumption compared to the standard LPCD mechanism. This can be further shown by considering t on，poll as a result of the detection performance. As an extreme case, consider a duration t representing a situation in which any load change will cause the reader to send a polling request. This time will be scaled by the probability of wake-up P(wake-up) to obtain t on，poll :

[0137]

[0138] with the following probabilities:

[0139] P(wake-up | tag) the probability of wake-up in case a card / tag enters the field,

[0140] P(wake-up | metal) the probability of wake-up in case an unrelated object enters the field,

[0141] P(wake-up | unloaded) the probability of wake-up in case the loading condition does not change,

[0142] P(tag) the prior probability of a card / tag entering the field,

[0143] P(metal) the prior probability of a metal object entering the field, and

[0144] P(unloaded) the prior probability of exhibiting a measurement outlier of a load change,

[0145] The total probability of wake-up is obtained by:

[0146]

[0147] Considering the following, by setting P(unloaded) = 0, we obtain:

[0148]

[0149] Additionally, if it is assumed that no prior knowledge is available about the relative frequency of tags and metal entering the field of the reader, then P(tag) = P(metal) = 1 / 2 is set accordingly. This yields the final equation for the probability of wake-up:

[0150] P(wake-up) = (P(wake-up | tag) + P(wake-up | metal)) - 1 / 2 (13)

[0151] An optimal LPCD mechanism would detect any tag, i.e. P(wake-up | tag) = 1, while never triggering a wake-up in case a metal object is placed in the field, i.e. P(wake-up | metal) = 0. However, the state-of-the-art is rather characterized by P(wake-up | metal) = 1, which means that P(wake-up) = 1, and Equation (10) and equation (13) show that the duration t on，poll can be reduced by lowering P(wake-up | metal), and thus the corresponding current consumption.

[0152] According to the present disclosure, it is proposed to analyze the dynamic behavior of the whole system rather than its steady-state behavior. To this end, the decay curve after the RF field of the reader is switched off is analyzed to gather further information about the type of object currently affecting the loading and tuning conditions of the reader. The basic principle behind this strategy is that NFC tags and cards can be considered as "tuned objects", which means that their antennas are tuned to a specific resonance frequency. This allows to separate the tuned objects from "untuned" objects that do not exhibit a specific resonance frequency. While tuned and untuned objects can have similar (or at least indistinguishable) effects on some parameters of the system, according to the present disclosure, it is claimed that resonance effects can be observed after the reader field has been switched off and that these effects can be used to separate the two classes of objects. It is further shown that the proposed method can also be used to distinguish NFC tags from each other. Thus, the proposed mechanism can be used to enhance the performance of any other LPCD method.

[0153] Some key features of the present disclosure include:

[0154] • Objects in the field of the reader can be classified to reduce current consumption.

[0155] • A method of analyzing the decay behavior of an NFC system is provided.

[0156] • The method involves estimating the decay behavior of the NFC system by a small number of complex-valued oscillations.

[0157] The method may also provide the possibility of distinguishing certain NFC tags in a reader field.

[0158] In the following text, the signal received after the RF field is cut off should be denoted by x[n], where n denotes the discrete-time index. The problem in the discrete-time domain is formulated without loss of generality because the same observations and modeling assumptions can also be made in continuous time. The signal x[n] is complex. Specifically, its real part corresponds to the I-channel ADC output (x... I [n]) and its imaginary part corresponds to the Q-channel ADC output (x Q [n]). This disclosure is based on performing the decomposition of x[n], that is, modeling x[n] as M superimposed components x. m [n] Specifically, M is the damped (and / or weighted) complex-valued oscillation component. The sum of the white noise terms d[n] is given by the following formula:

[0159] Where n∈[0, N-1] (1)

[0160] Where N is the number of observed samples, a m This represents the complex-valued weights (i.e., scaling and phase shift) associated with component m, and z m These are the oscillation parameters that determine the time-domain shape of the signal. m The angle corresponds to the frequency of the corresponding oscillation component and z m The absolute value of corresponds to the damping of the corresponding oscillation component.

[0161] In this decomposition, the superimposed components It can be expressed as the time-varying weighted oscillation component obtained by the following formula.

[0162]

[0163] The complex oscillation parameter z can be expressed by the following formula. m Oscillation function mapped to discrete-time index n:

[0164]

[0165] Oscillating components m and ∠z m linear frequency f m Connected as follows:

[0166]

[0167] Where f s Indicate the sampling frequency.

[0168] exist Figure 2This decomposition concept is illustrated in Figs. 1 to 2, which show exemplary oscillation components x m [n] and their reconstruction from the individual components.

[0169] Figure 2 comprising Figure 2 (a) and 2(b), Figure 2 (c) and 2(d), and Figure 2 (e) and 2(f). Thus, Figure 2 showing pairs of corresponding diagrams, each pair showing the real and imaginary parts of the following corresponding functions. Figure 2 (a) and 2(b) show Figure 2 the real part 210 and the imaginary part 220 of the first superposition component of the decomposition of the exemplary decay signal trace shown in (e) and 2(f). Figure 2 (c) and 2(d) show Figure 2 the real part 230 and the imaginary part 240 of the second superposition component of the decomposition of the exemplary decay signal trace shown in (e) and 2(f). Figure 2 (e) and 2(f) show the real part 250 and the imaginary part 260 of the exemplary decay signal trace x[n], respectively, as Figure 2 the sum of the first superposition component 210, 220, i.e. x1[n], shown in (a) and 2(b) and Figure 2 the second superposition component 230, 240, i.e. x2[n], shown in (c) and 2(d).

