CN116709509B - User equipment sensing and positioning method, system, electronic equipment and storage medium - Google Patents

User equipment sensing and positioning method, system, electronic equipment and storage medium Download PDF

Info

Publication number
CN116709509B
CN116709509B CN202310993102.7A CN202310993102A CN116709509B CN 116709509 B CN116709509 B CN 116709509B CN 202310993102 A CN202310993102 A CN 202310993102A CN 116709509 B CN116709509 B CN 116709509B
Authority
CN
China
Prior art keywords
node
round
training
global
verifier
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310993102.7A
Other languages
Chinese (zh)
Other versions
CN116709509A (en
Inventor
张荣辉
景晓军
解彦曦
齐万彬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN202310993102.7A priority Critical patent/CN116709509B/en
Publication of CN116709509A publication Critical patent/CN116709509A/en
Application granted granted Critical
Publication of CN116709509B publication Critical patent/CN116709509B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/009—Security arrangements; Authentication; Protecting privacy or anonymity specially adapted for networks, e.g. wireless sensor networks, ad-hoc networks, RFID networks or cloud networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/06—Authentication
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/12—Detection or prevention of fraud
    • H04W12/121—Wireless intrusion detection systems [WIDS]; Wireless intrusion prevention systems [WIPS]
    • H04W12/122—Counter-measures against attacks; Protection against rogue devices
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/30—Security of mobile devices; Security of mobile applications
    • H04W12/37—Managing security policies for mobile devices or for controlling mobile applications
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a sensing and positioning method, a sensing and positioning system, electronic equipment and a storage medium of user equipment, and relates to the technical field of wireless positioning, wherein the sensing and positioning method comprises the following steps: acquiring equipment information of target user equipment; the device information includes: the position of the base station and the direction angle formed by the user equipment and the base station; determining a location of the target user device based on the device information of the target user device and the positioning model; the positioning model is determined by each node in the distributed network based on federal learning and blockchain; the node comprises: a compute node, a verifier node, and a witness node. The invention uses federal learning and block chain to replace the traditional central server, and has higher anti-attack capability while maintaining high positioning precision.

