Biometric-based vehicle distributed identity authentication method and system

Abstract

The present application relates to the technical field of vehicle identity authentication, and discloses a vehicle distributed identity authentication method and system based on biological characteristics, which comprises the following steps: a user completes initial biological characteristic registration on a source vehicle, and biological characteristics are bound with random keys by a fuzzy binding algorithm to generate a binding factor and a biological characteristic hash value and store them in association; when the user uses a target vehicle, the target vehicle locates the source vehicle according to the user identity and requests verification. The source vehicle sends the binding factor and the biological characteristic hash value to the target vehicle; the target vehicle collects the current biological characteristics of the user, recovers the candidate key by decoding operation using the binding factor, and generates a temporary biological characteristic hash value; and the authentication is completed by comparing the temporary biological characteristic hash value with the received biological characteristic hash value. The method uses a decentralized hash transmission and fuzzy binding verification mechanism to realize efficient, safe and non-repeated cross-vehicle identity authentication on the premise of protecting the privacy of biological characteristics.

Description

A Biometric-Based Distributed Identity Authentication Method and System for Vehicles Technical Field

[0001] This invention relates to the field of vehicle identity authentication technology, and in particular to a vehicle distributed identity authentication method and system based on biometrics. Background Technology

[0002] In organizations with multiple vehicles, such as schools and businesses, the application of vehicle identification authentication is evolving from the traditional single-person, single-vehicle model to a collaborative, multi-person, multi-vehicle sharing model, requiring users to switch between different vehicles. Current vehicle authentication methods rely on independent biometric devices built into each vehicle. Users need to re-enter or verify their biometrics each time they use a different vehicle, which is inefficient in multi-vehicle environments.

[0003] Some attempts have been made to transmit authentication status between different devices through cloud synchronization or near-field communication, but this requires storing users' biometric information on a central server or transmitting replayable credentials between devices, which carries the risk of privacy leaks or replay attacks. Summary of the Invention

[0004] Therefore, the purpose of this invention is to overcome the shortcomings of low efficiency in identity authentication in multi-vehicle environments, reliance on centralized servers to store biometric information, and the vulnerability of privacy leaks and replay attacks caused by direct transfer of credentials between devices. The invention provides a biometric-based distributed vehicle identity authentication method and system that can achieve efficient, secure, and repetitive cross-vehicle identity authentication by using a decentralized hash transfer and fuzzy binding verification mechanism while protecting user biometric privacy.

[0005] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a vehicle distributed identity authentication method based on biometrics, comprising:

[0006] Obtain the initial biometric vector collected from the user on the source vehicle; bind the initial biometric vector and a random key using a fuzzy binding algorithm to obtain a binding factor; generate a biometric hash value based on the random key; and associate and store the binding factor, the biometric hash value, the user identity identifier, and the vehicle identifier.

[0007] When a user initiates a car rental request to the target vehicle, the user's identity identifier is obtained; the vehicle identifier associated with the user's identity identifier is mapped out based on the user's identity identifier; the source vehicle for verifying the user is determined based on the vehicle identifier; and a verification request is initiated to the source vehicle.

[0008] In response to the verification request, the source vehicle sends the binding factor and the biometric hash value to the target vehicle;

[0009] The target vehicle acquires the user's current biometric vector, and based on the received binding factor and the current biometric vector, obtains a candidate key through a decoding operation; and generates a temporary biometric hash value based on the candidate key.

[0010] The temporary biometric hash value is compared with the received biometric hash value;

[0011] If the comparison matches, the target vehicle is successfully authenticated, and the user is granted vehicle access.

[0012] Preferably, obtaining the user's initial biometric vector includes: acquiring the user's initial facial image through an in-vehicle image acquisition device; performing adaptive illumination equalization processing on the initial facial image to obtain an initial target image; and extracting features from the initial target image to obtain the initial biometric vector.