[0170] Figure 2 The decay signal trace 250, 250 shown in (e) and 2(f) is an exemplary decay signal trace in the sense that it represents a synthetically generated signal trace, but estimates a decay signal trace that has been measured for an NFC tag in an NFC reader device in an experimental setup. The decomposition into the first superposition component and the second superposition component has been obtained for the following set of characteristic parameters:

[0171] [a1] [a2] ​ ​ 150e iπ / 4 ]] 190e -iπ / 3 ]]> 0,8e -iπ / 2 ]] 0,6e iπ / 4 ]]>

[0172] According to the decomposition concept of the present disclosure, in addition, each superposition component x m [n] can be represented by a set of characteristic parameters from the associated weighted oscillation components stacked into a vector Θ m :

[0173]

[0174] wherein. T denotes a transposed vector. Furthermore, the parameter matrix is defined as:

[0175] Θ = [θ1 θ2 ... θ] M (7″)

[0176] It contains all the characteristic parameters of the superimposed components. See below. Figure 3 z is shown m For x m The influence of [n].

[0177] Figure 3 include Figure 3 (a) and 3(b), Figure 3 (c) and 3(d) and Figure 3 (e) and 3(f), and therefore, corresponding pairs of diagrams are shown, each pair showing a representation of the complex-valued oscillation parameters in the complex plane and a representation of the associated oscillation function varying with the discrete-time index. Specifically, Figure 3 (a) shows a representation 310 of the first exemplary complex-valued oscillation parameter z1, which can be expressed according to equation (4″) as follows: In the complex plane, the first value of the oscillating function for n=1 has a first damping |z1| and a first frequency ∠z1. Figure 3 (b) shows a representation 320 of an associated first exemplary oscillation function having a real part 322 and an imaginary part 324 that varies with the discrete-time index n. Figure 3 (c) shows a representation 330 of the second exemplary complex-valued oscillation parameter z2, which can be expressed according to equation (4″) as follows In the complex plane, for the first value of the oscillating function n=1, there is a second damping |z2|=|z1| equal to the first damping and a second frequency ∠z2=2∠z1 that is twice the first frequency. Figure 3 (d) shows a representation 340 of an associated second exemplary oscillation function with a real part 342 and an imaginary part 344 that varies with the discrete-time index n. Figure 3 (e) shows the representation 350 of the third exemplary complex-valued oscillation parameter z3, which can be expressed according to equation (4″) as follows In the complex plane, the first value of the oscillating function for n=1 has a third damping equal to half of the first damping. The third frequency, ∠z3, is twice the first frequency and equal to 2∠z1. Figure 3 (f) shows a representation 360 of an associated third exemplary oscillation function having a real part 362 and an imaginary part 364 that varies with the discrete-time index n.

[0178] In summary, Figure 3 The characteristic parameters of the oscillation parameters z1, z2, and z3 shown are set as follows:

[0179] a ​ ​ ​ 100 0,8e iπ / 4 ]] 0,8e iπ / 2 ]]> 0,4e iπ / 2 ]]>

[0180] The damped oscillation is characterized by the damping |z m | < 1. As Figure 3 illustrated in the complex plane representations 310, 330 and 350, the oscillation parameter z m is closer to the unit circle, the damping is smaller and the decay time is longer. Conversely, the oscillation parameter z m is further away from the unit circle and closer to its center, the damping is stronger and the decay time is shorter.

[0181] The oscillation frequency is characterized by the angle of z m , i.e. ∠z m . As Figure 3 illustrated in the complex plane representations 310, 330 and 350, the greater the frequency of the associated oscillation function (see corresponding representations 320, 340, 360), the greater the angle ∠z m of the oscillation parameter z m relative to the real axis of e.g. the complex plane representation 310, 330, 350.

[0182] Further in accordance with the present disclosure, it is considered that the scaling, phase, damping and frequency of the observed oscillation function after switching off the RF field will be characteristic for certain objects. Thus, the method proposed in accordance with the present disclosure attempts and can successfully classify objects in the reader field based on Θ.

[0183] Following this line of thought, the method proposed in accordance with the first aspect of the present disclosure for determining the type of an object in the RF field of an NFC reader device can be subdivided into the following main steps:

[0184] 1. Record N samples of ADC data after switching off the reader field.

[0185] 2. Estimate Θ from the N observed signal samples of x[n].

[0186] 3. Based on the estimate value classify any object in the reader field. (In this document, indicates the estimate value.)

[0187] For each of steps 2 and 3, a number of embodiments are known.

[0188] With respect to step 2, a non-exhaustive list of possibilities to estimate the oscillation parameter includes

[0189] - the Prony method,

[0190] - the function bundle method, and

[0191] - the (total) least squares estimator.

[0192] For a description of the Prony method, reference is made to the following document: Fernandez Rodriguez, A. &.-G.-J. (2018): “Coding Prony’s method in MATLAB and applying it to biomedical signal filtering. BMC Bioinformatics”.

[0193] For a description of the function bundle method, reference is made to the following document: Y. Hua, T. S. (1990): “Matrix pencil method for estimating parameters of exponentially damped / undamped sinusoids in noise”. IEEE Trans. Acoust. Speech Signal Process.

[0194] For a description of the (total) least squares estimator method, reference is made to the following documents: SW., C. (2000): “A two-stage discrimination of cardiac arrhythmias using a total least squares-based Prony modeling algorithm”. IEEE Trans Biomed Eng; and Markovsky I, V. H. (2007): “Overview of total least-squares methods”. Signal Processing.