Description

User equipment sensing and positioning method, system, electronic equipment and storage medium
Technical Field
The present invention relates to the field of wireless positioning technologies, and in particular, to a method, a system, an electronic device, and a storage medium for sensing and positioning a user device.
Background
In recent years, with the development of emerging technologies such as autopilot, industrial internet and smart city, a high-speed increase of service demands based on location is triggered, and the conventional global navigation satellite system has the defects of weak signal strength, poor penetration, easy interference and the like, and cannot provide high-precision location services in indoor, underground and other environments. With the rapid development of wireless communication technology, the wireless communication system has the advantages of wide coverage, low use cost, high reliability and the like, and the communication system is utilized for positioning, so that the integration of communication and positioning is realized, and the wireless communication system becomes a hot spot of current research. The presence and movement of a positioning object within the coverage of a wireless transmission signal will cause a change in some parameters of the received signal, such as: channel state information, angle of arrival, received signal strength, time of flight, etc., and these varying parameters include relevant location information. Meanwhile, along with the outstanding performance of the deep learning technology in the aspect of feature extraction and the improvement of calculation force, the development of the communication positioning calculation integrated technology is promoted.
In order to improve the positioning accuracy in the communication positioning calculation integrated network, on one hand, distributed calculation force cooperation is needed to share calculation pressure, and on the other hand, efficient data communication between nodes is needed. However, edge devices tend to be massive and unreliable, which presents a significant potential safety hazard to the integrated network of sensory computing. When geographically dispersed nodes in a large-scale edge network transmit data, the nodes are not only easy to attack by the network, but also have privacy disclosure risks, and face the problems of interference resistance, data protection of the edge nodes, tracing and the like. In addition, the effect of deep learning requires a large amount of data support, but the data is often mastered by different data parties, and for safety and privacy problems, it is not practical to directly integrate all data or parameters into a server, and there is also a possibility that the server is paralyzed and data information is tampered.
Disclosure of Invention
The invention aims to provide a sensing and positioning method, a sensing and positioning system, electronic equipment and a storage medium of user equipment, which have higher attack resistance while maintaining high positioning accuracy.
In order to achieve the above object, the present invention provides the following solutions:
a method of sensing and locating a user device, comprising:
Acquiring equipment information of target user equipment; the device information includes: the position of a base station and a direction angle formed by user equipment and the base station;
determining a location of the target user device based on the device information and a positioning model of the target user device; the positioning model is determined by each node in a distributed network based on federal learning and blockchain; the node comprises: a compute node, a verifier node, and a witness node.
Optionally, the determining of the positioning model includes:
initializing parameters of a global model;
performing global training and local training on parameters of an initial global model by using equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node to obtain the positioning model; the computing node, the verifier node, and the witness node are determined by partitioning all nodes in a distributed network based on a cumulative incentive.
Optionally, performing global training on parameters of the initial global model and a first round of global training in the local training by using device information and true positions of a plurality of training user devices, a computing node, a verifier node and a witness node, wherein l is greater than or equal to 1, and the method comprises the following steps:
For the ith compute node:
determining the parameters of the global model after the first round of global training as the parameters of the initial local model of the first round of global training of the ith computing node; all computing nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training; when l=1, the parameters of the global model of the first round of global training are the parameters of the initial global model, and the accumulated rewards after the first round of global training are the accumulated rewards after the global training obtained by initialization;
performing M rounds of local training on parameters of an initial local model of the first round of global training by using a local data set of the ith computing node and equipment information of a plurality of training user equipment to obtain parameters of the local model of the ith computing node after the first round of global training, and determining node rewards of the ith computing node in the first round of global training; the local data set comprises equipment information and real positions of all training user equipment corresponding to the nodes;
the method comprises the steps of performing package signature on parameters of a local model after the first round of global training of an ith computing node, the time used for the first round of global training of the ith computing node, the local data set size of the ith computing node and node rewards of the ith computing node in the first round of global training to obtain computing node transaction information after the ith signature, and sending the computing node transaction information after the ith signature to an associated verifier node; one verifier node is associated with at least one computing node;
After receiving all corresponding signed computing node transaction information, for a j-th verifier node:
broadcasting all received signed computing node transaction information to other verifier nodes, and verifying the legitimacy of the computing nodes; all validator nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
when the computing nodes are legal, voting is carried out on each legal computing node corresponding to the jth verifier node, and node rewards of the jth verifier node in the first round of global training are determined after the voting is finished;
the received signed calculation node transaction information and corresponding voting results of each legal calculation node, the j-th verifier node information and node rewards of the j-th verifier node in the first round of global training are packaged and signed to obtain the signed verifier node transaction information of the j-th verifier node, and the signed verifier node transaction information of the j-th verifier node is sent to the associated witness node; one witness node is associated with at least one verifier node;
after receiving all corresponding signed verifier node transaction information, for a kth witness node:
Verifying the transaction information of each signed verifier node, broadcasting the verification-passed signed verifier node transaction information to other witness nodes, signing based on the received verification-passed signed verifier node transaction information, the block head, the time stamp and the witness number of the kth witness node, generating the kth new block in the global training of the first round, and broadcasting the kth new block to all computing nodes, all verifier nodes and other witness nodes; all witness nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
verifying the kth witness node based on the kth new block by all the computing nodes, all the verifier nodes and other witness nodes, adding the kth new block into a blockchain when the kth witness node is legal, and determining node rewards of the kth witness node in the global training of the first round;
determining an effective local model from the local models of all computing nodes in the first round of global training according to voting results in the new block;
determining parameters of the global model after the first round of global training based on the parameters of the effective local model after all the first round of global training;
