CN118797598A - A lightweight user authentication method and system based on device fingerprint - Google Patents

Abstract

The present invention discloses a lightweight user authentication method and system based on device fingerprint, belonging to the technical field of electric power Internet of Things. The lightweight user authentication method based on device fingerprint of the present invention uses the trace of static random access memory SRAM as the device fingerprint by constructing a trace generation model, a cluster coupling model, an abnormal calculation model, and a user authentication model; clusters the data at the same time, selects objects that deviate from their cluster centers as trace objects, and measures the abnormal values of the trace objects using local abnormal value factors; and then obtains the ranking of user detection according to the abnormal value factor, and removes users with high abnormal values, thereby avoiding false detection, and obtaining lightweight legal user authentication, with high user authentication efficiency and good flexibility, so that the Internet of Things node can respond quickly, and is suitable for scenarios where a large number of requests need to be processed quickly, and the scheme is scientific, reasonable, and feasible.

Description

A lightweight user authentication method and system based on device fingerprint

Technical Field

The invention relates to a lightweight user authentication method and system based on device fingerprint, belonging to the technical field of electric power Internet of Things.

Background Art

Driven by the goal of reducing carbon emissions, the penetration rate of distributed energy in my country has steadily increased. A large number of IoT nodes have been introduced into the new power grid with distributed new energy as the main body to collect data on distributed energy access, transmission network, distribution network and operation and energy consumption of end users. The interactive business instruction interaction and information data transmission brought about by the collection process are mostly transmitted through wireless public networks. These interactive data are easily attacked by physical attacks, resulting in data tampering.

Furthermore, the Chinese patent (publication (announcement) number: CN110620820A) The present invention provides a ubiquitous power Internet of Things intelligent management system, including a device terminal and a management terminal; the device terminal is provided with an information collection module and an information sending module; the information collection module is used to collect the Internet of Things data of the device terminal; the information sending module is used to send the Internet of Things data to each node in the blockchain network to judge the legitimacy of the Internet of Things data. If the Internet of Things data is legal, the Internet of Things data is written into the management terminal and the blockchain respectively. By sending the Internet of Things data to each node in the blockchain network to judge the legitimacy of the Internet of Things data, if the Internet of Things data is legal, the Internet of Things data is written into the management terminal and the blockchain respectively to prevent data loss or change, and to prevent the traditional structured center mode from failing. If the terminal fails, the power grid data management is paralyzed and affects people's normal life, ensuring the normal operation of the power grid.

The above-mentioned method utilizes each node in the blockchain network to judge the legitimacy of IoT data, but the processing and storage capabilities of IoT nodes are low, and they usually have low-bandwidth communication channels. Therefore, the above-mentioned node processing solution has a large node resource overhead, making it impossible for IoT nodes to respond quickly. Therefore, the above-mentioned solution cannot be applied to scenarios that require rapid processing of a large number of requests, resulting in low user authentication efficiency and poor flexibility.

The information disclosed in this Background Art is only for understanding the background of the inventive concept and therefore it may include information that does not constitute the prior art.

Summary of the invention

In response to the above problem or one of the above problems, an object of the present invention is to provide a lightweight user authentication method based on device fingerprint. By constructing a trace generation model, a cluster coupling model, an abnormal calculation model, and a user authentication model, a lightweight legitimate user authentication is obtained. The user authentication efficiency is high and the flexibility is good, so that the Internet of Things nodes can respond quickly. It is suitable for scenarios that need to process a large number of requests quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The solution is scientific, reasonable, and feasible.

In response to the above problem or one of the above problems, the second object of the present invention is to provide a lightweight user authentication method and system based on device fingerprint, which uses static random access memory SRAM traces to characterize device fingerprints and use them as unique identifiers to accurately distinguish devices; at the same time, a clustering algorithm is used to cluster data, and objects that deviate from their cluster centers are selected as abnormal candidate sets, and local outlier factors are used to measure outliers in the abnormal candidate sets, which can improve the outlier overlap problem of the local outlier factor method; and the ranking of user detection is obtained through the outlier factor, and users with high outliers are eliminated by sorting the rankings in descending order, thereby avoiding false detection and obtaining lightweight legitimate user authentication.