[0195] As to step 3, a non-exhaustive list of possibilities as to how the classification can be performed includes:

[0196] - a Bayesian classifier,

[0197] - a support vector machine, and

[0198] - Artificial Neural Networks.

[0199] For a description of the Bayesian classifier method, reference is made to the following document: Devroye, L.; Gyorfi, L. & Lugosi, G. (1996): "A probabilistic theory of pattern recognition". Springer.

[0200] For a description of the Support Vector Machine method, reference is made to the following document: Christopher Bishop (2006): "Pattern Recognition and Machine Learning". Springer.

[0201] For a description of the Artificial Neural Networks method, reference is made to the following document: Schmidhuber, J. (2015): "Deep Learning in Neural Networks: An Overview". Neural Networks, vol 61 : p. 85-117.

[0202] For implementation purposes, it can be necessary and thus recommended to reduce the computational complexity of the oscillation parameter acquisition according to steps 2 and 3. One possible way to approximate the optimization of this step is to simplify it. However, the same number (or characteristic parameter) of θ m The determination of other parameters that reflect the same number (or characteristic parameter) of θ

[0203] According to a second aspect of the present disclosure, it is envisaged that a training set of numbers (characteristic parameters) like θ m should be acquired during a training phase for each object type for a specific NFC reader device, the classification should be enabled for said each object type.

[0204] In the following, reference is made to Figure 3 A formal generalization of the method proposed for determining the type of an object in the RF field of a NFC reader device according to the first aspect of the present disclosure is provided in the following. Subsequently, reference is made to Figure 4A formal description of a method for acquiring a training dataset for an NFC reader device for a plurality of training objects of an enabled training object type from an RF field of the NFC reader device according to the second aspect of the disclosure is provided.

[0205] Figure 5 A flow chart of a method 400 for determining a type of an object 150, 180, 190 arranged in an RF field 140 transmitted by an NFC reader device 100 according to an embodiment of the first aspect of the disclosure is shown. The type of the object 150, 180, 190 is from a predetermined group of object types. See Figure 4 The group of object types comprises at least one of an NFC tag 150, a metal object 180, and a no-load 190 for any object.

[0206] The method 400 starts at step 410 by switching on the RF field 140 of the NFC reader device 100 and continues at step 420 by arranging an object 150, 180 in the RF field 140 of the NFC reader device 100. The NFC reader device 100 can be an NFC-capable reader device and can be a battery-powered device. Subsequently, at step 430, the RF field 140 is switched off.

[0207] After switching off the transmitted RF field 140 at step 230, the method continues at step 440 by measuring and recording an attenuated signal trace of the front-end processed RF input signal 131 generated by the object 150, 180 in the RF field 140 of the NFC reader device 100, e.g. a signal trace, e.g. Figure 1 the traces 250 and 260 shown in Figs. 2(e) and 2(f), or Figure 2 the attenuated signal traces 612 and 622 shown in Fig. 6.

[0208] Subsequently, the method continues at step 450 by estimating a decomposition of the recorded attenuated signal trace generated by the object 150, 180 (see Figure 6 the traces 250 and 260 in Figs. 2(e) and 2(f), and Figure 2 the traces 612 and 622 in Fig. 6) into M superposed components (see e.g. Figure 6 the oscillating components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c), and 2(d)) of the predetermined decomposition scheme. The method then continues at step 460 by determining and storing, for each superposed component, a signal trace of the recorded attenuated signal trace (see Figure 2 the traces 250 and 260 shown in Figs. 2(e) and 2(f), and Figure 2The estimated characteristic parameter set associated with the estimated decomposed M superimposed components of the attenuated signal traces 612 and 622 shown in the figure, for example, the parameter θ for m = 1, 2, ..., M. m It should be noted that, according to the decomposition, the M superimposed components of the decomposition scheme (see, for example) Figure 6 Each superposition component in the oscillation components 210, 220, 230, 240 shown in (a), 2(b), 2(c), and 2(d) is defined by a predetermined superposition function and an associated set of characteristic parameters.

[0209] Once the estimated decomposition has been achieved at 450 and the estimated characteristic parameters have been stored at 460, the method proceeds to step 470, which includes: based on at least one of the stored estimated characteristic parameters of the M superimposed components according to the decomposition scheme, attempting to train the characteristic parameters of the M superimposed components on the training dataset (see datasets 710, 720, 730 in Figure 7, datasets 810, 820, 830 in Figure 8, and...) Figure 2 The datasets 1010 and 1020 in Figure 7 are used to classify the at least one estimated feature parameter. The training datasets (datasets 710, 720, and 730 in Figure 7, and datasets 810, 820, and 830 in Figure 8, and...) Figure 10 The datasets 1010 and 1020 are stored in a database 138 accessible by the P&E unit 124 of the NFC reader device 100, and have been obtained during the training phase using multiple training objects 150, 180, and 190 from each of the predetermined object type groups, as shown in the reference. Figure 10 The method 500 for an NFC reader device 100 is shown in the figure.

[0210] Subsequently, at step 480, it is checked whether the attempted classification was successful. If unsuccessful, see 480 for "No", then the method proceeds to the endpoint. If the attempted classification was successful, see 480 for "Yes", then the method proceeds to step 490, which involves determining the type of the objects 150, 180 to which the attenuated signal traces recorded in step 440 correspond based on the successful classification, and then proceeds to the endpoint.

[0211] Figure 5 A flowchart illustrating a method 500 for obtaining a training dataset for an NFC reader device 100 for a plurality of training objects 150, 180, 190 from a training object group, according to an embodiment of a second aspect of the present disclosure.