Judging whether the training ending condition is met; the training ending condition is that the global model after the first round of global training converges or the verification accuracy of the global model after the first round of global training reaches the preset accuracy;
if yes, determining the global model after the first round of global training as the positioning model;
if not, determining the accumulated rewards after the first round of global training of the corresponding node based on the node rewards of the first round of global training of each node and the accumulated rewards of the first round of global training of each node, thereby determining the computing node, the verifier node and the witness node of the first round of global training, and performing the first round of global training and the first round of global training until the training ending condition is met.
Optionally, performing M rounds of local training on parameters of the initial local model of the ith round of global training by using the local data set of the ith computing node and device information of a plurality of training user devices to obtain parameters of the local model of the ith round of global training of the ith computing node, where the method specifically includes:
determining parameters of the local model of the mth round of local training based on the equipment information of all training user equipment and the parameters of the local model of the mth-1 round of local training;
Judging whether M is smaller than M;
if so, performing the m+1st round of local training until m=m, and determining the parameters of the local model of the M round of local training as the parameters of the local model after the first round of global training.
Optionally, verifying the validity of the computing node specifically includes:
storing the computing node transaction information into a transaction pool of a j-th verifier node according to the time sequence of receiving the computing node transaction information;
and when the transaction pool of the jth verifier node is full or new computing node transaction information is not received within a preset time, verifying the legitimacy of the corresponding computing node based on the received signature of the computing node corresponding to the computing node transaction information.
Optionally, voting is performed on each legal computing node corresponding to the jth verifier node, which specifically includes:
verifying the received parameters of the local model after the first round of global training of each legal computing node based on the local data set of the jth verifier node, and determining the verification precision of the parameters of the local model after the first round of global training of the corresponding computing node;
verifying parameters of the global model of the first round of global training based on the local data set of the jth verifier node, and determining verification accuracy of the parameters of the global model of the first round of global training;
And voting the local model after the first round of global training of the corresponding computing node based on the verification precision of the parameters of the local model after the first round of global training of each computing node and the verification precision of the parameters of the global model after the first-1 round of global training.
Optionally, voting is performed on the local model after the first round of global training of the corresponding computing node based on the verification precision of the parameters of the local model after the first round of global training of each computing node and the verification precision of the parameters of the global model after the first-1 round of global training, which specifically comprises:
calculating the difference value between the verification precision of the parameters of the local model after the first round of global training and the verification precision of the parameters of the global model after the first-1 round of global training of each computing node;
judging whether the difference value is larger than a preset threshold value or not;
if yes, then the ticket is accepted;
if not, the anti-vote is thrown.
A sensing and positioning system for user equipment, comprising:
the device information acquisition module is used for acquiring the device information of the target user device; the device information includes: the position of a base station and a direction angle formed by user equipment and the base station;
a positioning module, configured to determine a location of the target user equipment based on the device information of the target user equipment and a positioning model; the positioning model is determined by each node in a distributed network based on federal learning and blockchain; the node comprises: a compute node, a verifier node, and a witness node.
An electronic device, comprising:
one or more processors;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the methods as described above.
A storage medium having stored thereon a computer program which, when executed by a processor, implements a method as described above.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention discloses a sensing and positioning method, a sensing and positioning system, electronic equipment and a storage medium of user equipment, wherein equipment information of target user equipment is firstly acquired; the device information includes: the position of the base station and the direction angle formed by the user equipment and the base station; determining a location of the target user device based on the device information of the target user device and the positioning model; the positioning model is determined using nodes in a distributed network based on federal learning and blockchain. The invention is constructed based on federal learning and a blockchain framework, is used for replacing a traditional central server and coordinating distributed computing nodes to carry out local training and global aggregation, ensures data security, effectively solves the problems of privacy disclosure and data security risk, and ensures higher anti-attack capability while maintaining high positioning precision.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a sensing and positioning method of a user equipment according to embodiment 1 of the present invention;
FIG. 2 is a training flow diagram of a federal learning and blockchain based positioning module.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention aims to provide a sensing and positioning method, a sensing and positioning system, electronic equipment and a storage medium of user equipment, which aim to maintain high positioning accuracy and simultaneously have higher attack resistance.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
Example 1
Fig. 1 is a flowchart of a sensing and positioning method of a ue according to embodiment 1 of the present invention. As shown in fig. 1, the sensing and positioning method of the user equipment in this embodiment includes:
step 101: and acquiring the equipment information of the target user equipment.
Wherein the device information includes: the location of the base station and the direction angle formed by the user equipment and the base station.
Step 102: the location of the target user device is determined based on the device information of the target user device and the positioning model.
Wherein the positioning model is determined using nodes in a distributed network based on federal learning and blockchain; the node comprises: a compute node, a verifier node, and a witness node.
Specifically, a distributed network based on federal learning and blockchain includes a computing node, a verifier node, and a witness node. Wherein C is i （i∈{1,2,…,N c }) represents the firstEach computing node N c Representing the total number of computing nodes, node C i The local data set is contained +.>。V j （j∈{1,2,…,N v }) represents->A plurality of verifier nodes N v Representing the total number of verifier nodes. W (W) k （k∈{1,2,…,N w -N) represents the kth witness node w Representing the total number of witness nodes, each compute node being associated with one verifier node, each verifier node being associated with one witness node. Initially, the related devices join the blockchain network to become distributed nodes, and randomly allocate computing nodes, verifier nodes and witness nodes to generate an creation block (i.e. a first block in the blockchain), wherein the creation block contains information such as the size of a local data set of each node, a public key for signing, the number of local training rounds, parameters of an initial global model and the like. Let->C during wheel training i 、V j 、W k Random association, computing node, verifier node, and witnessThe node satisfies N c +N v +N w =N D ，N D Representing the total number of devices corresponding to the nodes contained in the distributed network.