[0013] Preferably, obtaining the user's current biometric vector includes: acquiring the user's current facial image through an in-vehicle image acquisition device; performing adaptive illumination equalization processing on the current facial image to obtain a current target image; and extracting features from the current facial image to obtain the current biometric vector.

[0014] Preferably, the adaptive illumination equalization processing includes: partitioning the first image to obtain multiple image sub-regions; calculating the average pixel intensity of each image sub-region; calculating the weighted average of the average pixel intensities of all image sub-regions contained in the first image to obtain a global illumination reference value; calculating the deviation between the average pixel intensity of each image sub-region and the global illumination reference value to obtain a local illumination deviation value; performing gamma correction on the image sub-regions; wherein, if the local illumination deviation value is positive, the gamma value of the gamma correction is less than 1; if the local illumination deviation value is negative, the gamma value of the gamma correction is greater than 1; if the local illumination deviation value is 0, the gamma value of the gamma correction is equal to 1; and obtaining the target first image based on the image sub-regions after gamma correction processing.

[0015] Preferably, after gamma correction of the image sub-region, the method further includes performing contrast stretching transformation on the image sub-region; this includes: calculating the mean pixel intensity μ and standard deviation σ of the image sub-region after gamma correction; and setting an upper limit threshold I for the contrast stretching transformation. low and lower limit threshold I up :I low =max(0, μ-2×σ); I up =min(255, μ+2×σ); The intensity value I of each pixel within the image sub-region is mapped as follows: if I≤Ilow If I ≥ I, then the output pixel intensity is 0; up Then the output pixel intensity is 255; if I low <I<I up The output pixel intensity is: The image sub-regions after the contrast stretching transformation are stitched together to obtain the target first image.

[0016] Preferably, the binding factor is obtained by binding the initial biometric vector and the random key using a fuzzy binding algorithm, including: binarizing the initial biometric vector to obtain an initial binarized feature vector; performing error correction encoding on the random key to obtain an encoded key; and performing a bitwise XOR operation on the encoded key and the initial binarized feature vector to obtain the binding factor.

[0017] Preferably, based on the received binding factor and the current biometric vector, a candidate key is obtained through a decoding operation, including: binarizing the current biometric vector to obtain a current binarized feature vector; performing a bitwise XOR operation on the current binarized feature vector and the received binding factor to obtain a candidate encoded key; and performing error correction decoding on the candidate encoded key to obtain the candidate key.

[0018] Preferably, generating a biometric hash value based on the random key includes: performing a hash operation on the random key to obtain the biometric hash value; wherein the hash function of the hash operation is a non-cryptographic hash function.

[0019] Preferably, mapping the user identity identifier to the associated vehicle identifier includes: querying a locally stored distributed mapping table, which is updated based on authentication success events broadcast between vehicles; wherein each authentication success event contains a triplet of information: user identity identifier, source vehicle identifier, and timestamp; if multiple vehicle identifiers are found to correspond to the user identity identifier in the distributed mapping table, the nearest vehicle identifier is selected based on the timestamp.

[0020] Secondly, the present invention also provides a biometric-based distributed vehicle identity authentication system, comprising:

[0021] Multiple vehicle units, each vehicle unit comprising:

[0022] The biometric data acquisition module is used to collect users' biometric data;

[0023] The local storage module is used to store binding factors, biometric hash values, and user identity identifiers;

[0024] A communication module for communication between vehicles;

[0025] The processing module is configured to execute the biometric-based vehicle distributed identity authentication method.

[0026] A central management server, which communicates with the multiple vehicle units, is used to manage user access policies and trust chain configurations.

[0027] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0028] The biometric-based vehicle distributed identity authentication method and system described in this invention firstly uses a fuzzy binding algorithm to tolerate the inherent minor differences in the biometric collection process. As long as the current feature is sufficiently similar to the initial feature, the same random key can be recovered, thereby achieving biometric consistency verification across devices and time.

[0029] Secondly, during the entire verification process, the original biometric data and the key itself do not need to be transmitted over the network. Only the irreversible hash value and the binding factor used for unbinding are exchanged. This greatly protects the user's biometric privacy and prevents the data from being stolen and misused during transmission or storage.

[0030] Furthermore, by establishing a mapping between user identities and source vehicles, and transmitting and verifying hash values ​​between vehicles, decentralized trust transfer is achieved. A successful authentication result for one vehicle can be securely adopted by another vehicle through hash value comparison, breaking down information silos between vehicles.

[0031] Finally, the entire solution does not rely on a real-time online central server, but on direct communication between vehicles, which not only meets the real-time requirements of rapid authentication, but also enhances the robustness of the system in network outage or weak network environments. Attached Figure Description

[0032] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0033] Figure 1 is a flowchart of a vehicle distributed identity authentication method based on biometrics in a preferred embodiment of the present invention;

[0034] Figure 2 is a flowchart of the first part of the adaptive illumination equalization processing in a preferred embodiment of the present invention;

[0035] Figure 3 is a flowchart of the second part of the adaptive illumination equalization processing in a preferred embodiment of the present invention;

[0036] Figure 4 is a structural block diagram of the vehicle distributed identity authentication method based on biometrics in a preferred embodiment of the present invention. Detailed Implementation

[0037] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0038] Example 1: Referring to Figure 1, this embodiment of the invention discloses a vehicle distributed identity authentication method based on biometrics, including: obtaining an initial biometric vector collected by a user on a source vehicle; binding the initial biometric vector and a random key using a fuzzy binding algorithm to obtain a binding factor; generating a biometric hash value based on the random key; associating and storing the binding factor, biometric hash value, user identity identifier, and vehicle identifier; when a user initiates a vehicle usage request to a target vehicle, obtaining the user identity identifier; mapping the user identity identifier to the associated vehicle identifier, determining the source vehicle for verifying the user based on the vehicle identifier, and initiating a verification request to the source vehicle; the source vehicle responding to the verification request sending the binding factor and biometric hash value to the target vehicle; the target vehicle obtaining the user's current biometric vector, obtaining a candidate key through a decoding operation based on the received binding factor and current biometric vector; generating a temporary biometric hash value based on the candidate key; comparing the temporary biometric hash value with the received biometric hash value; if the comparison matches, the target vehicle successfully authenticates the user and grants the user vehicle usage permission.

[0039] The purpose of this invention is to overcome the shortcomings of low authentication efficiency in multi-vehicle environments, reliance on centralized servers to store biometric information, and the vulnerability of direct credential transfer between devices to privacy leaks and replay attacks. The implementation process is described below with reference to Figure 1:

[0040] Step 1: Initial User Registration and Credential Generation. This step is performed when a user first uses a vehicle in the fleet system (hereinafter referred to as the source vehicle).

[0041] Step S1.1: The user faces the vehicle's built-in camera and the original image containing the facial region is captured. Then, a face recognition algorithm is run to process the original image. Lightweight deep neural networks such as MobileFaceNet can be used to process the original image and extract the initial biological feature vector to represent the user's facial features.

[0042] Step S1.2: In the secure execution environment, a true random number generator is invoked to generate a random key K; subsequently, the fuzzy binding algorithm is executed, as follows:

[0043] The initial biometric vector is binarized to obtain an initial binarized feature vector; the random key is error-correcting encoded to obtain an encoded key; and the encoded key and the initial binarized feature vector are bitwise XORed to obtain a binding factor.

[0044] Specifically, the initial biological feature vector, obtained in the previous step, is converted into a binary string to obtain the initial binary feature vector B. An error-correcting code is used to encode the random key K, expanding the 256-bit random key K into a 511-bit codeword. In this embodiment, the error-correcting code is BCH(511, 256, t) code, where t represents the error-correcting capability, for example, t=30. The information bit length of the BCH(511, 256, t) code is 256 bits, and the codeword length is 511 bits. During the system design phase, the output layer of the feature extraction network is adjusted to make its output high-dimensional feature vector dimension 511; specifically, the last fully connected layer of the feature extraction network is adjusted to set the number of output nodes to 511.

[0045] Next, the binding factor δ is calculated: δ = E(K)⊕B; where E(K) represents the error-correcting code of the random key K; the symbol ⊕ represents the bitwise XOR operation. At this point, the user's initial biometrics are bound to the random key. The source vehicle securely stores the binding factor δ and the hash value H(K) of the random key K for subsequent local verification; the plaintext of the random key K and the original data of the initial biometric vector are then securely erased.

[0046] Step S1.3: Based on the random key K, calculate the biometric hash value H using a hash function. Specifically, perform a hash operation on the random key K using a non-cryptographic hash function to obtain the biometric hash value H. In this embodiment of the invention, the hash function used to obtain the biometric hash value H by hashing the random key K is not to protect the random key K, but to quickly generate a fixed identifier. Therefore, a non-cryptographic hash function with higher computational efficiency than a cryptographic hash function is used. This non-cryptographic hash function is selected to satisfy the system's requirements for the uniqueness and one-wayness of the hash value, provided that the random key K has cryptographically secure randomness. Examples include the xxHash algorithm, FarmHash algorithm, MurmurHash algorithm, and CityHash algorithm.

[0047] Step S1.4: Associate the binding factor δ, biometric hash value H, user identity identifier (such as employee ID or student ID) and source vehicle identifier (such as license plate number or unique code of vehicle equipment) and store them in the local security database.

[0048] Step 2: The target vehicle initiates a verification request. When a user requests to use the target vehicle, the target vehicle initiates the authentication process.

[0049] In step S2.1, the user sends their user identification identifier to the target vehicle via end-to-end wireless communication such as Bluetooth or NFC, or the target vehicle obtains the user identification identifier by reading the smart card carried by the user.

[0050] In step S2.2, the target vehicle queries its local cache or broadcasts a query to a small local network based on the received user identity identifier to determine which vehicle stores the user's credentials, that is, to determine the source vehicle associated with the user; after determination, the target vehicle sends a verification request to the source vehicle.

[0051] Step 3: The source vehicle responds to the request. Upon receiving the verification request, the source vehicle performs the following operations: retrieves the corresponding binding factor δ and biometric hash value H from the local database based on the user's identity identifier; and sends the binding factor δ and biometric hash value H to the target vehicle. To enhance security, in this embodiment, the source vehicle can use the target vehicle's temporary public key to encrypt this part of the data.

[0052] Step 4: Local verification of the target vehicle. The target vehicle executes the following steps in parallel upon receiving the response:

[0053] Step S4.1: Collect and process the user's current biometric data. Similar to step S1.1 above, collect the user's current facial image through the vehicle's onboard camera, extract the current biometric vector, and then binarize it into the current binarized feature vector B'.

[0054] Step S4.2, perform decoding operation to recover the random key K from the current binarized feature vector B': perform the following operations: perform a bitwise XOR operation on the current binarized feature vector and the received binding factor to obtain the candidate encoded key; perform error correction decoding on the candidate encoded key to obtain the candidate key.

[0055] Specifically, a bitwise XOR operation is performed on the current binarized feature vector B' and the received binding factor δ to calculate the candidate coding key C': C'=B'⊕δ.

[0056] Since δ = E(K)⊕B, the candidate encoded key C' = (B'⊕B)⊕E(K). Let the error e = B'⊕B, then C' = e⊕E(K). The target vehicle performs an error correction decoding operation on the candidate encoded key C' to obtain the candidate key K'. If the number of bit errors in the error e is within the error correction capability of the error-correcting code, then the decoding is successful and the candidate key K' = the random key K; otherwise, the decoding fails.

[0057] Step S4.3: Generate and compare hash values. If decoding is successful, calculate a temporary biometric hash value H', and then compare the temporary biometric hash value H' with the biometric hash value H received from the source vehicle bit by bit.