[0212] Method 500 begins at step 505 by pre-determining training groups 150, 180, and 190 based on object types. See also Figure 5The object type group includes at least one of NFC tag 150, metal object 180, and no-load 190 for any object.

[0213] Method 500 proceeds to step 510, which includes pre-determining the signal trace to be attenuated (see...). Figure 1 250 and 260 in (e) and 2(f), and Figure 2 The traces 612 and 622 in the figure are decomposed into M superimposed components (see example). Figure 6 The M-degree decomposition schemes (210, 220, 230, 240) shown in (a), 2(b), 2(c), and 2(d) are illustrated. It should be noted that each of the M superposition components 210, 220, 230, 240 is defined by a predetermined superposition function, which is determined by an associated set of characteristic parameters. Step 510 may further involve storing the decomposition scheme and indications of the M predetermined superposition functions, in particular, in a database 138 accessible to the P&E unit 124 of the NFC reader device 100.

[0214] Method 500 then performs the following steps 515 to 560 for each predefined object type.

[0215] In doing so, the method proceeds to step 515, which involves selecting an object type from the object type training group. Method 500 then proceeds to step 520, which involves providing the predetermined object type described in step 520. Figure 2 The system provides multiple different training objects 150, 180, and 190, and specifically provides training objects of a predetermined type, such as the first training object, the second training object, etc.

[0216] Method 500 then proceeds to step 525, which involves activating the RF field 140 of the NFC reader device 100, and continues at step 530, placing objects 150, 180 within the RF field 140 of the NFC reader device 100. The NFC reader device 100 may be an NFC-enabled reader device and may be a battery-powered device. Subsequently, at step 535, the RF field 140 is deactivated.

[0217] After cutting off the RF field 140 emitted by 535, the method continues at step 540 by measuring and recording the attenuated signal trace of the front-end processed RF input signal 131 generated by the training subjects 150, 180 in the RF field 140 of the NFC reader device 100, such as a signal trace. Figure 1 Traces 250 and 260 shown in (e) and 2(f), or Figure 2 The attenuation signal traces 612 and 622 are shown in the figure.

[0218] Subsequently, the method proceeds to step 545, involving estimating an M- degree decomposition of the recorded training attenuation signal traces (see e.g. traces 250 and 260 in Figs. 2(e) and 2(f) and traces 612 and 622 in Fig. 6) generated by said objects 150, 180 into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 6 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 2 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 6 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 2 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 2 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). Figure 2 (e) and 2(f) and traces 612 and 622 in Fig. 6) into M superposition components of the decomposition scheme (see e.g. superposition components 210, 220, 230, 240 shown in Figs. 2(a), 2(b), 2(c) and 2(d)). m The method additionally proceeds to step 555, involving storing the determined associated estimated characteristic training parameter set for each superposition component in the database 138.

[0219] The method then proceeds to step 560, involving checking whether there is another training object of the selected type. If this is the case, i.e. yes at 560, the method returns to step 520. If there is no training object of the selected type, meaning that training attenuation signal traces have been recorded and processed for all training objects of the selected type, i.e. no at 560, the method proceeds to step 565.

[0220] At step 565, it is checked whether there is another object type of the object type training group. If this is the case, i.e. yes at 565, the method returns to step 515. If there is no further type of training object, meaning that training attenuation signal traces have been recorded and processed for all training objects of the selected types, i.e. no at 565, the method proceeds to the end point.

[0221] The methods 400 and 500 can be executed in the NFC reader device 100. To this end, the NFC reader device 100 comprises a processing and evaluation (P&E) unit 124 and, in addition, a database 138 accessible to the P&E unit 124. The P&E unit 124 is capable of recording and storing the decay signal traces (250, 260; 612, 622) measured after switching off the RF field. The P&E unit 124 is in addition capable of executing the methods 400 and 500. The database 138 is capable of storing the training data sets 710, 720, 730; 810, 820, 830; 1010, 1020 that have been obtained during the training phase for the NFC reader device 100 according to the method 500. In particular, the database 138 is capable of storing the resolution scheme and the indication of the M predetermined superposition functions in step 510 and, in addition, the plurality of training decay signal traces measured in step 540 and the respective plurality of characteristic parameters determined for the resolution of the recorded training signal traces into the M predetermined superposition functions in step 550.

[0222] According to a first aspect of the present disclosure, the method 400 for determining the type of an object 150, 180, 190 arranged in the RF field 140 emitted by the NFC reader device 100 has been executed and tested using an experimental setup comprising an experimental NFC reader device, which comprises the following concepts: estimating 450 a classification of the recorded decay signal trace of an object arranged in the RF field 140 of the NFC reader device 100; attempting 470 a classification based on the characteristic parameters determined by the resolution; and, if possible, determining the object type based on the classification. And according to a second aspect of the present disclosure, the method 500 of obtaining a training data set for the NFC reader device 100 for a plurality of training objects 150, 180, 190 from a training group of object types has been executed and tested using an experimental setup comprising an experimental NFC reader device. These tests prove that the underlying concepts of the method comprising the resolution of the decay signal trace into superposition components and the attempt of a classification based on the characteristic parameters determined by said resolution are feasible and the concepts are validated.

[0223] In the following reference is made to Figure 6 The results and the validation of these tests of the underlying concepts of the method 400 for determining the type of an object 150, 180, 190 arranged in the RF field 140 emitted by the NFC reader device 100 and the method 500 of obtaining a training data set for the NFC reader device 100 for a plurality of training objects 150, 180, 190 from a training group of object types are described.