As an alternative embodiment, the determining of the positioning model includes:
parameters of the global model are initialized.
Performing global training and local training on parameters of an initial global model by utilizing equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node to obtain a positioning model; the compute node, the verifier node, and the witness node are determined by partitioning all nodes in the distributed network based on the cumulative rewards.
As an alternative embodiment, as shown in FIG. 2 (Y represents yes and N represents no in FIG. 2), the first round of global training, i.e., 1, of the initial global model parameters is performed by using the device information and the true positions of the plurality of training user devices, the computing node, the verifier node, and the witness node, and includes:
step 1: for the ith compute node:
step 11: determining the parameters of the global model after the first round of global training as the parameters of the initial local model of the first round of global training of the ith computing node; all computing nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training; when l=1, the parameters of the global model of the first round of global training are the parameters of the initial global model, and the accumulated rewards after the first round of global training are the accumulated rewards after the global training obtained by initialization.
Step 12: local data set using ith compute nodeM rounds of local training are carried out on the parameters of the initial local model of the first round of global training by the equipment information of the plurality of training user equipment to obtain the parameters of the local model of the ith computing node after the first round of global training, and the node rewards of the ith computing node in the first round of global training are determined The method comprises the steps of carrying out a first treatment on the surface of the The local data set comprises the equipment information and the real positions of all training user equipment corresponding to the nodes.
As an alternative embodiment, the local data set of the ith computing node is utilizedAnd performing M rounds of local training on the parameters of the initial local model of the first round of global training by using the equipment information of the plurality of training user equipment to obtain the parameters of the local model of the first round of global training of the ith computing node, wherein the method specifically comprises the following steps:
the parameters of the local model of the mth round of local training are determined based on the device information of all training user devices and the parameters of the local model of the mth-1 round of local training.
And judging whether M is smaller than M.
If so, performing the m+1st round of local training until m=m, and determining the parameters of the local model of the M round of local training as the parameters of the local model after the first round of global training.
Specifically, the update formula of the parameters of the local model is:
。
wherein γ represents a learning rate; i represents an i-th computing node;representing parameters of a local model of an ith computing node after m-1 times of local training in a first round of global training; />Representing parameters of a local model of an ith computing node after m times of local training in a first round of global training; / >Is a gradient operator; />Representation localData set->And the loss function between the predicted position coordinates and the actual position coordinates of the training user equipment determined by the local model after m times of local training in the first round of global training by the ith computing node is specifically expressed as follows:
。
wherein, the liquid crystal display device comprises a liquid crystal display device,representing the real position coordinates of the q-th training user equipment,/->The predicted position coordinates of the qth training user equipment are expressed, and n1 is the total number of training user equipment.
Specifically, the ith compute node rewards the node in the first round of global trainingIs of the size ofK1 represents a data reward coefficient satisfying 0<k1<1，/>Representing local data set +.>Is of a size of (a) and (b).
Step 13: the method comprises the steps of performing package signature on parameters of a local model after the first round of global training of an ith computing node, the time used for the first round of global training of the ith computing node, the local data set size of the ith computing node and node rewards of the ith computing node in the first round of global training to obtain computing node transaction information after the ith signature, and sending the computing node transaction information after the ith signature to an associated verifier node; one verifier node is associated with at least one computing node.
Step 2: after receiving all corresponding signed computing node transaction information, for a j-th verifier node:
step 21: broadcasting all received signed computing node transaction information to other verifier nodes, and verifying the legitimacy of the computing nodes; all validator nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training. The other verifier nodes are the verifier nodes other than the jth verifier node among all the verifier nodes.
As an optional implementation manner, verifying the validity of the computing node specifically includes:
and storing the computing node transaction information into a transaction pool of the j-th verifier node according to the time sequence of receiving the computing node transaction information.
And when the transaction pool of the jth verifier node is full or new computing node transaction information is not received within a preset time, verifying the legitimacy of the corresponding computing node based on the received signature of the computing node corresponding to the computing node transaction information.
And when the computing nodes are legal, voting is carried out on each legal computing node corresponding to the jth verifier node, and node rewards of the jth verifier node in the first round of global training are determined after the voting is finished.
Specifically, the jth verifier node rewards the node in the first round of global trainingIs Size (T) which represents the number of node transaction information in the transaction pool of the j-th verifier node.
As an optional implementation manner, voting is performed on each legal computing node corresponding to the jth verifier node, which specifically includes:
local data set based on jth verifier nodeVerifying the parameters of the received local model after the first round of global training of each legal computing node, and determining the verification accuracy of the parameters of the local model after the first round of global training of the corresponding computing node +.>。
Verifying parameters of the global model of the first round of global training based on the local data set of the jth verifier node, and determining verification accuracy of the parameters of the global model of the first round of global training。
And voting the local model after the first round of global training of the corresponding computing node based on the verification precision of the parameters of the local model after the first round of global training of each computing node and the verification precision of the parameters of the global model after the first-1 round of global training.
In practice, voting is to perform parameter verification in the local model, where parameter verification is to check out the existence of malicious nodes or malicious attacks, and each verifier node is assumed to have equal voting rights.
As an optional implementation manner, voting is performed on the local model after the first round of global training of the corresponding computing node based on the verification precision of the parameters of the local model after the first round of global training of each computing node and the verification precision of the parameters of the global model after the first-1 round of global training, and specifically the method comprises the following steps:
and calculating the difference value between the verification precision of the parameters of the local model after the first round of global training and the verification precision of the parameters of the global model after the first-1 round of global training of each computing node.
And judging whether the difference value is larger than a preset threshold value.
If so, a vote is awarded indicating that the parameters of the local model after the first round of global training are valid on the validation set (i.e., the local data set of the jth validator node).
If not, the anti-vote is thrown.
Step 22: the received signed calculation node transaction information and corresponding voting results of each legal calculation node, the j-th verifier node information and node rewards of the j-th verifier node in the first round of global training are packaged and signed to obtain the signed verifier node transaction information of the j-th verifier node, and the signed verifier node transaction information of the j-th verifier node is sent to the associated witness node; one witness node is associated with at least one verifier node.
Step 3: after receiving all corresponding signed verifier node transaction information, for a kth witness node:
step 31: verifying the transaction information of each signed verifier node, broadcasting the verification-passed signed verifier node transaction information to other witness nodes, signing based on the received verification-passed signed verifier node transaction information, the block head, the time stamp and the witness number of the kth witness node, generating the kth new block in the global training of the first round, and broadcasting the kth new block to all computing nodes, all verifier nodes and other witness nodes; all witness nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training.
Specifically, the method for generating the new block based on the agent rights and interests proving (Delegated Proof of Stake, DPOS) consensus mechanism comprises the following specific steps: witness nodes need to construct a preliminary block before they can be ready to generate a new block. The prepared block includes not only the verifier node transaction information taken from the transaction pool of the witness node, but also information such as block size, block header, witness number, time stamp, and the like. Wherein the block size represents the total volume of all data of the new block, which is equal to the number of transaction information of the verifier node; the block header is the thumbnail information of the whole new block and is used for identifying the whole new block; the verifier node transaction information is used for solving the defects of calculation resource waste and long consensus time existing in the traditional block generation based on a workload consensus mechanism.
Step 4:
step 41: all compute nodes, all verifier nodes, and other witness nodes verify the kth witness node based on the kth new chunk, add the kth new chunk to the blockchain when the kth witness node is legitimate, and determine node rewards for the kth witness node for global training on the first round.
Step 42: and determining an effective local model from the local models of all computing nodes in the first round of global training according to voting results in the new block.
Specifically, if len (for) > len (against), the node model parameter is considered valid, and the corresponding local model is determined to be a valid local model; if len (for). Ltoreq.len (agains), then the local model is considered to have potential malicious attacks on the first round, which should be culled. len (for) is the number of approved tickets and len (agains) is the number of disagreeed tickets.
Step 43: and determining the parameters of the global model after the first round of global training based on the parameters of the effective local model after all the first round of global training.
Specifically, step 43 includes:
firstly, weighting the effective local model after the first round of global training by using a principal component analysis method, and constructing a parameter matrix, namely compressing the parameters of each effective local model (the parameters of the local model are a matrix with a plurality of rows and a plurality of columns) into one column of the parameter matrix, then transposing the whole parameter matrix, and supposing that the parameter matrix is deformed to be To eliminate the influence of dimension, parameters are normalized first, namely:
。
wherein, the liquid crystal display device comprises a liquid crystal display device,representing the values of the matrix h, row a, column b, normalized by the parameters,/for the elements of the matrix h, row a, column b, and column b>Representing the mean value of the b-th column element in the matrix h, s b Represents standard deviation of element b, h ab Representing the values of the elements of row a and column b of matrix h.
Secondly, calculating the correlation between different nodes to obtain a correlation coefficient matrix, wherein the element r of the row a and the column b in the correlation coefficient matrix ab The expression of (2) is:
。
where n1 represents the total number of rows in matrix h, c represents the row number, h ca Representing the values of the elements of row c and column a of matrix h, h cb Representing the values of the elements of row c and column b of matrix h,representing the mean of the elements of column a in matrix h.
Thirdly, singular value decomposition is carried out on the correlation coefficient matrix to obtain score vectors of parameters of each effective local model, and normalization is carried out to obtain global aggregation weight of the d effective local model:
。
e is the total number of the effective local models, E and d are the serial numbers of the effective local models, SC d Score vector for the d-th valid local model, SC e Is the score vector of the e-th valid local model.
Finally, the global model generated by the first round of aggregation (i.e., the parameters of the global model after the first round of global training) can be expressed as:
。
Wherein, the liquid crystal display device comprises a liquid crystal display device,representing the d effective book of the first round of global training after M times of local trainingModel parameters->Parameters representing the global model after the first round of global training.
Step 44: judging whether the training ending condition is met; the training ending condition is that the global model after the first round of global training converges or the verification precision of the global model after the first round of global training reaches the preset precision.
Step 45: if so, determining the global model after the first round of global training as a positioning model.
Step 46: if not, determining the accumulated rewards after the first round of global training of the corresponding node based on the node rewards of the first round of global training of each node and the accumulated rewards of the first round of global training of each node, thereby determining the computing node, the verifier node and the witness node of the first round of global training, and performing the first round of global training and the first round of global training until the training ending condition is met.
Example 2
The sensing and positioning system of the user equipment in this embodiment includes:
the device information acquisition module is used for acquiring the device information of the target user device; the device information includes: the location of the base station and the direction angle formed by the user equipment and the base station.
The positioning module is used for determining the position of the target user equipment based on the equipment information of the target user equipment and the positioning model; the positioning model is determined by each node in the distributed network based on federal learning and blockchain; the node comprises: a compute node, a verifier node, and a witness node.
Example 3
An electronic device, comprising:
one or more processors.
A storage device having one or more programs stored thereon.
The one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method as in embodiment 1.
Example 4
A storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the method as in embodiment 1.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different manner from other embodiments, so that identical and similar parts of each embodiment are mutually referred to. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.