In response to the above problem or one of the above problems, the third object of the present invention is to provide a lightweight user authentication system based on device fingerprint. By setting up the collection layer, access layer and application layer, the authentication of lightweight legitimate users is realized. The authentication efficiency is high and the flexibility is good, so that the Internet of Things nodes can respond quickly. It is suitable for scenarios where a large number of requests need to be processed quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The solution is scientific, reasonable and feasible.

To achieve one of the above purposes, the first technical solution of the present invention is:

A lightweight user authentication method based on device fingerprint comprises the following steps:

In the first step, the trace generation model built in advance is used to collect trace samples of static random access memory (SRAM) and construct a trace data set to characterize device fingerprints.

The second step is to cluster the trace data set using the cluster coupling model constructed in advance to obtain one or more trace objects that deviate from their cluster center;

The third step is to use the anomaly calculation model constructed in advance to measure the anomaly information of one or more trace objects based on the local outlier factor;

The fourth step is to complete the authentication of lightweight users based on abnormal information through the user authentication model built in advance.

After continuous exploration and experimentation, the present invention constructs a trace generation model, a cluster coupling model, an abnormal calculation model, and a user authentication model, and uses the trace of a static random access memory SRAM as a device fingerprint, that is, a unique identifier of a device to achieve accurate distinction of the device; at the same time, the data is clustered, and objects that deviate from their cluster centers are selected as trace objects. The outlier value of the trace object is measured by a local outlier factor, which can improve the outlier overlap problem of the local outlier factor method; and then the ranking of user detection is obtained according to the outlier factor, and the ranking is sorted in descending order to eliminate users with high outliers, thereby avoiding false detection and obtaining lightweight legal user authentication, with high user authentication efficiency and good flexibility; therefore, the node resource overhead of the scheme of the present invention is small, so that the Internet of Things node can respond quickly, and is suitable for scenarios where a large number of requests need to be processed quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The scheme is scientific, reasonable, and feasible.

Furthermore, since the static random access memory SRAM exists in the onboard controller of the IoT node in the acquisition layer and is used to store the acquired data, and each node has a static random access memory, the trace of the static random access memory SRAM can be used as a device fingerprint feature, that is, a unique identifier, to achieve accurate distinction of devices.

Furthermore, since the local outlier factor (LOF) can be used to detect outliers, but cannot handle overlapping outliers, the present invention introduces cluster coupling model clustering, uses the cluster coupling model to analyze the user's static random access memory SRAM trace, lists the users far away from the cluster center as abnormal candidates, and uses the local abnormal factor to calculate the abnormal degree of the abnormal candidate, so as to authenticate the legitimate user.

As the preferred technical measures:

In the first step, the method of constructing the trace data set through the trace generation model is as follows:

Step 11: Determine the static random access memory SRAM of all user devices of the collection layer to be authenticated;

Step 12: Collect power-on SRAM trace samples from each SRAM by turning the power on and off;

Step 13: Assemble the trace samples to obtain a trace data set.

As the preferred technical measures:

In the second step, the method of clustering the trace data set using the cluster coupling model is as follows:

Step 21: Obtain the trace data set and initialize the clustering parameters;

Step 22: Use the Gaussian mixture algorithm to process the trace data set to obtain corresponding trace clusters;

Step 23: Based on the trace clustering, the number of trace clusters of the trace data set is calculated;

Step 24: Calculate the probability that each trace sample belongs to each trace cluster;

Step 25: Maximize the probability that the trace sample belongs to a certain trace cluster, and divide the trace sample into a certain trace cluster;

Step 26: Iterate steps 3, 4, and 5 until all trace samples are clustered.

As the preferred technical measures:

In the second step, the method of obtaining the trace object deviating from its cluster center by using the cluster coupling model is as follows:

S21: After clustering is completed, for any trace cluster, the cluster center is calculated;

S22: Calculate the distance between each trace sample in the trace cluster and the center of each cluster to obtain deviation data;

S23: based on the deviation data, obtaining one or more trace objects deviating from the cluster center;

S24: Aggregate one or more trace samples that deviate from the cluster center to form a candidate set of trace anomaly data.

As the preferred technical measures:

The formula for calculating the distance of a trace sample to the center of its respective cluster is as follows:

Among them, ^xj is the cluster center corresponding to the trace sample, is the trace sample, is the distance from the trace sample to the center of each cluster;

Or/and, the method to obtain the trace object that deviates from its cluster center is as follows:

The trace anomaly dataset A is calculated for all trace clusters as follows:

To satisfy The trace samples of are regarded as trace objects that deviate from their cluster centers and are included in the trace anomaly dataset A;

Where |Dj| is the size of the jth cluster, |D| is the trace dataset, and δ is the value of n in the local outlier factor. is the triple standard deviation from the jth cluster.

As the preferred technical measures:

In the third step, the method of using the abnormal calculation model to measure abnormal information is as follows:

Step 31: for the trace sample x in the trace data set, calculate the n nearest neighbors N _n (x) closest to the trace sample x in the trace data set, and the nth nearest distance dist _n (y) between the trace sample in the trace data set and the trace sample y;

Step 32: According to the nth nearest distance dist _n (y), calculate the reachable distance Rd(x, y) between two trace samples x and y in the trace data set. The calculation formula is as follows:

Rd(x,y)=max{dist(x,y),dist _n (y)};

Where dist(x,y) represents the Euclidean distance between two trace samples x and y;

Step 33: According to the reachable distance Rd(x,y), calculate the local reachable density _ρn (x) of the trace sample x, and the calculation formula is as follows:

Where n is the nth nearest neighbor of x, It represents the reciprocal of the average distance between trace sample x and trace sample y when selecting a trace sample y from the nearest neighbors of trace sample x;

Step 34: Based on the local reachability density _ρn (x), calculate the ratio of the local reachability density of the trace sample x to the trace sample y in its nearest neighbor corresponding to a specific n, _LOFn (x), which is used to measure the outlier degree of x and its nearest neighbor. The calculation formula is as follows:

Among them, ρ _n (y) represents the local reachability density between the trace sample y and a trace sample selected from its nearest neighbors;

Step 35: The local reachable density ratio _LOFn (x) is used as the outlier of the trace sample x, and multiple outliers are arranged in order to obtain outlier information.

As the preferred technical measures:

In the fourth step, the method for completing the authentication of lightweight users through the user authentication model is as follows:

Step 41: sorting the trace samples in the trace anomaly data set according to the size of the anomaly value in the anomaly information to obtain user ranking data;

Step 42: Based on the user ranking data, users with high average rankings are eliminated to obtain the authorized user authentication result.

To achieve one of the above purposes, the second technical solution of the present invention is:

A lightweight user authentication method based on device fingerprint includes the following contents:

Collect trace samples of static random access memory (SRAM) and construct a trace data set to characterize device fingerprints. Perform clustering on the trace data set and select trace objects that deviate from their cluster centers as trace anomaly candidate sets.

Based on the local outlier factor, measure the outliers in the trace anomaly candidate set;

Based on outliers, lightweight user authentication is completed.

The present invention characterizes the device fingerprint by using the static random access memory (SRAM) trace, and uses it as a unique identifier to accurately distinguish the device. At the same time, the present invention clusters the data using a clustering algorithm, selects objects that deviate from their cluster centers as an abnormal candidate set, and uses a local abnormal value factor to measure the abnormal values in the abnormal candidate set, thereby improving the abnormal value overlap problem of the local abnormal value factor method. The ranking of user detection is obtained through the abnormal value factor, and the ranking is sorted in descending order to eliminate users with high abnormal values, thereby avoiding false detection and obtaining lightweight legal user authentication.

To achieve one of the above purposes, the third technical solution of the present invention is:

A lightweight user authentication system based on device fingerprint, which adopts the above-mentioned lightweight user authentication method based on device fingerprint, includes a collection layer, an access layer and an application layer;

The collection layer is used to collect node data of distributed energy, transmission network equipment, distribution network equipment and end users. It deploys IoT nodes and trace extraction modules.

The access layer is equipped with 4G and/or 5G base stations for transmitting node data;

The application layer is the power grid business cloud, which is used to process node data. It deploys an abnormal calculation module based on cluster coupling model and outlier detection, as well as a user authentication module that can eliminate abnormal users.

The node data collected by the IoT nodes is connected to the power grid business cloud through 4G and/or 5G base stations to perform outlier detection and eliminate abnormal users, thereby achieving lightweight user authentication.

The present invention realizes lightweight authentication of legitimate users by setting up a collection layer, an access layer and an application layer. The authentication efficiency is high and the flexibility is good, so that the Internet of Things nodes can respond quickly. It is suitable for scenarios where a large number of requests need to be processed quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The solution is scientific, reasonable and feasible.

As the preferred technical measures:

The trace extraction module is used to extract the static random access memory SRAM traces of all user devices in the acquisition layer, and collect the trace samples of the power-on static random access memory SRAM from each static random access memory SRAM by turning the power on and off.

Or/and, an anomaly calculation module, used to cluster all trace samples using a cluster coupling model, select trace objects that deviate from their cluster centers as trace anomaly candidate sets, and measure the outliers in the trace anomaly candidate sets using a local outlier factor.

or/and, a user authentication module, which is used to obtain the user ranking at the detection time by using outlier detection. If the average ranking of the detection is always at the front, the user corresponding to the outlier is eliminated to complete the user authentication;

Or/and, the node data at least includes operation data and energy consumption data;

Or/and, the IoT nodes include smart devices and/or smart meters.

To achieve one of the above purposes, the fourth technical solution of the present invention is:

An electronic device comprising:

one or more processors;

A storage device for storing one or more programs;

When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned lightweight user authentication method based on device fingerprint.

To achieve one of the above purposes, the fifth technical solution of the present invention is:

A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned lightweight user authentication method based on device fingerprint.

Compared with the prior art solutions, the present invention has the following beneficial effects:

After continuous exploration and experimentation, the present invention constructs a trace generation model, a cluster coupling model, an abnormal calculation model, and a user authentication model, and uses the trace of a static random access memory SRAM as a device fingerprint, that is, a unique identifier of a device to achieve accurate distinction of the device; at the same time, the data is clustered, and objects that deviate from their cluster centers are selected as trace objects. The outlier value of the trace object is measured by a local outlier factor, which can improve the outlier overlap problem of the local outlier factor method; and then the ranking of user detection is obtained according to the outlier factor, and the ranking is sorted in descending order to eliminate users with high outliers, thereby avoiding false detection and obtaining lightweight legal user authentication, with high user authentication efficiency and good flexibility; therefore, the node resource overhead of the scheme of the present invention is small, so that the Internet of Things node can respond quickly, and is suitable for scenarios where a large number of requests need to be processed quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The scheme is scientific, reasonable, and feasible.

Furthermore, the present invention characterizes the device fingerprint by using the static random access memory SRAM trace, and uses it as a unique identifier to accurately distinguish the device; at the same time, the present invention clusters the data using a clustering algorithm, selects objects that deviate from their cluster centers as an abnormal candidate set, and uses a local outlier factor to measure the outliers in the abnormal candidate set, which can improve the outlier overlap problem of the local outlier factor method; and obtains the ranking of user detection through the outlier factor, and uses the descending order of the ranking to eliminate users with high outliers, thereby avoiding false detection and obtaining lightweight legal user authentication.

Furthermore, the present invention realizes lightweight authentication of legitimate users by setting up a collection layer, an access layer and an application layer. The authentication efficiency is high and the flexibility is good, so that the Internet of Things nodes can respond quickly. It is suitable for scenarios that need to process a large number of requests quickly, and is particularly suitable for low-bandwidth communication channels with low processing and storage capabilities. The solution is scientific, reasonable and feasible.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a flow chart of a lightweight user authentication method based on device fingerprint of the present invention;

FIG2 is a schematic diagram of a novel power grid architecture in which legal user authentication is introduced in the present invention;

FIG3 is another schematic flow chart of a lightweight user authentication method based on device fingerprint of the present invention;

FIG. 4 is a schematic diagram of a flow chart of calculating LOF according to the present invention.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the detailed description of the present invention below. Those skilled in the art can fully understand the present invention without the description of these details.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention pertains. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "or/and" as used herein includes any and all combinations of one or more of the associated listed items.

As shown in FIG1 , the first specific embodiment of the lightweight user authentication method based on device fingerprint of the present invention is as follows:

A lightweight user authentication method based on device fingerprint comprises the following steps:

In the first step, the trace generation model built in advance is used to collect trace samples of static random access memory (SRAM) and construct a trace data set to characterize device fingerprints.

The second step is to cluster the trace data set using the cluster coupling model constructed in advance to obtain one or more trace objects that deviate from their cluster center;

The third step is to use the anomaly calculation model constructed in advance to measure the anomaly information of one or more trace objects based on the local outlier factor;

The fourth step is to complete the authentication of lightweight users based on abnormal information through the user authentication model built in advance.

A second specific embodiment of the lightweight user authentication method based on device fingerprint of the present invention:

A lightweight user authentication method based on device fingerprint comprises the following steps:

The first step is to collect trace samples of static random access memory (SRAM) through the trace generation model built in advance, and construct a trace dataset to characterize the device fingerprint.

The second step is to cluster the trace data set using the cluster coupling model constructed in advance, and select the trace objects that deviate from their cluster centers as the trace anomaly candidate set;

The third step is to use the anomaly calculation model constructed in advance to measure the outliers in the trace anomaly candidate set based on the local outlier factor;

The fourth step is to complete the authentication of lightweight users based on outliers through the user authentication model built in advance.

The third specific embodiment of the lightweight user authentication method based on device fingerprint of the present invention:

A lightweight user authentication method based on device fingerprint includes the following contents:

Collect the static random access memory SRAM trace samples of the IoT device when it is powered on to form a data set D, use the cluster coupling model to cluster the data, select the objects that deviate from their cluster center as the abnormal candidate set A, use the local outlier factor to measure the outliers in A, and obtain the ranking at the detection time. If the average ranking of the detection is always in the front (the first 3% of the entire data volume), remove the outlier and obtain user authentication. The present invention uses the static random access memory SRAM trace of the IoT node as the only legal identifier of the device, and uses lightweight clustering and local outlier detection to obtain the authentication of legal users. The scheme is scientific, reasonable, and ingenious.

Static random access memory SRAM exists in the onboard controller of the IoT node in the acquisition layer and is used to store the collected data. The SRAM structure is divided into a data part, a high variance stack, and a low variance stack. The mean and variance of the high variance stack of the device are very different, and can be used as a device fingerprint to uniquely calibrate the device. The static random access memory SRAM trace shows the size of the high variance stack of the static random access memory SRAM. The high variance stack consists of bytes. The bytes are converted into corresponding integer values and reduced by 255 times to the range of [0,1].

The method for selecting objects that are away from their cluster center is as follows:

S1. Calculate the cluster center x ^j first;

S2. Then calculate the distance from the trace sample to the center of each cluster. The calculation formula is as follows:

S3. Calculate the triple standard deviation d _c ^j from the jth cluster;

S4. Classify the abnormal candidate set A that meets the following judgment conditions

Wherein, |Dj| is the size of the jth cluster, |D| is the size of the entire data set D, and δ is a parameter, which is 0.05 in this embodiment.

Since IoT nodes have low processing and storage capabilities and usually have low-bandwidth communication channels, this limits the use of security solutions such as public key encryption algorithms based on cryptography. In the new power grid architecture that introduces legitimate user authentication (as shown in Figure 2), each node has a static random access memory SRAM, so the trace of the static random access memory SRAM is used as the fingerprint feature of the device.

Meanwhile, the local outlier factor (LOF) can be used to detect outliers, but it cannot handle overlapping outliers. Therefore, the present invention introduces cluster coupling model clustering, uses the cluster coupling model to analyze the user's static random access memory SRAM trace, lists the users far away from the cluster center as abnormal candidates, and uses the local abnormal factor to calculate the abnormal degree of the abnormal candidate, thereby authenticating the legitimate user.

As shown in FIG3 , a fourth specific embodiment of the lightweight user authentication method based on device fingerprint of the present invention is as follows:

A lightweight user authentication method based on device fingerprint is proposed. It is necessary to collect static random access memory (SRAM) trace samples of each IoT node in the new power grid acquisition layer as device fingerprints, establish a lightweight user authentication method based on device fingerprints, and authenticate legitimate users through the authentication method. The specific implementation steps are as follows:

Step 1: Obtain the trace data set D of the static random access memory SRAM and initialize the relevant parameters n,δ;

Step 2: Obtain the number of clusters and the corresponding clusters of the Gaussian mixture algorithm clustering, and the acquisition method includes the following:

The obtained device fingerprints of IoT nodes (trace data set D) are divided into k cluster groups, that is, into k Gaussian distributions, i.e. clusters. The k Gaussian distributions are mixed and superimposed to form a Gaussian mixture distribution (cluster). Each Gaussian distribution has a variance, a mean, and an external mixing coefficient.

Step 3: Calculate the number of Gaussian models (clusters); Step 4: Calculate the probability that each trace sample belongs to each cluster. The specific method is as follows:

The expectation maximization algorithm is used to find 3k parameters, and the Bayesian formula is used to find the posterior probability that each trace sample belongs to a Gaussian distribution (cluster);

Step 5: Maximize the probability that the trace sample belongs to a certain cluster. The specific method is as follows:

The Bayesian formula is used to calculate the posterior probability that each node data (trace sample) belongs to the first, second, ..., up to the kth Gaussian distribution. The node data is divided into the corresponding cluster with the largest posterior probability.

Step 6: Iterate steps 3 and 4 until convergence;

Step 7: For any cluster, calculate the cluster center and the distance from each object in the cluster to its own cluster center;

Step 8: Obtain an abnormal data candidate set A consisting of objects that deviate from the center;

Step 9: Calculate the abnormal data candidate set for all clusters;

Step 10: Calculate the local outlier factor LOF;

Step 11: Obtain the ranking of abnormal data candidate sets according to LOF;

Step 12: Eliminate users with high average rankings and obtain the authentication result of authorized users, where the authorized users refer to legitimate IoT nodes connected to the power grid and recognized by the power grid.

The present invention also provides a novel power grid system for lightweight legitimate user authentication based on the power-on static random access memory (SRAM) trace samples of IoT devices, clustering using cluster coupling models and local outlier factor measurement. The system mainly includes a collection layer, which deploys IoT nodes (smart devices, smart meters, etc.) to collect data on distributed energy access, transmission and distribution networks, and operation and energy consumption of end users. At the access layer, the collected data is accessed through 4G and 5G base stations to the new power grid service cloud at the application layer.

In this embodiment, the specific implementation steps of calculating LOF are as follows (as shown in FIG4 ):

S1: The node establishes a communication connection with the base station;

S2: IoT nodes use Gaussian model to obtain clusters;

S3: For any cluster, calculate the cluster center ^xj and the distance from each object to its own sub-cluster center ^xj Calculate the triple standard deviation of the jth cluster

S4: Satisfaction The objects are included in the abnormal data set A;

Among them, |Dj| is the size of the jth cluster, |D| is the trace data set, and δ is the value taken according to n in LOF;

S5: Calculate A for all clusters;

S6: For an object x in a dataset D, calculate the n nearest neighbors N _n (x) closest to x in D, and the nth nearest distance dist _n (y) between the object in D and y;

S7: Calculate the reachable distance Rd(x,y) between two objects x and y in the data set D. The calculation formula is as follows:

Rd(x,y)=max{dist(x,y),dist _n (y)};

Among them, dist(x,y) represents the Euclidean distance between two objects x and y.

S8: Calculate the local reachability density ρ _n (x) of x, which is calculated as follows:

Where n is the nth nearest neighbor of x, It represents the reciprocal of the average reachable distance between x and y when selecting an object y from the nearest neighbors of x.

S9: Calculate the ratio of the local reachability density of x to the object y in its nearest neighbor corresponding to a specific n, LOF _n (x), which is used to measure the outlier degree of x and its nearest neighbor. The calculation formula is as follows:

Among them, ρ _n (y) represents the local reachability density between y and an object selected from its nearest neighbors.

The first specific embodiment of the lightweight user authentication system based on device fingerprint of the present invention:

A lightweight user authentication system based on device fingerprint, which adopts the above-mentioned lightweight user authentication method based on device fingerprint, includes a collection layer, an access layer and an application layer;

The collection layer is used to collect node data of distributed energy, transmission network equipment, distribution network equipment and end users. It deploys IoT nodes and trace extraction modules.

The access layer is equipped with 4G and/or 5G base stations for transmitting node data;

The application layer is the power grid business cloud, which is used to process node data. It deploys an abnormal calculation module based on cluster coupling model and outlier detection, as well as a user authentication module that can eliminate abnormal users.

The node data collected by the IoT nodes is connected to the power grid business cloud through 4G and/or 5G base stations to perform outlier detection and eliminate abnormal users, thereby achieving lightweight user authentication.

The second specific embodiment of the lightweight user authentication system based on device fingerprint of the present invention:

A lightweight user authentication system based on device fingerprint mainly includes three parts: static random access memory (SRAM) trace extraction module, anomaly calculation module based on cluster coupling model and outlier detection, and user authentication module for eliminating abnormal users.

Static Random Access Memory SRAM trace extraction module: Completes the static random access memory SRAM trace extraction of all user devices in the acquisition layer, and manually collects power-on static random access memory SRAM trace samples from each device by turning the power on and off.

Anomaly calculation module based on cluster coupling model and outlier detection: completes clustering of data using cluster coupling model, selects objects that deviate from their cluster centers as anomaly candidate set A, and measures outliers in A using local outlier factors.

User authentication module: Use the above-mentioned outlier detection to obtain the ranking at the detection time. If the average ranking of the detection is always at the front, remove the outlier and obtain user authentication.

An embodiment of a device applying the method of the present invention:

An electronic device comprising:

one or more processors;

A storage device for storing one or more programs;

When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned lightweight user authentication method based on device fingerprint.

A computer medium embodiment using the method of the present invention:

A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned lightweight user authentication method based on device fingerprint.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, and computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

The present application is described by flowcharts or/and block diagrams of the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each process or/and box in the flowchart or/and block diagram and the combination of the processes or/and boxes in the flowchart or/and block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate a device for implementing the functions specified in one process or multiple processes in the flowchart or/and one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims (10)

1. A lightweight user authentication method based on device fingerprint, characterized in that: The following steps are involved: In the first step, the trace generation model built in advance is used to collect trace samples of static random access memory (SRAM) and construct a trace data set to characterize device fingerprints. The second step is to cluster the trace data set using the cluster coupling model constructed in advance to obtain one or more trace objects that deviate from their cluster center; The third step is to use the anomaly calculation model constructed in advance to measure the anomaly information of one or more trace objects based on the local outlier factor; The fourth step is to complete the authentication of lightweight users based on abnormal information through the user authentication model built in advance. 2. A lightweight user authentication method based on device fingerprint as claimed in claim 1, characterized in that: In the first step, the method of constructing the trace data set through the trace generation model is as follows: Step 11: Determine the static random access memory SRAM of all user devices of the collection layer to be authenticated; Step 12: Collect power-on SRAM trace samples from each SRAM by turning the power on and off; Step 13: Assemble the trace samples to obtain a trace data set. 3. A lightweight user authentication method based on device fingerprint as claimed in claim 1, characterized in that: In the second step, the method of clustering the trace data set using the cluster coupling model is as follows: Step 21: Obtain the trace data set and initialize the clustering parameters; Step 22: Use the Gaussian mixture algorithm to process the trace data set to obtain corresponding trace clusters; Step 23: Based on the trace clustering, the number of trace clusters of the trace data set is calculated; Step 24: Calculate the probability that each trace sample belongs to each trace cluster; Step 25: Maximize the probability that the trace sample belongs to a certain trace cluster, and divide the trace sample into a certain trace cluster; Step 26: Iterate steps 3, 4, and 5 until all trace samples are clustered. 4. A lightweight user authentication method based on device fingerprint as claimed in claim 1, characterized in that: In the second step, the method of obtaining the trace object deviating from its cluster center by using the cluster coupling model is as follows: S21: After clustering is completed, for any trace cluster, the cluster center is calculated; S22: Calculate the distance between each trace sample in the trace cluster and the center of each cluster to obtain deviation data; S23: based on the deviation data, obtaining one or more trace objects deviating from the cluster center; S24: Aggregate one or more trace samples that deviate from the cluster center to form a candidate set of trace anomaly data. 5. A lightweight user authentication method based on device fingerprint as claimed in claim 4, characterized in that: The formula for calculating the distance of a trace sample to the center of its respective cluster is as follows: Among them, ^xj is the cluster center corresponding to the trace sample, is the trace sample, is the distance from the trace sample to the center of each cluster; Or/and, the method to obtain the trace object that deviates from its cluster center is as follows: The trace anomaly dataset A is calculated for all trace clusters as follows: To satisfy The trace samples of are regarded as trace objects that deviate from their cluster centers and are included in the trace anomaly dataset A; Where |D ^j | is the size of the jth cluster, |D| is the trace dataset, and δ is the value of n in the local outlier factor. is the triple standard deviation from the jth cluster. 6. A lightweight user authentication method based on device fingerprint as claimed in claim 1, characterized in that: In the third step, the method of using the abnormal calculation model to measure abnormal information is as follows: Step 31: for the trace sample x in the trace data set, calculate the n nearest neighbors N _n (x) closest to the trace sample x in the trace data set, and the nth nearest distance dist _n (y) between the trace sample in the trace data set and the trace sample y; Step 32: According to the nth nearest distance dist _n (y), calculate the reachable distance Rd(x, y) between two trace samples x and y in the trace data set. The calculation formula is as follows: Rd(x,y)=max{dist(x,y),dist _n (y)}; Where dist(x,y) represents the Euclidean distance between two trace samples x and y; Step 33: According to the reachable distance Rd(x,y), calculate the local reachable density _ρn (x) of the trace sample x, and the calculation formula is as follows: Where n is the nth nearest neighbor of x, It represents the reciprocal of the average distance between trace sample x and trace sample y when selecting a trace sample y from the nearest neighbors of trace sample x; Step 34: Based on the local reachability density _ρn (x), calculate the ratio of the local reachability density of the trace sample x to the trace sample y in its nearest neighbor corresponding to a specific n, _LOFn (x), which is used to measure the outlier degree of x and its nearest neighbor. The calculation formula is as follows: Among them, ρ _n (y) represents the local reachability density between the trace sample y and a trace sample selected from its nearest neighbors; Step 35: The local reachable density ratio _LOFn (x) is used as the outlier of the trace sample x, and multiple outliers are arranged in order to obtain outlier information. 7. A lightweight user authentication method based on device fingerprint as claimed in claim 1, characterized in that: In the fourth step, the method for completing the authentication of lightweight users through the user authentication model is as follows: Step 41: sorting the trace samples in the trace anomaly data set according to the size of the anomaly value in the anomaly information to obtain user ranking data; Step 42: Based on the user ranking data, users with high average rankings are eliminated to obtain the authorized user authentication result. 8. A lightweight user authentication method based on device fingerprint, characterized in that: Includes the following: Collect trace samples of static random access memory (SRAM) and construct a trace data set to characterize device fingerprints. Perform clustering on the trace data set and select trace objects that deviate from their cluster centers as trace anomaly candidate sets. Based on the local outlier factor, measure the outliers in the trace anomaly candidate set; Based on outliers, lightweight user authentication is completed. 9. A lightweight user authentication system based on device fingerprint, characterized in that: A lightweight user authentication method based on device fingerprint as described in any one of claims 1 to 8 is adopted, which includes a collection layer, an access layer and an application layer; The collection layer is used to collect node data of distributed energy, transmission network equipment, distribution network equipment and end users. It deploys IoT nodes and trace extraction modules. The access layer is equipped with 4G and/or 5G base stations for transmitting node data; The application layer is the power grid business cloud, which is used to process node data. It deploys an abnormal calculation module based on cluster coupling model and outlier detection, as well as a user authentication module that can eliminate abnormal users. The node data collected by the IoT nodes is connected to the power grid business cloud through 4G and/or 5G base stations to perform outlier detection and eliminate abnormal users, thereby achieving lightweight user authentication. 10. A lightweight user authentication system based on device fingerprint as claimed in claim 9, characterized in that: The trace extraction module is used to extract the static random access memory SRAM traces of all user devices in the acquisition layer, and collect the trace samples of the power-on static random access memory SRAM from each static random access memory SRAM by turning the power on and off; or/and, an anomaly calculation module, used for clustering all trace samples using a cluster coupling model, selecting trace objects deviating from their cluster centers as trace anomaly candidate sets, and measuring the outliers in the trace anomaly candidate sets using a local outlier factor; or/and, a user authentication module, which is used to obtain the user ranking at the detection time by using outlier detection. If the average ranking of the detection is always at the front, the user corresponding to the outlier is eliminated to complete the user authentication; Or/and, the node data at least includes operation data and energy consumption data; Or/and, the IoT nodes include smart devices and/or smart meters.