[0224] Figures 6 to 10Including left figure 610 and right figure 620, and thus, a pair of figures 610, 620 are shown, respectively illustrating the output of front-end processed attenuated signal traces 612, 622 with N=30 sample values ​​according to an embodiment of the first aspect of the present disclosure, and their approximate values ​​614, 624 decomposed to M=3 degrees. Figure 6 The left side diagram 610 shows the output of the analog-to-digital converter (ADC) in the I channel of the received signal processing chain 122, i.e., ADC I, and its approximate value decomposed to M=3 degrees according to equation (2) for N=30 sample values. Figure 6 The diagram on the right shows the output of the analog-to-digital converter (ADC) in the Q channel of the received signal processing chain, i.e., the ADC, and its approximate value decomposed to M=3 degrees for N=30 sample values. The object in the RF field of the NFC reader device in the experimental setup is a metal plate, specifically a 60mm × 40mm aluminum plate, positioned 5mm from the antenna of the NFC reader device.

[0225] exist Figure 6 The attenuation signal traces 612 and 622 can be seen to be decomposed at M=3 degrees (i.e., for m=1, 2, and 3, only three superimposed components are involved). The decompositions 614 and 624 can be fitted to the measured attenuation signals 612 and 622 and provide good approximations 614 and 624 for the real part (I channel) and imaginary part (Q channel) of the attenuation signal trace.

[0226] In the sense of method 400 according to the first aspect of this disclosure, as shown in FIG7, it has been examined whether it is possible to analyze the attenuated signal trace after cutting off the reader field (more specifically, for the scaling of the underlying weighted oscillation function |a| for m = 1, 2, 3). m |、Phase ∠a m Damping | z m |and frequency ∠z m To separate the object type of the metal object 180 from the object type of the NFC tag 150, measurements were performed on the metal plate 180 and the NFC tag 150 at distances from the antenna of the NFC reader device ranging from 5 mm to 100 mm. For all measured attenuation signal traces, a decomposition according to model order M=3 was estimated, and superposition components according to model orders m=1, 2, and 3 were fitted to the measured attenuation signal traces. In this analysis, for simplicity, the complex-valued weights a of the individual weighted oscillation components are ignored. m; however, the complex weights can also be characteristic of certain objects. The resulting parametric representation of z Figure 3 , 7B in the complex plane is shown in FIGS. m

[0227] FIG. 7 includes Figure 7A , 7B and 7C, and thus, shows the respective figure triplets 710, 720, 730, showing the M=3 degree resolved oscillation parameters fitted to the respective plurality of measured and recorded attenuation signal traces using the functional beam approach, for a plurality of unloaded RF fields (representations 711, 721, 731), for a plurality of RF fields loaded with NFC tags (representations 712, 722, 732), and for a plurality of RF fields loaded with metallic objects (representations 713, 723, 733), respectively.

[0228] Figure 7A shows the figure triplet 710 showing the first oscillation parameter z1 of the M=3 superposition components obtained from the operation of fitting the M-degree decomposition to the plurality of measured and recorded training attenuation signal traces, for a plurality of training unloaded RF fields (representation 711), for a plurality of RF fields loaded with training NFC tags (representation 712), and for a plurality of RF fields loaded with training metallic objects (representation 713), respectively. Figure 7A shows the figure triplet 720 showing the second oscillation parameter z2 of the M=3 superposition components obtained from the operation of fitting the M-degree decomposition to the plurality of measured and recorded training attenuation signal traces, for a plurality of training unloaded RF fields (representation 721), for a plurality of RF fields loaded with training NFC tags (representation 722), and for a plurality of RF fields loaded with training metallic objects (representation 723), respectively. Figure 7B shows the figure triplet 730 showing the third oscillation parameter z3 of the M=3 superposition components obtained from the operation of fitting the M-degree decomposition to the plurality of measured and recorded training attenuation signal traces, for a plurality of training unloaded RF fields (representation 731), for a plurality of RF fields loaded with training NFC tags (representation 732), and for a plurality of RF fields loaded with training metallic objects (representation 733), respectively.

[0229] In Figure 7C ​The point cloud or cluster of oscillation parameters associated with the oscillation component for m = 1, i.e. z1, is seen to be very similar for the metal object (indicated 713) and the unloaded case (indicated 711 ). Also for the NFC tag (indicated 712), a first point cloud or cluster is seen which is very similar to the point clouds or clusters of the metal object (indicated 713) and the unloaded case (indicated 711 ). This can mean that in case a classification would be attempted based on characteristic parameters of this oscillation component, it can not be possible to distinguish the presence of a metal object from the unloaded case based on characteristic parameters of the oscillation function associated with the superposition component for m = 1. This can additionally mean that it can not be possible to distinguish the presence of an NFC tag from the presence of a metal object and the unloaded case, which presence of an NFC tag would result in an oscillation function associated with the superposition component for m = 1 by characteristic parameters in the decomposition of the attenuated signal trace such that the oscillation function falls in the first point cloud.

[0230] However, for the NFC tag (indicated 712), a second additional point cloud or cluster is seen which is closer to the unit circle, meaning that the damping |z1 | is smaller than for those points of the first point cloud which are closer to the origin of the complex plane. This means that characteristic parameters of the oscillation parameters associated with m = 1 can be used for classification of object types in the RF field of an NFC reader, wherein an NFC tag can be distinguished from a metal object and the unloaded case.