Claims (8)

1. A method of sensing and locating a user device, the method comprising:
acquiring equipment information of target user equipment; the device information includes: the position of a base station and a direction angle formed by user equipment and the base station;
determining a location of the target user device based on the device information and a positioning model of the target user device; the positioning model is determined by each node in a distributed network based on federal learning and blockchain; the node comprises: a computing node, a verifier node, and a witness node;
the determining process of the positioning model comprises the following steps:
initializing parameters of a global model;
performing global training and local training on parameters of an initial global model by utilizing equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node to obtain a positioning model; the computing node, the verifier node, and the witness node are determined by partitioning all nodes in the distributed network based on the cumulative rewards;
performing global training on parameters of an initial global model and a first round of global training in local training by using equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node, wherein l is more than or equal to 1, and the method comprises the following steps:
For the ith compute node:
determining the parameters of the global model after the first round of global training as the parameters of the initial local model of the first round of global training of the ith computing node; all computing nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training; when l=1, the parameters of the global model of the first round of global training are the parameters of the initial global model, and the accumulated rewards after the first round of global training are the accumulated rewards after the global training obtained by initialization;
performing M rounds of local training on parameters of an initial local model of the first round of global training by using a local data set of the ith computing node and equipment information of a plurality of training user equipment to obtain parameters of the local model of the ith computing node after the first round of global training, and determining node rewards of the ith computing node in the first round of global training; the local data set comprises equipment information and real positions of all training user equipment corresponding to the nodes;
the method comprises the steps of performing package signature on parameters of a local model after the first round of global training of an ith computing node, the time used for the first round of global training of the ith computing node, the local data set size of the ith computing node and node rewards of the ith computing node in the first round of global training to obtain computing node transaction information after the ith signature, and sending the computing node transaction information after the ith signature to an associated verifier node; one verifier node is associated with at least one computing node;
After receiving all corresponding signed computing node transaction information, for a j-th verifier node:
broadcasting all received signed computing node transaction information to other verifier nodes, and verifying the legitimacy of the computing nodes; all validator nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
when the computing nodes are legal, voting is carried out on each legal computing node corresponding to the jth verifier node, and node rewards of the jth verifier node in the first round of global training are determined after the voting is finished;
the received signed calculation node transaction information and corresponding voting results of each legal calculation node, the j-th verifier node information and node rewards of the j-th verifier node in the first round of global training are packaged and signed to obtain the signed verifier node transaction information of the j-th verifier node, and the signed verifier node transaction information of the j-th verifier node is sent to the associated witness node; one witness node is associated with at least one verifier node;
after receiving all corresponding signed verifier node transaction information, for a kth witness node:
Verifying the transaction information of each signed verifier node, broadcasting the verification-passed signed verifier node transaction information to other witness nodes, signing based on the received verification-passed signed verifier node transaction information, the block head, the time stamp and the witness number of the kth witness node, generating the kth new block in the global training of the first round, and broadcasting the kth new block to all computing nodes, all verifier nodes and other witness nodes; all witness nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
verifying the kth witness node based on the kth new block by all the computing nodes, all the verifier nodes and other witness nodes, adding the kth new block into a blockchain when the kth witness node is legal, and determining node rewards of the kth witness node in the global training of the first round;
determining an effective local model from the local models of all computing nodes in the first round of global training according to voting results in the new block;
determining parameters of the global model after the first round of global training based on the parameters of the effective local model after all the first round of global training;
Judging whether the training ending condition is met; the training ending condition is that the global model after the first round of global training converges or the verification precision of the global model after the first round of global training reaches the preset precision;
if yes, determining the global model after the first round of global training as a positioning model;
if not, determining the accumulated rewards after the first round of global training of the corresponding node based on the node rewards of the first round of global training of each node and the accumulated rewards of the first round of global training of each node, thereby determining the computing node, the verifier node and the witness node of the first round of global training, and performing the first round of global training and the first round of global training until the training ending condition is met.
2. The method for sensing and positioning user equipment according to claim 1, wherein performing M rounds of local training on parameters of an initial local model of the first round of global training by using a local data set of the ith computing node and device information of a plurality of training user equipments, to obtain parameters of the local model of the first round of global training of the ith computing node, specifically includes:
determining parameters of the local model of the mth round of local training based on the equipment information of all training user equipment and the parameters of the local model of the mth-1 round of local training;
Judging whether M is smaller than M;
if so, performing the m+1st round of local training until m=m, and determining the parameters of the local model of the M round of local training as the parameters of the local model after the first round of global training.
3. The method for sensing and positioning a user device according to claim 1, wherein verifying the validity of the computing node comprises:
storing the computing node transaction information into a transaction pool of a j-th verifier node according to the time sequence of receiving the computing node transaction information;
and when the transaction pool of the jth verifier node is full or new computing node transaction information is not received within a preset time, verifying the legitimacy of the corresponding computing node based on the received signature of the computing node corresponding to the computing node transaction information.
4. The method for sensing and positioning a user device according to claim 1, wherein voting is performed on each legal computing node corresponding to a jth verifier node, specifically comprising:
verifying the received parameters of the local model after the first round of global training of each legal computing node based on the local data set of the jth verifier node, and determining the verification precision of the parameters of the local model after the first round of global training of the corresponding computing node;
Verifying parameters of the global model of the first round of global training based on the local data set of the jth verifier node, and determining verification accuracy of the parameters of the global model of the first round of global training;
