NFC device fast pairing method based on neural network optimization

Abstract

The present invention relates to the field of wireless communication technology, and discloses a method for fast pairing of NFC devices based on neural network optimization, comprising the following steps: S101, when the transmitter and the receiver of the NFC device are started, each is initialized and the unique identification information of the transmitter and the receiver is loaded respectively; S102, the receiver scans the wireless radio frequency identification tag signal of the transmitter through the built-in NFC module, extracts the second candidate unique identification information and environmental signal characteristics of the transmitter, and the receiver predicts the signal distance and signal stability score of the target device based on a preset signal characteristic prediction model, which is used to confirm whether the target device is the transmitter; S103, the unique identification information of the transmitter and the receiver is bidirectionally verified and exchanged through the NFC protocol, and the receiver confirms whether the unique identification information of the transmitter matches the target device, and if the match is successful, writes it into the pairing data area of ​​the receiver. The present invention realizes efficient, accurate and safe pairing of NFC devices in complex environments.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communications, and in particular relates to a method for quickly pairing NFC devices based on neural network optimization. Background Art

[0002] In modern smart scenes, fast pairing and communication of NFC devices have become the key to achieving efficient work. For example, in smart classrooms, teachers' microphones need to be paired with receivers quickly and accurately to achieve seamless sound amplification or recording and broadcasting functions; in the performance field, efficient pairing of microphones with audio receiving devices can significantly reduce audio acquisition costs and improve configuration flexibility. However, current technologies still have many limitations in device pairing: 1. Traditional pairing is inefficient. Existing device pairing methods usually rely on manual selection of devices or screening based on a fixed device list. This method is prone to mispairing in complex environments where multiple devices coexist, and the pairing efficiency is low; 2. Lack of dynamic signal analysis capabilities. Current pairing technologies are mostly based on fixed signal parameters (such as MAC addresses) for device identification, which is difficult to handle signal interference or dynamically changing environments, making it difficult to accurately identify the target device; 3. Insufficient security. Existing technologies lack a reliable verification mechanism in the device identification verification process, and are vulnerable to security threats such as man-in-the-middle attacks, forged device identification, or signal tampering. Summary of the invention

[0003] The present invention provides a method for quickly pairing NFC devices based on neural network optimization, which solves the technical problems that the related technology has low efficiency, poor adaptability and insufficient security in the device pairing process, and is difficult to meet the requirements of fast and accurate pairing in complex environments.

[0004] The present invention provides a method for quickly pairing NFC devices based on neural network optimization, comprising the following steps:

[0005] S101, when the transmitter and the receiver of the NFC device are started, they are initialized and the unique identification information of the transmitter and the receiver is loaded respectively, the unique identification information of the transmitter is stored in the storage unit of the transmitter, and the receiver includes: a neural network module, a pairing data area and historical pairing data;

[0006] S102, the receiving end scans the radio frequency identification tag signal of the transmitting end through the built-in NFC module, extracts the second candidate unique identification information and environmental signal characteristics of the transmitting end, and the receiving end predicts the signal distance and signal stability score of the target device based on a preset signal characteristic prediction model to confirm whether the target device is the transmitting end, wherein the signal characteristic prediction model is stored in the neural network module built into the receiving end, the environmental signal characteristics include signal strength, signal propagation time and frequency, and the signal stability score ranges from 0 to 1;

[0007] S103, bidirectionally verify and exchange the unique identification information of the transmitter and the unique identification information of the receiver through the NFC protocol, and the receiver confirms whether the unique identification information of the transmitter matches the target device. If the match is successful, it is written into the pairing data area of ​​the receiver.

[0008] Further, extracting the second candidate unique identification information of the transmitting end includes:

[0009] S201, using an NFC module to scan and read multiple signal data blocks of a wireless radio frequency identification tag, and extract multiple first candidate unique identification information;

[0010] S202, counting the multiple unique identification information, and using a multiple voting method to select the unique identification information with the largest number of occurrences as the second candidate unique identification information;

[0011] S203, for the first candidate unique identification information with the same number of occurrences, use the first calculation formula to calculate the credibility corresponding to each first candidate unique identification information, and use the first candidate unique identification information with the highest credibility as the second candidate unique identification information, wherein the first calculation formula is:

[0012] , represents the credibility of the jth first candidate unique identification information, , and represent the first weight coefficient, the second weight coefficient and the third weight coefficient respectively, i and j are the indexes of the first candidate unique identification information, represents the weight coefficient of the i-th first candidate unique identification information, represents the signal strength corresponding to the i-th first candidate unique identification information, represents a set of first candidate unique identification information, represents the number of occurrences of the jth first candidate unique identification information, Indicates the number of times the j-th first candidate unique identification information appears in the historical records.

[0013] Further, the signal strength, signal propagation time and frequency of the target device in the historical pairing data are extracted and represented by a first feature sequence, and a signal feature prediction model is constructed based on a gated recurrent unit according to the first feature sequence and the environmental signal characteristics of the current target device. The structure of the signal feature prediction model includes: an input layer, a first hidden layer, a second hidden layer, a splicer and a classification layer;

[0014] The input layer is used to input the first feature sequence and the environmental signal features of the current target device;

[0015] The first hidden layer is used to input a first feature sequence and output a first hidden feature;

[0016] The second hidden layer is used to input the environmental signal feature of the current target device and output the second hidden feature;

[0017] The splicer is used to splice the first hidden feature and the second hidden feature to obtain a final hidden feature;

[0018] The classification layer includes a first classifier and a second classifier, and the final hidden features are input into the first classifier and the second classifier respectively. The classification space of the first classifier represents the signal distance of the target device, and the classification space of the second classifier represents the signal stability score of the target device.

[0019] Furthermore, the first hidden layer is constructed based on the gated recurrent unit, and the first hidden layer includes N first units, where N represents the number of sequence units of the first feature sequence, each first unit outputs a corresponding hidden feature, and the Nth first unit outputs the Nth hidden feature and uses it as the first hidden feature.

[0020] Furthermore, the calculation formula of the second hidden layer is:

[0021] ;

[0022] in, represents the second hidden feature, concat represents the concatenation operation, Indicates that and To perform horizontal splicing, and denote the fourth weight coefficient and the fifth weight coefficient respectively, Indicates the environmental signal characteristics of the current target device. represents the signal strength at the first preset position, R represents the signal strength of the current target device, Represents the path loss exponent.

[0023] Furthermore, the model is trained using the mean square error function as the loss function of the signal feature prediction model, the signal distance and signal stability score obtained by the signal feature prediction model are used as predicted values, and the distance and signal stability score in the historical paired data are used as the true value of the loss function.

[0024] Furthermore, the specific steps of S103 include:

[0025] S301, the receiving end initiates handshake communication with the target device through the NFC module;

[0026] S302, the transmitting end reads and sends its unique identification information to the receiving end;

[0027] S303, after receiving the unique identification information of the transmitting end, the receiving end performs verification based on the message authentication code combined with the challenge response mechanism;

[0028] S304, after the receiving end completes the verification, the unique identification information of the transmitting end is written into the pairing data area.

[0029] Furthermore, the specific steps of S303 include:

[0030] S401, the receiving end sends a random challenge value to the transmitting end;

[0031] S402, the transmitter calculates the message authentication code of the transmitter using a second calculation formula based on its unique identification information and the shared key in combination with the random challenge value. The second calculation formula is:

[0032] ,in, Represents the message authentication code of the transmitter, Indicates the HMAC algorithm with the shared key K. represents the random challenge value, Indicates the unique identification information of the transmitter. , and represent the sixth weight coefficient, the seventh weight coefficient and the eighth weight coefficient for encoding the signal characteristic, respectively, represents the signal propagation time, represents a minimum positive value, represents frequency, || represents concatenation operation;

[0033] S403, the transmitting end sends its message authentication code and its unique identification information to the receiving end;

[0034] S404, the receiving end calculates a message authentication code of the receiving end using a second calculation formula according to the random challenge value, the unique identification information of the transmitting end and the shared key;

[0035] S405, when the message authentication code of the transmitting end is equal to the message authentication code of the receiving end, the verification is successful.