[0231] With regard to Figure 7A , the point cloud of oscillation parameters associated with the oscillation component for m = 2, i.e. z2, is seen to be very similar for the unloaded case (indicated 721 ), the metal object (indicated 722) and the NFC tag (indicated 723), which means that characteristic parameters of the oscillation parameters associated with m = 2 can be difficult to use for classification. Similarly, with regard to Figure 7B , the point cloud of oscillation parameters associated with the component for m = 3, i.e. z2, is seen to be very similar for the unloaded case (indicated 731 ), the metal object (indicated 732) and the NFC tag (indicated 733), and can be difficult to use for classification between these object types.

[0232] The point clouds seen in Figs. 7A, 7B, 7C, 7D, 7E and 7F can be resolved with respect to the distance from the antenna of the NFC reader device at which the respective object has been arranged in the RF field. The resolution results for the distance are shown in Fig. 8. Figure 7C , 7B and 7C are resolved with respect to the distance from the antenna of the NFC reader device at which the respective object has been arranged in the RF field. The resolution results for the distance are shown in Fig. 8.

[0233] Fig. 8 comprises Figure 7A , 8BAnd 8C, and therefore, corresponding diagram pairs are shown below, respectively showing multiple RF fields loaded with an NFC tag varying with the distance from the NFC tag to the antenna of the NFC reader device, and multiple RF fields loaded with a metal object varying with the distance from the NFC tag to the antenna of the NFC reader device, the absolute values ​​of the first to third oscillation parameters corresponding to the M=3 degree decomposition of the corresponding multiple measured and recorded attenuation signal traces, and the angles of the first to third oscillation parameters corresponding to the M=3 degree decomposition.

[0234] Figure 8A The corresponding diagrams 811-812 and 813-814 show the absolute value of the first oscillation parameter z1 for the M=3 degree corresponding decomposition of the NFC tag (Diagram 811) and the angle of the first oscillation parameter z1 for the M=3 degree corresponding decomposition of the NFC tag (Diagram 813) and the metal object (Diagram 814), respectively. Figure 8A The corresponding diagrams 821-822 and 823-824 show the absolute value of the second oscillation parameter z2 for the M=3 degree corresponding decomposition of the NFC tag (Diagram 821) and the angle of the second oscillation parameter z2 for the M=3 degree corresponding decomposition of the NFC tag (Diagram 823) and the metal object (Diagram 824), respectively. Figure 8B The corresponding diagrams 831-832 and 833-834 show the absolute value of the third oscillation parameter z3 for the M=3 degree decomposition of the NFC tag (Diagram 831) and the angle of the third oscillation parameter z3 for the M=3 degree decomposition of the NFC tag (Diagram 833) and the metal object (Diagram 834), respectively.

[0235] exist Figure 8C In (Figure 811), it can be seen that Figure 8A Points closer to the unit circle in (represented by 712) can be associated with NFC tags immediately adjacent to the NFC reader device. Lower damping and the associated longer decay time are a result of the resonant characteristics of the NFC tag. This is confirmed by observations of the frequency variance (i.e., z0). m Angle ∠z m See Figure 7A The lower left panel (813) decreases at closer distances, indicating a stronger performance of the NFC tag's resonant frequency in the decaying signal. Figure 8A The results shown indicate that, specifically, when object types such as NFC tags and metal objects are positioned near the RF antenna, i.e. at a close distance of less than about 25 mm from the RF antenna, the presence of an NFC tag can be distinguished from the presence of a metal object.

[0236] In addition,Figure 8A and 8B It can be seen in

[0237] In Figure 8A is shown Figure 9 the corresponding time-domain shape of the individual components in

[0238] Figure 7A In the respective complex plane, the figures are shown which include figures showing the respective plurality of oscillation parameters z1, z2, z3 for the respective decomposition of the training attenuation signal trace for the training NFC tag (see representations 910, 920, 930) and the plurality of oscillation parameters z1 for the respective decomposition of the training attenuation signal trace for the training metal plate (see representation 940), as well as figures showing the associated oscillation functions as a function of the discrete time index (or sample index) in the time domain.

[0239] In representation 910 and figure 950, it can be seen that the unit circle is approached for the oscillation parameter associated with m = 1 for the NFC tag (the characteristic parameter 910 in representation 910 corresponds to an oscillation function with less strong damping and thus exhibits a longer decay time (in figure 950) compared to the characteristic parameter of the metal object closer to the center of the unit circle (in representation 940) which corresponds to an oscillation function with stronger damping and thus exhibits a short decay time (in figure 960) and thus no sustained oscillation after a longer time (higher sample index).

[0240] Further experiments have been conducted with more different NFC tags (three different tags) and more different metal plates (four different metal plates). The results are shown in Figure 9

[0241] Figure 10 Figure 10 ​A respective set 1010 and 1020 of complex plane plots 1011, 1012, 1013, 1014 for a plurality of (specifically, three) training NFC tags and 1021, 1022, 1023, 1024 for a plurality of (specifically, four) training metal objects is shown. The first set 1010 shows a respective plurality of oscillation parameters z1, z2, z3 for a respective decomposition of the training attenuation signal traces for a plurality of training tags arranged at four different distances (i.e. far field (see plot 1011), 50 mm (see plot 1012), 30 mm (see plot 1013) and 5 mm (see plot 1013)) in the RF field of the NFC reader device. The second set 1020 of complex plane plots shows a respective plurality of oscillation parameters z1, z2, z3 for a respective decomposition of the training attenuation signal traces for a plurality of training metal objects arranged at four different distances (i.e. far field (see plot 1021), 50 mm (see plot 1022), 30 mm (see plot 1023) and 5 mm (see plot 1024)) in the RF field of the NFC reader device.

[0242] Specifically, from the plots 1014 and 1024 for objects near the antenna (5 mm) it can be seen that the characteristic parameters resulting from the decomposition for the model order M = 3, specifically the oscillation parameters, are not only different between NFC tags and metal objects, but also among different NFC tags, see the distinguishable point clouds for z1 and z2 in plot 1014.

[0243] The results shown in plot 1014, i.e. the possibility to distinguish between different NFC tags, also makes it possible to classify certain NFC tags in order to optimize the settings of the NFC reader device for specific communication scenarios or similar applications with different NFC tags.

[0244] It is important to note that the method for determining the type of an object arranged in the RF field transmitted by the NFC reader device 100 according to the first aspect of the present disclosure can be used on top of any other state-of-the-art LPCD mechanism and method to improve object recognition performance and power saving efficiency.

[0245] In this presentation, example embodiments have been presented with respect to selected combinations of details. However, a person of ordinary skill in the art will appreciate that many other example embodiments can be practiced including different selected combinations of these details. The appended claims are intended to cover all possible example embodiments.

[0246] As a supplement, it should be noted that "have" or "include" does not exclude other elements or steps, and "one" or "a" does not exclude multiple. In addition, it should be noted that the features or steps described above with reference to one of the above-mentioned embodiment examples can also be used in combination with other features or steps of other embodiment examples described above. The element symbol in the claim is not understood as limiting.

Claims

1. A method (400) for determining the type of objects (150, 180, 190) arranged in a radio frequency (RF) field (140) emitted by an NFC reader device (100), characterized in that, The types of objects (150, 180, 190) are from a predefined group of object types. The NFC reader device (100) includes an RF antenna (130) and a received signal processing chain (128, 122). The RF antenna (130) is used to transmit an RF field (140) and to receive an RF input signal (131). The received signal processing chain (128, 122) is used for front-end processing of the received RF input signal (131) and to provide a front-end processed received signal (133)x corresponding to the received RF input signal (131) as the output of the received signal processing chain (128, 122). And the method (400) described therein includes the following steps: a) For an object (150, 180, 190) in an RF field (140) emitted from the NFC reader device (100), after the emitted RF field (140) is cut off (230), the attenuation signal traces (250, 260; 612, 622) of the received signal (133) generated by the object (150, 180, 190) in the RF field (140) are measured and recorded (440). b) Estimate (450) the recorded attenuation signal traces (250, 260; 612, 622) generated by the object according to a predetermined decomposition scheme to an M-degree decomposition of M superimposed components (210, 220, 230, 240) of the decomposition scheme, and further, for each superimposed component, determine (460) and store the associated estimated characteristic parameter set of the superimposed component (210, 220, 230, 240), wherein each of the M superimposed components (210, 220, 230, 240) is defined by a predetermined superposition function, which is in turn determined by the characteristic parameter set; c) Based on at least one of the estimated characteristic parameters determined according to the M superimposed components (210, 220, 230, 240) of the decomposition scheme, attempt (470) to classify the at least one characteristic parameter among the characteristic parameters with respect to the training dataset (710, 720, 730; 810, 820, 830; 1010, 1020) of the training characteristic parameters of the M superimposed components; as well as d) If the classification of at least one of the estimated characteristic parameters is successful (480, Y), then based on the successful classification, determine (490) the type of the object (150, 180, 190) to which the attenuation signal traces (250, 260; 612, 622) recorded in step a) are targeted.

2. The method (400) according to claim 1, characterized in that, It has at least one of the following characteristics: i) The group of object types said therein includes at least one of NFC tags (150), metal objects (180), and no-load (190) for any object; ii) The NFC reader device (100) wherein the NFC reader device (100) has a processing and evaluation P&E unit (124) for processing and evaluating the received signal (133) after front-end processing, wherein in step a), the attenuated signal traces (250, 260; 612, 622) are recorded and stored in the P&E unit (124), and wherein steps b), c) and d) are performed in the P&E unit (124). iii) The training datasets (710, 720, 730; 810, 820, 830; 1010, 1020) are stored in a database (138) accessible to the P&E unit (124) and have been obtained during the training phase (500) for the NFC reader device (100) using multiple training objects (150, 180, 190) from each object type in the predetermined object type group; iv) The measurement and recording of the attenuation signal traces (250, 260; 612, 622) in step a) includes: starting at or after the cutoff (230) of the RF field (140), recording the temporal dependence of the attenuation signal traces (250, 260; 612, 622) during a time interval starting after a predetermined time delay after the cutoff (230) of the RF field (140), wherein the attenuation signal value associated with the attenuation signal trace is a complex value; v) Wherein the estimation (450) in step b) according to the predetermined decomposition scheme of the M-degree decomposition includes fitting the M-degree decomposition to the recorded attenuation signal traces (250, 260; 612, 622) by changing the associated characteristic parameters of the estimated decomposition of the M superimposed components (210, 220, 230, 240).

3. The method (400) according to claim 1 or claim 2, characterized in that, At least characteristic iv), wherein the measurement and recording (440) of the attenuated signal traces (250, 260; 612, 622) in step a) includes: recording the temporal dependence of the attenuated signal traces (250, 260; 612, 622) in the form of time-series signal values ​​x[n], where n denotes a discrete-time index, and x[n] is a complex value and has: - The real part (450, 612) real(x[n]) and the imaginary part (460, 622) imag(x[n]), and / or -Absolute value |x[n]| and angle x[n].