[0058] If the temporary biometric hash value H' matches the biometric hash value H, the target vehicle recognizes the user's identity as legitimate and authentication is successful; the target vehicle then grants the user the corresponding user-side permissions. Simultaneously, the target vehicle can broadcast the successful authentication to the fleet network, updating other vehicles' understanding of the user's availability.

[0059] This invention implements a fuzzy binding algorithm through error-correcting coding and XOR operations. Its core principle is to utilize the fault-tolerance capability of error-correcting codes, treating biometric differences as noise in the communication process. During verification, as long as the noise is within a correctable range, the original key can be recovered without error. This solves the problem of natural differences in biometric vectors collected from the same user at different times and in different vehicles.

[0060] The entire cross-vehicle authentication process does not directly compare biometric features or transmit the original key. Instead, it compares hash values ​​generated based on random keys. The one-way nature of the hash function ensures that even if the biometric hash value H is intercepted during transmission, attackers cannot deduce the random key K or biometric features, effectively protecting privacy.

[0061] The source vehicle stores credentials (δ, H) derived from the user's biometrics. The target vehicle autonomously verifies whether the user has biometrics that match the credentials through fuzzy commitment. This enables distributed trust transfer between vehicles without relying on a central server, breaking down information silos.

[0062] In the above embodiment, the initial biometric vector is extracted from the user's facial image captured by the vehicle's onboard camera, and the current biometric vector is extracted from the user's facial image captured by the target vehicle's onboard camera. The inconsistency in lighting conditions between the two acquisitions of user facial images will affect the consistency of the biometric vectors extracted subsequently, thereby reducing the reliability of subsequent identity authentication.

[0063] To address this issue, in the embodiments of this invention, the same adaptive illumination equalization processing method is used to perform adaptive illumination equalization processing on the initial facial image and the current facial image. Subsequently, feature extraction is performed on the image after adaptive illumination equalization processing to obtain the initial biometric vector and the current biometric vector.

[0064] Referring to Figure 2, the adaptive illumination equalization processing includes: dividing the first image into multiple image sub-regions; calculating the average pixel intensity of each image sub-region; calculating the weighted average of the average pixel intensities of all image sub-regions contained in the first image to obtain a global illumination reference value; calculating the deviation between the average pixel intensity of each image sub-region and the global illumination reference value to obtain a local illumination deviation value; performing gamma correction on the image sub-regions; wherein, if the local illumination deviation value is positive, the gamma value of the gamma correction is less than 1; if the local illumination deviation value is negative, the gamma value of the gamma correction is greater than 1; if the local illumination deviation value is 0, the gamma value of the gamma correction is equal to 1; and obtaining the target first image based on the image sub-regions after gamma correction processing.

[0065] Referring to Figure 3, after gamma correction of the image sub-region, the method further includes contrast stretching transformation of the image sub-region; this includes: calculating the mean μ and standard deviation σ of the pixel intensity of the image sub-region after gamma correction; and setting an upper limit threshold I for the contrast stretching transformation. low and lower limit threshold I up :I low =max(0, μ-2×σ); I up =min(255, μ+2×σ); Map the intensity value I of each pixel within the image sub-region as follows: If I≤I low If I ≥ I, then the output pixel intensity is 0; up Then the output pixel intensity is 255; if I low <I<I up The output pixel intensity is: The image sub-regions after contrast stretching transformation are stitched together to obtain the target first image.

[0066] It should be noted that when processing the initial facial image, the first image here is the initial facial image; when processing the current facial image, the first image here is the current facial image.

[0067] In a specific implementation, the first image is divided into multiple non-overlapping or partially overlapping image sub-regions. For each image sub-region, the average intensity value of all pixels is calculated, reflecting the local illumination level of the image sub-region. Based on the average intensity values ​​of all image sub-regions, a global illumination reference value is calculated through weighted averaging. The global illumination reference value represents the overall brightness level of the first image; specifically, the central region of the first image, which is usually the main area of ​​the face, is given a higher weight.

[0068] Next, for each image sub-region, the difference between its average intensity value and the global illumination reference value is calculated to obtain the local illumination deviation value for that region; a positive difference indicates that the region is too bright, and a negative difference indicates that it is too dark. Then, gamma correction is performed on each image sub-region. The gamma value for gamma correction is dynamically adjusted according to the local illumination deviation value: for overly bright regions (i.e., positive deviation value), a gamma value less than 1 is used for compression to reduce its brightness; for overly dark regions (i.e., negative deviation value), a gamma value greater than 1 is used for expansion to increase its brightness; for regions with normal illumination, the gamma value is equal to 1 and remains unchanged. The aim is to non-linearly correct the brightness distortion caused by uneven illumination, making the brightness response of different regions more consistent.

[0069] Subsequently, a contrast stretching transformation is performed. First, the mean pixel intensity μ and standard deviation σ of the image sub-region after gamma correction are calculated. Then, a lower limit threshold I for contrast stretching is set. low = max(0, μ - 2×σ) and upper limit threshold I up =min(255, μ+2 ×σ). Linearly map the intensity values ​​I of pixels in the image sub-region to the interval [0, 255]: if I ≤ I low If I ≥ I, then the mapping is 0; up If I, then it is mapped to 255; low <I<I up The mapping is then (I - I) low )×255 / (I up -I low The purpose of this step is to extend the effective contrast of each image sub-region to the entire dynamic range, enhancing the distinguishability of features. Finally, all processed image sub-regions are stitched together according to their original positions to form the illumination-equalized target image.

[0070] In this embodiment of the invention, adaptive illumination equalization directly addresses the most challenging issue of variable illumination in the vehicle environment. Instead of simple global adjustment, it achieves adaptive correction of local illumination conditions by partitioning, calculating local deviations, and dynamically adjusting gamma values ​​and contrast stretching parameters. This makes the brightness and contrast distribution of different images acquired under varying illumination conditions more consistent and normalized after processing. Secondly, since the input to the feature extraction model is an equalized, more stable target image, the initial biometric vectors extracted from these images exhibit higher intra-class compactness (i.e., feature vectors of the same user are more similar under different illuminations) and inter-class discriminability (i.e., feature vector differences between different users are more pronounced). This fundamentally improves the reliability of subsequent fuzzy binding based on biometric vectors, reduces authentication failures caused by illumination issues, and thus achieves reliable and convenient cross-vehicle authentication.

[0071] In the above embodiments, the present invention does not rely on a real-time online central server. The present invention also provides a solution for quickly and reliably locating the source vehicle without the intervention of a central server. Specifically, mapping the user's identity identifier to the associated vehicle identifier includes: querying a locally stored distributed mapping table, which is updated based on successful authentication events broadcast between vehicles; wherein each successful authentication event includes a triplet of information: user identity identifier, source vehicle identifier, and timestamp; if multiple vehicle identifiers corresponding to the user identity identifier are found in the distributed mapping table, the nearest vehicle identifier is selected based on the timestamp.

[0072] In practice, each authorized vehicle in the fleet first maintains a local data structure, namely a distributed mapping table. Each record in the distributed mapping table is essentially a triple: user identifier, source vehicle identifier, and timestamp. The user identifier is the user's unique identifier, the source vehicle identifier is the unique identifier of the vehicle that most recently successfully authenticated the user, and the timestamp is the time of that successful authentication. The distributed mapping table is updated by authentication success events broadcast between vehicles. Whenever a vehicle (as the target vehicle) successfully authenticates a user, it generates and broadcasts an authentication success event message, which contains at least the aforementioned triple information. Other vehicles in the fleet network listen for the broadcast message via inter-vehicle communication. Upon receiving the broadcast message, they extract the triple information from the message and use it to update their local distributed mapping table. The update logic is: using the user identifier as the key, check if a record for that user already exists in the local table. If it does not exist, the new triple record is directly inserted. If the record already exists, the newly received timestamp is compared with the timestamp in the stored record. The old record is overwritten only if the new timestamp is later (i.e., updated). The new triplet information (containing the new source vehicle identifier and the new timestamp) is used to overwrite the old record.

[0073] When a user requests a ride to a new target vehicle, the target vehicle needs to locate the source vehicle that can verify the user's identity. The steps are as follows: The target vehicle obtains the user's identity identifier from the ride request; then, it queries its locally stored distributed mapping table, using the user's identity identifier as the key. Several scenarios may occur: First, only one record matching the user's identity identifier exists in the table. In this case, the target vehicle can directly read the source vehicle identifier from that record, which is the target vehicle to initiate the verification request. Second, multiple records with the same user identity identifier but different source vehicle identifiers and timestamps exist in the table. This indicates that the user may have been successfully authenticated on multiple different vehicles in the past. According to this embodiment, the target vehicle will select the record with the latest timestamp based on the timestamp field and use its source vehicle identifier as the current source vehicle to contact. Third, no record matching the user's identity identifier is found in the table. This may occur when the user is first authenticated in the fleet network, and the successful authentication event has not yet propagated to the current target vehicle, or when the target vehicle has just joined the network.

[0074] The method for rapidly locating the source vehicle described above reuses the authentication success event broadcast to transmit the user-vehicle mapping relationship, achieving self-organized maintenance of the distributed mapping table without increasing additional communication overhead. When the source vehicle needs to be located, the target vehicle only needs to perform a fast local table lookup, achieving extremely low location latency and meeting the requirements of rapid response for vehicle access control. The adoption of a timestamp-based most recent authentication source selection strategy aligns with user movement patterns, effectively improving the hit rate of authentication requests and overall process efficiency. This distributed scheme is resistant to single points of failure and network partitions, and can adapt to dynamic changes in the fleet and user behavior patterns.

[0075] Example 2: Referring to Figure 4, this embodiment of the invention also provides a vehicle distributed identity authentication system based on biometrics, including:

[0076] Multiple vehicle units, each vehicle unit comprising:

[0077] The biometric data acquisition module is used to collect users' biometric data;

[0078] The local storage module is used to store binding factors, biometric hash values, and user identity identifiers;

[0079] A communication module for communication between vehicles;

[0080] The processing module is configured to execute the above-described biometric-based distributed vehicle identity authentication method;

[0081] The central management server communicates with multiple vehicle units and is used to manage user access policies and trust chain configurations.

[0082] The embodiments of the present invention are based on the same inventive concept as Embodiment 1 and have the same beneficial effects, which will not be repeated here.

[0083] In summary, the biometric-based vehicle distributed identity authentication method and system described in this invention can achieve efficient, secure, and repetitive cross-vehicle identity authentication by using a decentralized hash passing and fuzzy binding verification mechanism, while protecting the user's biometric privacy.

[0084] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A biometric-based distributed vehicle identity authentication method, characterized in that, include: Obtain the initial biometric vector collected by the user on the source vehicle; bind the initial biometric vector and a random key using a fuzzy binding algorithm to obtain a binding factor; and generate a biometric hash value based on the random key. The binding factor, the biometric hash value, the user identity identifier, and the vehicle identifier are associated and stored together. When a user initiates a car rental request to the target vehicle, the user identity identifier is obtained. The user identity identifier is mapped to the associated vehicle identifier, the source vehicle for verifying the user is determined based on the vehicle identifier, and a verification request is initiated to the source vehicle. In response to the verification request, the source vehicle sends the binding factor and the biometric hash value to the target vehicle. The target vehicle acquires the user's current biometric vector, and based on the received binding factor and the current biometric vector, obtains a candidate key through a decoding operation; and generates a temporary biometric hash value based on the candidate key. The temporary biometric hash value is compared with the received biometric hash value; if they match, the target vehicle is successfully authenticated, and the user is granted vehicle access. The initial biometric vector and the random key are bound together using a fuzzy binding algorithm to obtain a binding factor. This process includes: binarizing the initial biometric vector to obtain an initial binarized feature vector; performing error-correcting encoding on the random key to obtain an encoded key; performing a bitwise XOR operation on the encoded key and the initial binarized feature vector to obtain the binding factor; and obtaining a candidate key through decoding based on the received binding factor and the current biometric vector. This includes binarizing the current biometric vector to obtain a current binarized feature vector; performing a bitwise XOR operation on the current binarized feature vector and the received binding factor to obtain a candidate encoded key; and performing error-correcting decoding on the candidate encoded key to obtain the candidate key.

2. The vehicle distributed identity authentication method based on biometrics according to claim 1, characterized in that, Obtaining the user's initial biometric vector includes: acquiring the user's initial facial image through an in-vehicle image acquisition device; performing adaptive illumination equalization processing on the initial facial image to obtain an initial target image; and extracting features from the initial target image to obtain the initial biometric vector.

3. The biometric-based distributed vehicle identity authentication method according to claim 1, characterized in that, Obtaining the user's current biometric vector includes: acquiring the user's current facial image through an in-vehicle image acquisition device; performing adaptive illumination equalization processing on the current facial image to obtain a current target image; and extracting features from the current facial image to obtain the current biometric vector.

4. The biometric-based distributed vehicle identity authentication method according to claim 2 or 3, characterized in that, The adaptive illumination equalization processing includes: partitioning the first image to obtain multiple image sub-regions; calculating the average pixel intensity of each image sub-region; calculating the weighted average of the average pixel intensities of all image sub-regions contained in the first image to obtain a global illumination reference value; calculating the deviation between the average pixel intensity of each image sub-region and the global illumination reference value to obtain a local illumination deviation value; performing gamma correction on the image sub-regions; wherein, if the local illumination deviation value is positive, the gamma value of the gamma correction is less than 1; if the local illumination deviation value is negative, the gamma value of the gamma correction is greater than 1; if the local illumination deviation value is 0, the gamma value of the gamma correction is equal to 1; and obtaining the target first image based on the image sub-regions after gamma correction processing.

5. The biometric-based distributed vehicle identity authentication method according to claim 4, characterized in that, After gamma correction of the image sub-region, the method further includes contrast stretching transformation of the image sub-region; the contrast stretching transformation includes: calculating the mean μ and standard deviation σ of the pixel intensity of the image sub-region after gamma correction; and setting an upper limit threshold I for the contrast stretching transformation. low and lower limit threshold I up :I low =max(0, μ-2×σ); I up =min(255, μ+2×σ); The intensity value I of each pixel within the image sub-region is mapped as follows: if I≤I low If I ≥ I, then the output pixel intensity is 0; up Then the output pixel intensity is 255; if I low <I<I up The output pixel intensity is: The image sub-regions after the contrast stretching transformation are stitched together to obtain the target first image.

6. The biometric-based distributed vehicle identity authentication method according to claim 1, characterized in that, Generating a biometric hash value based on the random key includes: performing a hash operation on the random key to obtain the biometric hash value; wherein the hash function of the hash operation is a non-cryptographic hash function.

7. The vehicle distributed identity authentication method based on biometrics according to claim 1, characterized in that, Mapping the user's identity identifier to the associated vehicle identifier includes: querying a locally stored distributed mapping table, which is updated based on authentication success events broadcast between vehicles; wherein each authentication success event contains a triplet of information: user identity identifier, source vehicle identifier, and timestamp; if multiple vehicle identifiers are found to correspond to the user identity identifier in the distributed mapping table, the nearest vehicle identifier is selected based on the timestamp.

8. A biometric-based distributed vehicle identity authentication system, characterized in that, include: Multiple vehicle units, each vehicle unit comprising: a biometric acquisition module for acquiring user biometric data; a local storage module for storing binding factors, biometric hash values, and user identity identifiers; a communication module for inter-vehicle communication; a processing module configured to execute the biometric-based vehicle distributed identity authentication method as described in any one of claims 1-7; and a central management server communicatively connected to the multiple vehicle units for managing user access policies and trust chain configurations.

Images