[0036] The beneficial effects of the present invention are as follows: the present invention uses a built-in NFC module combined with a neural network prediction model to quickly extract the unique identification information and signal characteristics of the target device, and uses the signal distance and stability score to confirm the target device, thereby greatly improving the device pairing speed;

[0037] The present invention adopts a message authentication code combined with a challenge response mechanism, and ensures the uniqueness and tamper resistance of each verification through a random challenge value and a nested HMAC algorithm, prevents man-in-the-middle attacks and forged device identification, and enhances the integrity and authenticity of data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flow chart of the NFC device fast pairing method based on neural network optimization of the present invention. DETAILED DESCRIPTION

[0039] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0040] like Figure 1 As shown, the NFC device fast pairing method based on neural network optimization includes the following steps:

[0041] S101, when the transmitter and the receiver of the NFC device are started, they are initialized and the unique identification information of the transmitter and the receiver is loaded respectively, the unique identification information of the transmitter is stored in the storage unit of the transmitter, and the receiver includes: a neural network module, a pairing data area and historical pairing data;

[0042] S102, the receiving end scans the wireless radio frequency identification tag signal of the transmitting end through the built-in NFC module, extracts the second candidate unique identification information and environmental signal characteristics of the transmitting end, and the target device is the transmitting end. The receiving end predicts the signal distance and signal stability score of the target device based on a preset signal characteristic prediction model to confirm whether the target device is the transmitting end, wherein the signal characteristic prediction model is stored in the neural network module built into the receiving end, and the environmental signal characteristics include signal strength, signal propagation time and frequency, and the signal stability score ranges from 0 to 1;

[0043] S103, bidirectionally verify and exchange the unique identification information of the transmitter and the unique identification information of the receiver through the NFC protocol, and the receiver confirms whether the unique identification information of the transmitter matches the target device. If the match is successful, it is written into the pairing data area of ​​the receiver.

[0044] In one embodiment of the present invention, for a smart classroom scenario, during the pairing process of a microphone and a receiver, when the transmitter and receiver of the NFC device are initialized, the receiver loads its own MAC ID and simultaneously scans and identifies the MAC IDs of multiple transmitter devices. Through the uniqueness of the MAC ID, the system can quickly confirm the speaker's microphone and pair it with it.

[0045] In one embodiment of the present invention, the unique identification information is the MAC ID of the device, which is pre-assigned by the device manufacturer or dynamically generated based on the hardware characteristics of the device.

[0046] In one embodiment of the present invention, the specific steps of extracting the unique identification information of the target device include:

[0047] S201, using an NFC module to scan and read multiple signal data blocks of a wireless radio frequency identification tag, and extract multiple first candidate unique identification information;

[0048] S202, counting the multiple unique identification information, and using a multiple voting method to select the unique identification information with the largest number of occurrences as the second candidate unique identification information;

[0049] S203, for the first candidate unique identification information with the same number of occurrences, use the first calculation formula to calculate the credibility corresponding to each first candidate unique identification information, and use the first candidate unique identification information with the highest credibility as the second candidate unique identification information, wherein the first calculation formula is:

[0050] , represents the credibility of the jth first candidate unique identification information, , and They represent the first weight coefficient, the second weight coefficient and the third weight coefficient respectively. The first weight coefficient is used to measure the contribution of the signal strength in the first candidate unique identification information to the credibility calculation. The second weight coefficient is used to adjust the proportion of the number of times the candidate unique identification information appears in the scan data to its credibility. The third weight parameter is used to reflect the weighted effect of the number of times the candidate unique identification information appears in the historical records on the credibility calculation. i and j are the indexes of the first candidate unique identification information. represents the weight coefficient of the i-th first candidate unique identification information, represents the signal strength corresponding to the i-th first candidate unique identification information, represents a set of first candidate unique identification information, represents the number of occurrences of the jth first candidate unique identification information, Indicates the number of times the j-th first candidate unique identification information appears in the historical records.

[0051] In one embodiment of the present invention, the specific steps of extracting the environmental signal characteristics of the target device include: the NFC module activates the wireless radio frequency identification tag through the antenna and captures the returned wireless signal, and decodes the wireless signal through the ISO / IEC14443 protocol to obtain the signal strength, signal propagation time and frequency.

[0052] In one embodiment of the present invention, the signal strength, signal propagation time and frequency of the target device in the historical pairing data are extracted and represented by a first feature sequence, and a signal feature prediction model is constructed based on a gated recurrent unit according to the first feature sequence and the environmental signal features of the current target device. The structure of the signal feature prediction model includes: an input layer, a first hidden layer, a second hidden layer, a splicer and a classification layer;

[0053] The input layer is used to input the first feature sequence and the environmental signal features of the current target device;

[0054] The first hidden layer is used to input a first feature sequence and output a first hidden feature;

[0055] The second hidden layer is used to input the environmental signal feature of the current target device and output the second hidden feature;

[0056] The splicer is used to splice the first hidden feature and the second hidden feature to obtain a final hidden feature;

[0057] The classification layer includes a first classifier and a second classifier, and the final hidden features are input into the first classifier and the second classifier respectively. The classification space of the first classifier represents the signal distance of the target device, and the classification space of the second classifier represents the signal stability score of the target device.

[0058] In one embodiment of the present invention, when the signal stability score output by the signal feature prediction model is greater than a first preset threshold, the device is used as the target device, otherwise, each first candidate unique identification information is judged in order of credibility whether the signal stability score meets the requirement of being greater than the first preset threshold.

[0059] In one embodiment of the present invention, the first hidden layer is constructed based on a gated recurrent unit, and the first hidden layer includes N first units, where N represents the number of sequence units of the first feature sequence, each first unit outputs a corresponding hidden feature, and the Nth first unit outputs the Nth hidden feature and uses it as the first hidden feature.

[0060] In one embodiment of the present invention, the calculation formula of the second hidden layer is:

[0061] ;

[0062] in, represents the second hidden feature, concat represents the concatenation operation, Indicates that and Perform horizontal splicing, specifically, and For row vectors, the horizontal splicing operation is performed end to end according to the dimension of the row vector, and the original arrangement order is maintained after splicing. and They represent the fourth weight coefficient and the fifth weight coefficient respectively. The fourth weight parameter is used to adjust the weight ratio between the environmental signal characteristics of the current target device and the signal strength at the preset position. The fifth weight parameter is used to balance the influence of the signal strength of the current target device and the path loss index in the calculation of the second hidden layer characteristics. Indicates the environmental signal characteristics of the current target device. represents the signal strength at the first preset position, R represents the signal strength of the current target device, Represents the path loss exponent.

[0063] In one embodiment of the present invention, the mean square error function is used as the loss function of the signal feature prediction model to train the model, the signal distance and signal stability score obtained by the signal feature prediction model are used as prediction values, and the distance and signal stability score in the historical paired data are used as the true value of the loss function.

[0064] In one embodiment of the present invention, the specific steps of S103 include:

[0065] S301, the receiving end initiates handshake communication with the target device through the NFC module;

[0066] S302, the transmitting end reads and sends its unique identification information to the receiving end;

[0067] S303, after receiving the unique identification information of the transmitting end, the receiving end performs verification based on the message authentication code combined with the challenge response mechanism;

[0068] S304, after the receiving end completes the verification, the unique identification information of the transmitting end is written into the pairing data area.

[0069] In one embodiment of the present invention, the specific steps of S303 include:

[0070] S401, the receiving end sends a random challenge value to the transmitting end;

[0071] S402, the transmitter calculates the message authentication code of the transmitter using a second calculation formula based on its unique identification information and the shared key in combination with the random challenge value. The second calculation formula is:

[0072] ,in, Represents the message authentication code of the transmitter, Indicates the HMAC algorithm with the shared key K. represents the random challenge value, Indicates the unique identification information of the transmitter. , and They respectively represent the sixth, seventh and eighth weight coefficients of the signal characteristic coding, the sixth weight coefficient is used to adjust the importance of the signal propagation time, the seventh weight coefficient is used to enhance the robustness of the message authentication code calculation, and the eighth weight coefficient is used to adjust the influence of the frequency information in the message authentication code signal characteristic coding. represents the signal propagation time, represents a minimum positive value, preferably, Set to , represents frequency, || represents concatenation operation;

[0073] S403, the transmitting end sends its message authentication code and its unique identification information to the receiving end;

[0074] S404, the receiving end calculates a message authentication code of the receiving end using a second calculation formula according to the random challenge value, the unique identification information of the transmitting end and the shared key;

[0075] S405, when the message authentication code of the transmitting end is equal to the message authentication code of the receiving end, the verification is successful.

[0076] In one embodiment of the present invention, verification is performed based on a message authentication code combined with a challenge response mechanism, which greatly improves security. By introducing a random challenge value, the uniqueness of each verification content is ensured. By nesting HMAC to increase complexity, the ability to resist man-in-the-middle attacks and tampering is enhanced. In addition, the signal strength, signal propagation time and frequency of the device are combined, and the device characteristics in different scenarios are adapted through nonlinear transformation and weight adjustment, making the verification process more robust and accurate.

[0077] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. A fast pairing method for NFC devices based on neural network optimization, characterized in that: The following steps are involved: S101, when the transmitter and the receiver of the NFC device are started, they are initialized and the unique identification information of the transmitter and the receiver is loaded respectively, the unique identification information of the transmitter is stored in the storage unit of the transmitter, and the receiver includes: a neural network module, a pairing data area and historical pairing data; S102, the receiving end scans the radio frequency identification tag signal of the transmitting end through the built-in NFC module, extracts the second candidate unique identification information and environmental signal characteristics of the transmitting end, and the receiving end predicts the signal distance and signal stability score of the target device based on a preset signal characteristic prediction model to confirm whether the target device is the transmitting end, wherein the signal characteristic prediction model is stored in the neural network module built into the receiving end, the environmental signal characteristics include signal strength, signal propagation time and frequency, and the signal stability score ranges from 0 to 1; S103, bidirectionally verifying and exchanging the unique identification information of the transmitter and the unique identification information of the receiver through the NFC protocol, and the receiver confirms whether the unique identification information of the transmitter matches the target device, and if the match is successful, writes it into the pairing data area of ​​the receiver; The specific steps of S103 include: S301, the receiving end initiates handshake communication with the target device through the NFC module; S302, the transmitting end reads and sends its unique identification information to the receiving end; S303, after receiving the unique identification information of the transmitting end, the receiving end performs verification based on the message authentication code combined with the challenge response mechanism; S304, after the receiving end completes the verification, the unique identification information of the transmitting end is written into the pairing data area; The specific steps of S303 include: S401, the receiving end sends a random challenge value to the transmitting end; S402, the transmitter calculates the message authentication code of the transmitter using a second calculation formula based on its unique identification information and the shared key in combination with the random challenge value. The second calculation formula is: ,in, Represents the message authentication code of the transmitter, Indicates the HMAC algorithm with the shared key K. represents the random challenge value, Indicates the unique identification information of the transmitter. , and represent the sixth weight coefficient, the seventh weight coefficient and the eighth weight coefficient for encoding the signal characteristic, respectively, represents the signal propagation time, represents a minimum positive value, represents frequency, || represents concatenation operation; S403, the transmitting end sends its message authentication code and its unique identification information to the receiving end; S404, the receiving end calculates a message authentication code of the receiving end using a second calculation formula according to the random challenge value, the unique identification information of the transmitting end and the shared key; S405, when the message authentication code of the transmitting end is equal to the message authentication code of the receiving end, the verification is successful.

2. The NFC device fast pairing method based on neural network optimization according to claim 1 is characterized in that: Extracting the second candidate unique identification information of the transmitting end, the specific steps include: S201, using an NFC module to scan and read multiple signal data blocks of a wireless radio frequency identification tag, and extract multiple first candidate unique identification information; S202, counting the multiple unique identification information, and using a multiple voting method to select the unique identification information with the largest number of occurrences as the second candidate unique identification information; S203, for the first candidate unique identification information with the same number of occurrences, use the first calculation formula to calculate the credibility corresponding to each first candidate unique identification information, and use the first candidate unique identification information with the highest credibility as the second candidate unique identification information, wherein the first calculation formula is: , represents the credibility of the jth first candidate unique identification information, , and represent the first weight coefficient, the second weight coefficient and the third weight coefficient respectively, i and j are the indexes of the first candidate unique identification information, represents the weight coefficient of the i-th first candidate unique identification information, represents the signal strength corresponding to the i-th first candidate unique identification information, represents a set of first candidate unique identification information, represents the number of occurrences of the jth first candidate unique identification information, Indicates the number of times the j-th first candidate unique identification information appears in the historical records.

3. The NFC device fast pairing method based on neural network optimization according to claim 1 is characterized in that: The signal strength, signal propagation time and frequency of the target device in the historical pairing data are extracted and represented by a first feature sequence, and a signal feature prediction model is constructed based on a gated recurrent unit according to the first feature sequence and the environmental signal characteristics of the current target device. The structure of the signal feature prediction model includes: an input layer, a first hidden layer, a second hidden layer, a splicer and a classification layer; The input layer is used to input the first feature sequence and the environmental signal features of the current target device; The first hidden layer is used to input a first feature sequence and output a first hidden feature; The second hidden layer is used to input the environmental signal feature of the current target device and output the second hidden feature; The splicer is used to splice the first hidden feature and the second hidden feature to obtain a final hidden feature; The classification layer includes a first classifier and a second classifier, and the final hidden features are input into the first classifier and the second classifier respectively. The classification space of the first classifier represents the signal distance of the target device, and the classification space of the second classifier represents the signal stability score of the target device.

4. The NFC device fast pairing method based on neural network optimization according to claim 3 is characterized in that: The first hidden layer is constructed based on the gated recurrent unit. The first hidden layer includes N first units, where N represents the number of sequence units of the first feature sequence. Each first unit outputs a corresponding hidden feature, and the Nth first unit outputs the Nth hidden feature and uses it as the first hidden feature.

5. The NFC device fast pairing method based on neural network optimization according to claim 3 is characterized in that: The calculation formula for the second hidden layer is: ； in, represents the second hidden feature, concat represents the concatenation operation, Indicates that and To perform horizontal splicing, and denote the fourth weight coefficient and the fifth weight coefficient respectively, Indicates the environmental signal characteristics of the current target device. represents the signal strength at the first preset position, R represents the signal strength of the current target device, Represents the path loss exponent.

6. The NFC device fast pairing method based on neural network optimization according to claim 3 is characterized in that: The model is trained using the mean square error function as the loss function of the signal feature prediction model. The signal distance and signal stability score obtained by the signal feature prediction model are used as predicted values, and the distance and signal stability score in the historical paired data are used as the true value of the loss function.

Images