4. The method (400) according to claim 3, characterized in that, The M-degree decomposition of the attenuated signal traces (250, 260; 612, 622) recorded by the estimation (450) in step b) includes: modeling the signal value x[n] of the time series as M superimposed components (320, 340, 360) according to the following formula. Add the sum of the terms d[n] representing white noise: ， Where N is the number of signal samples measured, n is the discrete-time index corresponding to the measured signal sample, and m is the identifier of the superimposed components (320, 340, 360). The index, where M is the allowed superposition component (320, 340, 360) in the decomposition. The number of integers, where n, N, m, and M are integers.

5. The method (400) according to claim 4, characterized in that, The M superimposed components (320, 340, 360) This indicates that the attenuated signal traces (250, 260; 612, 622) are decomposed into a set of superimposed functions, which are selected from the following clusters: oscillating function set; sine function set; pulse function set; and step function set.

6. The method (400) according to claim 1, characterized in that, It has at least one of the following characteristics: i) wherein estimating the M-degree decomposition of the recorded attenuated signal traces (250, 260; 612, 622) in step b) and determining the associated estimated characteristic parameter set comprises fitting the M-degree decomposition to the recorded attenuated signal traces (250, 260; 612, 622) by changing the characteristic parameters of the associated estimated characteristic parameter set of the superimposed components of the estimated decomposition, and wherein determining the associated estimated characteristic parameter set involves a mathematical method selected from the group of Proni methods, function bundle methods, and (total) least squares estimators; ii) Wherein the attempt (470) classification described in step c) involves mathematical methods selected from a group consisting of: Bayesian classification, support vector machines and artificial neural networks.

7. A method (500) for obtaining training datasets (710, 720, 730; 810, 820, 830; 1010, 1020) for an NFC reader device (100) from a database (138) for multiple training objects (150, 180, 190) from an object type training group, characterized in that, The training dataset (710, 720, 730; 810, 820, 830; 1010, 1020) includes multiple recorded training attenuation signal traces (250, 260; 612, 622) for each training object in at least one training object (150, 180, 190) selected from the training groups of the object types. The NFC reader device (100) includes an RF antenna (130) and a received signal processing chain (128, 122). The RF antenna (130) is used to transmit an RF field (140) and to receive an RF input signal (131). The received signal processing chain (128, 122) is used for front-end processing of the received RF input signal (131) and to provide a front-end processed received signal (133)x corresponding to the received RF input signal (131) as the output of the received signal processing chain (128, 122). And the method (500) described therein includes the following steps: 1) Predetermine (505) object (150, 180, 190) type training groups; 2) A predetermined (510) M-degree decomposition scheme is determined to decompose the attenuation signal traces (250, 260; 612, 622) into M superimposed components (210, 220, 230, 240), and the decomposition scheme and the indications of the M predetermined superimposed functions are stored in a database (138), wherein each superimposed component of the M superimposed components (210, 220, 230, 240) is defined by a predetermined superimposed function, and the predetermined superimposed function is determined by an associated characteristic parameter set; 3) For each predefined object type: 3.1) Provide multiple different training objects (150, 180, 190) of the predetermined object type described in (520). 3.2) Each of the multiple different training objects (150, 180, 190) provided is sequentially (530) brought into the RF field emitted by the NFC reader device. 3.3) For each training object (150, 180, 190) in the emitted RF field (140) of the NFC reader device (100), the following steps are performed: 3.3.1) After the RF field of the NFC reader device is cut off, the training attenuation signal traces (250, 260; 612, 622) of the received signals (133) generated by the training objects (150, 180, 190) in the RF field (140) after front-end processing are measured and recorded (540) in the database (138). 3.3.2) Estimate (545) the recorded training attenuation signal traces (250, 260; 612, 622) are decomposed into M superimposed components (210, 220, 230, 240) according to a predetermined decomposition scheme, and then for each superimposed component, determine (550) the associated determined characteristic training parameter set of the superimposed components of the estimated decomposed training attenuation signal traces (250, 260; 612, 622) generated by the training objects (150, 180, 190) in the RF field (140) and store (555) it in the database (138).

8. The method (500) according to claim 7, characterized in that, It has at least one of the following characteristics: i) The training group of the object types mentioned above includes at least one of NFC tags (150), metal objects (180), and no-load (190) for any object; ii) The NFC reader device (100) wherein the NFC reader device (100) has a processing and evaluation P&E unit (124) for processing and evaluating the received signal (133) after front-end processing. iii) wherein the NFC reader device (100) has a database (138) that can be accessed by the P&E unit (124). iv) wherein in step 2), the instructions for the decomposition scheme and the M predetermined superposition functions are stored in a database (138) accessible to the P&E unit (124), wherein in step 3.3.1), the training attenuation signal traces (250, 260; 612, 622) are stored in the database (138), and wherein in step 3.3.2), the feature training parameters are stored in the database (138); v) Wherein the measurement and recording (540) described in step 3.3.1) in the database (138) includes: The temporal dependencies (250, 260; 612, 622) of the training decay signal traces (250, 260; 610, 620) are recorded in the form of training signal values ​​x[n] of the time series, where n denotes the discrete-time index, and x[n] is a complex value and has: - The real part (450; 612) real(x[n]) and the imaginary part (460; 622) imag(x[n]), and / or -Absolute value |x[n]| and angle x[n].

9. An NFC-enabled reader device configured to perform the method (400) according to any one of claims 1 to 6 and / or configured to perform the method (500) according to any one of claims 7 to 8.

10. A machine-readable non-transitional storage medium for storing a computer program product, the computer program product comprising instructions that, when executed on a processor, microprocessor, or computer-controlled data processing system (124, 164), perform the method (400) according to any one of claims 1 to 6 and / or the method (500) according to any one of claims 7 to 8.

Images