and voting the local model after the first round of global training of the corresponding computing node based on the verification precision of the parameters of the local model after the first round of global training of each computing node and the verification precision of the parameters of the global model after the first-1 round of global training.
5. The method for sensing and positioning user equipment according to claim 4, wherein voting is performed on the local model after the first round of global training of the corresponding computing node based on the verification accuracy of the parameters of the local model after the first round of global training of the computing node and the verification accuracy of the parameters of the global model after the first-1 round of global training, specifically comprising:
calculating the difference value between the verification precision of the parameters of the local model after the first round of global training and the verification precision of the parameters of the global model after the first-1 round of global training of each computing node;
judging whether the difference value is larger than a preset threshold value or not;
if yes, then the ticket is accepted;
if not, the anti-vote is thrown.
6. A sensing and positioning system for user equipment, the system comprising:
The device information acquisition module is used for acquiring the device information of the target user device; the device information includes: the position of a base station and a direction angle formed by user equipment and the base station;
a positioning module, configured to determine a location of the target user equipment based on the device information of the target user equipment and a positioning model; the positioning model is determined by each node in a distributed network based on federal learning and blockchain; the node comprises: a computing node, a verifier node, and a witness node;
the determining process of the positioning model comprises the following steps:
initializing parameters of a global model;
performing global training and local training on parameters of an initial global model by utilizing equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node to obtain a positioning model; the computing node, the verifier node, and the witness node are determined by partitioning all nodes in the distributed network based on the cumulative rewards;
performing global training on parameters of an initial global model and a first round of global training in local training by using equipment information and real positions of a plurality of training user equipment, a computing node, a verifier node and a witness node, wherein l is more than or equal to 1, and the method comprises the following steps:
For the ith compute node:
determining the parameters of the global model after the first round of global training as the parameters of the initial local model of the first round of global training of the ith computing node; all computing nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training; when l=1, the parameters of the global model of the first round of global training are the parameters of the initial global model, and the accumulated rewards after the first round of global training are the accumulated rewards after the global training obtained by initialization;
performing M rounds of local training on parameters of an initial local model of the first round of global training by using a local data set of the ith computing node and equipment information of a plurality of training user equipment to obtain parameters of the local model of the ith computing node after the first round of global training, and determining node rewards of the ith computing node in the first round of global training; the local data set comprises equipment information and real positions of all training user equipment corresponding to the nodes;
the method comprises the steps of performing package signature on parameters of a local model after the first round of global training of an ith computing node, the time used for the first round of global training of the ith computing node, the local data set size of the ith computing node and node rewards of the ith computing node in the first round of global training to obtain computing node transaction information after the ith signature, and sending the computing node transaction information after the ith signature to an associated verifier node; one verifier node is associated with at least one computing node;
After receiving all corresponding signed computing node transaction information, for a j-th verifier node:
broadcasting all received signed computing node transaction information to other verifier nodes, and verifying the legitimacy of the computing nodes; all validator nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
when the computing nodes are legal, voting is carried out on each legal computing node corresponding to the jth verifier node, and node rewards of the jth verifier node in the first round of global training are determined after the voting is finished;
the received signed calculation node transaction information and corresponding voting results of each legal calculation node, the j-th verifier node information and node rewards of the j-th verifier node in the first round of global training are packaged and signed to obtain the signed verifier node transaction information of the j-th verifier node, and the signed verifier node transaction information of the j-th verifier node is sent to the associated witness node; one witness node is associated with at least one verifier node;
after receiving all corresponding signed verifier node transaction information, for a kth witness node:
Verifying the transaction information of each signed verifier node, broadcasting the verification-passed signed verifier node transaction information to other witness nodes, signing based on the received verification-passed signed verifier node transaction information, the block head, the time stamp and the witness number of the kth witness node, generating the kth new block in the global training of the first round, and broadcasting the kth new block to all computing nodes, all verifier nodes and other witness nodes; all witness nodes in the first round of global training are determined based on the cumulative rewards after the first-1 round of global training;
verifying the kth witness node based on the kth new block by all the computing nodes, all the verifier nodes and other witness nodes, adding the kth new block into a blockchain when the kth witness node is legal, and determining node rewards of the kth witness node in the global training of the first round;
determining an effective local model from the local models of all computing nodes in the first round of global training according to voting results in the new block;
determining parameters of the global model after the first round of global training based on the parameters of the effective local model after all the first round of global training;
Judging whether the training ending condition is met; the training ending condition is that the global model after the first round of global training converges or the verification precision of the global model after the first round of global training reaches the preset precision;
if yes, determining the global model after the first round of global training as a positioning model;
if not, determining the accumulated rewards after the first round of global training of the corresponding node based on the node rewards of the first round of global training of each node and the accumulated rewards of the first round of global training of each node, thereby determining the computing node, the verifier node and the witness node of the first round of global training, and performing the first round of global training and the first round of global training until the training ending condition is met.
7. An electronic device, comprising:
one or more processors;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-5.
8. A storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the method of any of claims 1 to 5.
CN202310993102.7A 2023-08-09 2023-08-09 User equipment sensing and positioning method, system, electronic equipment and storage medium Active CN116709509B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310993102.7A CN116709509B (en) 2023-08-09 2023-08-09 User equipment sensing and positioning method, system, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310993102.7A CN116709509B (en) 2023-08-09 2023-08-09 User equipment sensing and positioning method, system, electronic equipment and storage medium

Publications (2)

Publication Number Publication Date
CN116709509A CN116709509A (en) 2023-09-05
CN116709509B true CN116709509B (en) 2023-10-17

Family

ID=87831603

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310993102.7A Active CN116709509B (en) 2023-08-09 2023-08-09 User equipment sensing and positioning method, system, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN116709509B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111879305A (en) * 2020-06-16 2020-11-03 华中科技大学 Multi-mode perception positioning model and system for high-risk production environment
CN112422650A (en) * 2020-11-05 2021-02-26 徐康庭 Building positioning method, building positioning device, building positioning equipment and computer readable storage medium
CN113794675A (en) * 2021-07-14 2021-12-14 中国人民解放军战略支援部队信息工程大学 Distributed Internet of things intrusion detection method and system based on block chain and federal learning
CN114731615A (en) * 2019-11-21 2022-07-08 高通股份有限公司 UE-based positioning
CN115460537A (en) * 2021-06-08 2022-12-09 中移(上海)信息通信科技有限公司 Base station positioning method, device and network equipment

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114731615A (en) * 2019-11-21 2022-07-08 高通股份有限公司 UE-based positioning
CN111879305A (en) * 2020-06-16 2020-11-03 华中科技大学 Multi-mode perception positioning model and system for high-risk production environment
CN112422650A (en) * 2020-11-05 2021-02-26 徐康庭 Building positioning method, building positioning device, building positioning equipment and computer readable storage medium
CN115460537A (en) * 2021-06-08 2022-12-09 中移(上海)信息通信科技有限公司 Base station positioning method, device and network equipment
CN113794675A (en) * 2021-07-14 2021-12-14 中国人民解放军战略支援部队信息工程大学 Distributed Internet of things intrusion detection method and system based on block chain and federal learning

Also Published As

Publication number Publication date
CN116709509A (en) 2023-09-05

Similar Documents

Publication Publication Date Title
US9122853B2 (en) Location determination for user authentication
CN108109011A (en) A kind of anti-cheat method of advertisement and computing device
US10015171B1 (en) Authentication using metadata from posts made to social networking websites
CN104903902A (en) Systems and methods for authenticating user based on biometric model associated with user
CN112416986B (en) User portrait realizing method and system based on hierarchical personalized federal learning
CN106716471A (en) Systems and methods for identifying mobile devices
CN105245345A (en) High reliability perception data collection algorithm based on mobile perception user anonymity reputation in crowd sensing
CN111860865B (en) Model construction and analysis method, device, electronic equipment and medium
US20150309962A1 (en) Method and apparatus for modeling a population to predict individual behavior using location data from social network messages
CN104508657B (en) Mediation computing device and the method associated for producing semantic label
CN107563798A (en) Prize-winning data processing method and device
CN104852916A (en) Social engineering-based webpage verification code recognition method and system
CN104281795A (en) Mouse action based password fault tolerance method
CN101984620B (en) Codebook generating method and convert communication system
CN112182645A (en) Quantifiable privacy protection method, equipment and medium for destination prediction
CN109121133B (en) Location privacy protection method and device
CN113886817A (en) Host intrusion detection method and device, electronic equipment and storage medium
CN116709509B (en) User equipment sensing and positioning method, system, electronic equipment and storage medium
CN113064972A (en) Intelligent question and answer method, device, equipment and storage medium
KR102470639B1 (en) Improving Data Integrity with Trusted Proof-of-Code Tokens
CN106993027B (en) Remote data storage location verification method
CN113010865B (en) Big data base component safety management method and system of intelligent education platform
CN106992862A (en) Electric endorsement method and device based on the Big Dipper and digital certificate
CN113435502B (en) Site flow determination method, device, equipment and storage medium
Liu et al. An in-situ authentication with privacy preservation scheme for accident response in Internet of Vehicles

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant