CN115604314A - Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning - Google Patents

Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning Download PDF

Info

Publication number
CN115604314A
CN115604314A CN202211212353.9A CN202211212353A CN115604314A CN 115604314 A CN115604314 A CN 115604314A CN 202211212353 A CN202211212353 A CN 202211212353A CN 115604314 A CN115604314 A CN 115604314A
Authority
CN
China
Prior art keywords
model
diagnosis
ciphertext
parameters
medical diagnosis
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.)
Pending
Application number
CN202211212353.9A
Other languages
Chinese (zh)
Inventor
孔庆磊
陈波
庞艳华
温书娜
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.)
Shenzhen Graduate School Harbin Institute of Technology
Original Assignee
Shenzhen Graduate School Harbin Institute of Technology
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 Shenzhen Graduate School Harbin Institute of Technology filed Critical Shenzhen Graduate School Harbin Institute of Technology
Priority to CN202211212353.9A priority Critical patent/CN115604314A/en
Publication of CN115604314A publication Critical patent/CN115604314A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
    • G06F21/45—Structures or tools for the administration of authentication
    • G06F21/46—Structures or tools for the administration of authentication by designing passwords or checking the strength of passwords
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/14—Relay systems
    • H04B7/15—Active relay systems
    • H04B7/185—Space-based or airborne stations; Stations for satellite systems
    • H04B7/1851—Systems using a satellite or space-based relay
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/04—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks
    • H04L63/0428—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload
    • H04L63/0435—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload wherein the sending and receiving network entities apply symmetric encryption, i.e. same key used for encryption and decryption
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/008—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols involving homomorphic encryption
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/32—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
    • H04L9/3247—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving digital signatures

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Medical Informatics (AREA)
  • Health & Medical Sciences (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Astronomy & Astrophysics (AREA)
  • Mathematical Physics (AREA)
  • General Health & Medical Sciences (AREA)
  • Radio Relay Systems (AREA)

Abstract

The invention provides a privacy protection diagnosis method of a low-orbit satellite diagnosis model based on joint learning, which comprises the steps of collecting medical diagnosis data of a target area; constructing a medical diagnosis model through a disease model server of a ground station, and transmitting the medical diagnosis model to a ground data collection point of the target area through a satellite network; training the medical diagnosis model according to the medical diagnosis data to obtain a model updating parameter; and carrying out encryption protection on the model updating parameters, and transmitting the encrypted model updating parameters back to the ground station through the satellite network. By the method, a data privacy protection mechanism suitable for high-dynamic, low-power and complex topological environment of the satellite Internet of things is designed, and the problems of data security collection and privacy model aggregation are solved.

Description

Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning
Technical Field
The invention belongs to the field of privacy protection.
Background
In recent years, a great deal of research work has been carried out on security and privacy protection problems in malaria patient networks, expert scholars from the communication industry, computer science major, and network security fields in the aspect of protecting the internet of things security of malaria patients and the privacy of data related to the internet of things security. Data values gathered by terrestrial access points are correlated with geographic location and time information, and location information includes personal sensitive information of the user. In the existing research, the following key technologies are mainly covered aiming at the data privacy protection problem in the internet of things architecture:
anonymization techniques. The anonymous privacy protection technology requires that the distribution of the sensitive attribute values in all equivalence classes is the same as the probability distribution of all data in the data set, namely, when the sensitive attribute value of the target user does not change, an attacker cannot obtain privacy information from the data set.
Differential privacy techniques. The differential privacy technology aims to maximize the accuracy of data query results and simultaneously reduce the opportunity of identifying records of the data query results when the data set is queried. Namely, random noise is added to ensure that the data query is publicly visible, and the query result of the information is not changed by individuals.
Data encryption technology. Among many security policies, encryption techniques can ensure the security and privacy of data related between user devices or processes in a malicious environment. The existing encryption data protection strategy mainly focuses on a data transmission stage, a data storage stage and a data processing stage.
However, in the field of communication technology of satellite internet of things, the following four important problems are mainly faced: firstly, the propagation path of the satellite is long, so that the propagation delay of the satellite network is too long; secondly, the problem of difficult terminal cooperation and centralized scheduling of large-scale low-orbit satellites is solved; thirdly, the problem of network congestion caused by common transmission of massive Internet of things equipment is solved; and fourthly, the low power consumption of the satellite Internet of things equipment is realized. At present, a great deal of research work is carried out in satellite access research both at home and abroad in the aspects of satellite time slot and satellite non-time slot random access. In the aspect of solving the access of massive user terminals, the problem that the reliability and timeliness of information transmission are ensured needs to be solved. Starting the satellite relay random access system later and not generating a corresponding satellite application scene; meanwhile, the problem of high system complexity exists. In the field of architecture design of the low-earth-orbit satellite internet of things system, the communication access field and the network architecture design field of the satellite internet of things are mainly concentrated, and the safety requirements of the low-earth-orbit satellite internet of things on data safety and privacy protection are not considered. Based on the above analysis, how to design a data privacy protection mechanism suitable for high dynamic, low power and complex topological environment of the satellite internet of things is an important problem to be solved urgently.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, the invention aims to provide a privacy protection diagnosis method of a low-orbit satellite diagnosis model based on joint learning, which is used for solving the problems of data security collection and privacy model aggregation in a high-dynamic heterogeneous environment of a low-orbit satellite internet of things.
In order to achieve the above object, a first embodiment of the present invention provides a privacy-preserving diagnosis method based on a joint learning low-earth orbit satellite diagnosis model, including:
acquiring medical diagnosis data of a target area;
constructing a medical diagnosis model through a disease model server of a ground station, and transmitting the medical diagnosis model to a ground data collection point of the target area through a satellite network;
training the medical diagnosis model according to the medical diagnosis data to obtain a model updating parameter;
and carrying out encryption protection on the model updating parameters, and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
In addition, the privacy protection diagnosis method based on the jointly-learned low-earth orbit satellite diagnosis model according to the above embodiment of the present invention may further have the following additional technical features:
further, in an embodiment of the present invention, before constructing the medical diagnosis model by the disease model server of the ground station, the method further includes:
the parameters of the medical diagnostic model are protected by a symmetric homomorphic cryptographic cryptosystem and a verifiable secret sharing scheme.
Further, in an embodiment of the present invention, the cryptographically protecting the model update parameters and transmitting the encrypted model update parameters back to the ground station via the satellite network includes:
generating a secret shared secret c by the ground data collection point b The secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k through the surface data collection point b For the secret sharing value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model from the ground data collection points
Figure BDA0003871683780000021
Encrypting the parameters of the medical diagnosis model by the ground data collection point by using the public key in the symmetric homomorphic encryption cryptosystem
Figure BDA0003871683780000031
Information to be calculated by using the ground data collection point satellite node
Figure BDA0003871683780000032
Is sent back to the ground station。
Further, in an embodiment of the present invention, after transmitting the encrypted model update parameters back to the ground station through the satellite network, the method further includes:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (short-range encryption) cryptosystem through the disease model server, and obtaining an aggregation result;
generating a new message by the ground station:
Figure BDA0003871683780000033
and transmitting the message to the disease model server;
decrypting the received one-time ciphertext H (x | | | task) by the disease model server j ||t)·f(id b )；
Then the correctness of the decrypted disposable ciphertext is verified through the disease model server, if the correctness is right, the disposable ciphertext H (x | | task) is calculated through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure BDA0003871683780000034
Finally pass through
Figure BDA0003871683780000035
To decrypt the ciphertext to obtain
Figure BDA0003871683780000036
I.e. the updated model parameters.
In order to achieve the above object, a second embodiment of the present invention provides a privacy-preserving diagnosis apparatus based on a joint learning low-earth orbit satellite diagnosis model, including the following modules:
the acquisition module is used for acquiring medical diagnosis data of a target area;
the construction module is used for constructing a medical diagnosis model through a disease model server of a ground station and transmitting the medical diagnosis model to a ground data collection point of the target area through a satellite network;
the training module is used for training the medical diagnosis model according to the medical diagnosis data to obtain a model updating parameter;
and the transmission module is used for carrying out encryption protection on the model updating parameters and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
Further, in an embodiment of the present invention, the building module is further configured to:
the parameters of the medical diagnostic model are protected by a symmetric homomorphic cryptographic cryptosystem and a verifiable secret sharing scheme.
Further, in an embodiment of the present invention, the transmission module is further configured to:
generating a secret shared secret c by the ground data collection point b The secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k through the surface data collection point b For the secret sharing value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model from the ground data collection points
Figure BDA0003871683780000041
Encrypting the parameters of the medical diagnosis model by the ground data collection point by using the public key in the symmetric homomorphic encryption cryptosystem
Figure BDA0003871683780000042
Information to be calculated by using the terrestrial data collection point satellite node
Figure BDA0003871683780000043
And sent back to the ground station.
Further, in an embodiment of the present invention, the transmission module further includes an aggregation unit, configured to:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (short-range encryption) cryptosystem through the disease model server, and obtaining an aggregation result;
generating a new message by the ground station:
Figure BDA0003871683780000044
and transmitting the message to the disease model server;
decrypting the received one-time ciphertext H (x | | | task) by the disease model server j ||t)·f(id b )；
Then the correctness of the decrypted disposable ciphertext is verified through the disease model server, if the correctness is right, the disposable ciphertext H (x | | task) is calculated through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure BDA0003871683780000045
Finally pass through
Figure BDA0003871683780000046
To decrypt the ciphertext to obtain
Figure BDA0003871683780000047
I.e. the updated model parameters.
To achieve the above object, a third aspect of the present invention provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the method for privacy-preserving diagnosis based on a jointly learned low-orbit satellite diagnosis model as described above.
To achieve the above object, a fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a privacy-preserving diagnosis method based on a joint learning low-orbit satellite diagnosis model as described above.
The privacy protection diagnosis method of the low-earth orbit satellite diagnosis model based on the joint learning is based on the symmetric homomorphic encryption technology, achieves security model aggregation and user authentication in a model aggregation process of a distributed training process of the low-earth orbit satellite Internet of things, and meanwhile achieves high computing efficiency by using an SHE (short-range encryption algorithm) password system. The scheme of the invention evaluates the privacy protection diagnosis model in an actual malaria database, and is compared with the traditional model protection scheme using a paillier password system, and the scheme is far superior to the traditional scheme in the aspect of computational complexity.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a schematic diagram of a privacy protection diagnosis method of a low earth orbit satellite diagnosis model based on joint learning according to an embodiment of the present invention.
Fig. 2 is a schematic diagram illustrating an update process of a low earth orbit satellite diagnostic model according to an embodiment of the present invention.
Fig. 3 is a schematic diagram illustrating comparison of computational complexity between the present solution and the conventional solution according to an embodiment of the present invention.
Fig. 4 is a schematic diagram of a privacy protection diagnosis apparatus based on a joint learning low-earth orbit satellite diagnosis model according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative and intended to explain the present invention and should not be construed as limiting the present invention.
The privacy protection diagnosis method based on the joint learning low-orbit satellite diagnosis model according to the embodiment of the invention is described below with reference to the accompanying drawings.
Traditional diagnostic methods rely primarily on laboratory and expertise and are not adaptable to some infectious diseases that occur in remote tropical regions like malaria. In these areas, the ground network infrastructure is not fully covered, and there is also a lack of adequate medical doctor resources and a lack of adequate case data in a single area.
Fig. 1 is a schematic flowchart of a privacy protection diagnosis method based on a joint learning low-earth orbit satellite diagnosis model according to an embodiment of the present invention.
As shown in fig. 1, the privacy-preserving diagnosis method based on the joint learning low-earth orbit satellite diagnosis model includes the following steps:
s101: acquiring medical diagnosis data of a target area;
s102: constructing a medical diagnosis model through a disease model server of a ground station, and transmitting the medical diagnosis model to a ground data collection point of a target area through a satellite network;
s103: training a medical diagnosis model according to the medical diagnosis data to obtain model updating parameters;
s104: and carrying out encryption protection on the model updating parameters, and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
The invention provides a privacy data model aggregation protocol based on federal learning, which is based on a scene that a low earth orbit satellite network distributed joint diagnosis data model is constructed. Due to the fact that the low-earth orbit satellite moves at a high speed and the energy supply and computing capacity of the participating nodes are limited, the privacy aggregation of the model is achieved by the aid of a symmetric homomorphic encryption method. Meanwhile, in order to ensure that the content of a single model is not leaked in the model aggregation result, the minimum number of users participating in model training is further specified by using a secret sharing protocol.
In particular, the homomorphic encryption technique is an encryption method that allows a specific algebraic operation to be performed on encrypted content by effectively operating on the associated ciphertext.
1) Addition homomorphism attribute: the product of the two ciphertexts will be decrypted as the sum of their respective plaintexts:
D(E(m 1 )*E(m 2 )mod n 2 )＝m 1 +m 2 mod n，
2) Number-multiplied homomorphism attribute:
Figure BDA0003871683780000061
the homomorphic encryption method consists of a quadruplet: h = { KeyGen, enc, dec, eval },
KeyGen: representing a key generation function, symmetric homomorphic encryption, generating only one key sk = KeyGen (g);
enc: representing an encryption function, inputting a public key sk and a plaintext m in an encryption process in symmetric homomorphic encryption, and generating a ciphertext c = Encsk (m);
and Dec: representing a decryption function, in symmetric homomorphic encryption, the privacy key sk and the ciphertext c are used as inputs to generate the associated plaintext m = Decsk (c);
eval: representing the evaluation function.
Federal learning is a distributed machine learning technology, is used for solving the data island problem, and can help a plurality of organizations to use data and model machine learning under the condition of meeting the requirements of privacy protection, data safety and government regulations. The System mainly comprises three major parts, namely a data source (Database), a federal learning System (System/model) and a User (User). In the federal learning system, a plurality of data sources firstly process the obtained data, jointly establish a federal learning system model, and then transmit an output result to a user.
The invention designs an artificial intelligent diagnosis model (the input is the symptoms and physical conditions of a patient, and the output is a diagnosis result device) suitable for malaria infectious diseases. The initial model based on the ground gateway station is M 0 Want to update the model parameters M with the malaria database 0 . The system model mainly comprises three parts: 1) medical model servers and ground gateway stations, 2) low earth orbit satellite constellations and 3) ground access points and medical data collection points. Patient medical record entry due to single data collection pointMultiple data collection points are required to jointly implement model training. In each round of model updating process, a ground gateway station firstly calls a plurality of low-orbit satellites to participate in the transfer process of an initial model, transfers initial parameters to a target ground access point for model training, and returns a training result to the ground station through an inter-satellite link after the model training is finished.
During each model update iteration t, the disease model server sends model parameters M t To the Ground Station (GS), which then recognizes the ground access point A j Of a group of satellites S t He first transmits the model parameters M t For satellite Sat a The satellite will transmit M t Target ground access point AP b (b∈A j ) Then, AP b The model parameters M will be updated by locally collected disease data t Obtaining model updates
Figure BDA0003871683780000071
In addition, AP b Will be transmitted through satellite
Figure BDA0003871683780000072
To a ground station. After the ground station receives the signals, the pair
Figure BDA0003871683780000073
And integrating and sending the integration to a medical model server.
The system main body of the scheme of the invention is divided into three parts: 1) Initializing a system; 2) Encrypting the model; 3) And aggregating the privacy protection model.
Further, in an embodiment of the present invention, before constructing the medical diagnosis model by the disease model server of the ground station, the method further includes:
the parameters of the medical diagnosis model are protected by a symmetric homomorphic encryption cryptosystem and a verifiable secret sharing scheme.
Wherein a Symmetric Homomorphic Encryption (SHE) cryptosystem and a verifiable secret sharing scheme are combined to realize model parameter protection. In the system initialization phase, we assume thatOne Trust Authority (TA) may direct the entire system. In this cryptosystem initialization, TA first selects a large prime number p, and a generator α ∈ GF (p). Then, TA selects a sequence (a) n-1 ，a n-2 ，...，a 1 ，a 0 ) Belongs to GF (p) to construct n-1 degree polynomial f on GF (p) 0 (x)＝a n-1 ·x n -1 +a n-2 ·x n-2 +...+a 1 ·x+a 0 The secret value is a 0 . Meanwhile, TA selects a one-way hash function H: {0,1} * →GF(p)。
In the process of initializing the model server, the model server selects a secret number beta E [1,p-1]Calculating the value alpha β And passes this value to TA. Upon receiving this value, the TA calculates the value of the model server
Figure BDA0003871683780000074
And a is s Back into the model server. TA also selects a secret value x, which is stored in local memory.
In the stage of the registration of the ground station GS, a random number z is first selected s ∈[1，p-1]As a secret key and publishes a public key
Figure BDA0003871683780000075
At the same time, secret parameters k of the SHE cryptosystem are obtained 0 ，k 1 ，k 2 The GS selects two large prime numbers q s ，p s And | p s |＝|q s |＝k 0 -bits, generating a key
Figure BDA0003871683780000076
Figure BDA0003871683780000077
Is a random number
Figure BDA0003871683780000081
GS selects two random numbers
Figure BDA0003871683780000082
And calculates the key pk in the SHE cryptographic system s ＝(E(0) 1 ，E(0) 2 ，N s ＝p s ·q s ). Meanwhile, the GS selects a one-way hash function
Figure BDA0003871683780000083
In addition, the GS also publishes its system public parameter pp s ＝(Z s ，k 0 ，k 1 ，k 2 ，pk s ，H s )。
In the registration stage of ground data collection points, collecting point identity id i E.g. GF (p), the collection point first selects a random number z i ∈[1，p-1]And issues a public key
Figure BDA0003871683780000084
At the same time, the TA generates an identity-based secret share f (id) i )＝a n-1 ·(id i ) n-1 +a n-2 ·(id i ) n-2 ...+a 1 ·id i +a 0 And transmits it to the access point. The model server will also distribute this secret value to him.
Further, in an embodiment of the present invention, the encrypting the model update parameters and transmitting the encrypted model update parameters back to the ground station through the satellite network includes:
generation of secret shared secret c by ground data collection points b Secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k through a ground data collection point b For secret sharing value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model via ground data collection points
Figure BDA0003871683780000085
Point utilization through surface data collectionPublic key encryption medical diagnosis model parameter in symmetric homomorphic encryption cryptosystem
Figure BDA0003871683780000086
Information computed by using terrestrial data collection point satellite nodes
Figure BDA0003871683780000087
And sent back to the ground station.
Specifically, in the t-th iteration, the diagnostic model is tasked j Updating, diagnosing model server transmission model parameter M t The target access point setting A j To the ground station GS, as satellite Sat a (a∈S t And satisfy | S t | ≧ n) loitering above GS, GS will check Sat a And entrusted task j To Sat a . When Sat a Run to AP b (b∈A j ) At access point, sat a Will transmit the model parameter M t ||pp s ||A s To AP b 。AP b Model parameters may be derived using collected malaria diagnostic data
Figure BDA0003871683780000088
AP b Each independent model update is protected according to the following steps
Figure BDA0003871683780000089
First step, AP b Using the public key Z of the ground station s And its own private key z b Protecting the one-time ciphertext to generate a secret shared ciphertext:
Figure BDA0003871683780000091
second step, AP b Selecting a random value k b ∈[1，p-1]For secret sharing H (x | | | task) j ||t)·f(id b ) And (3) signing:
Figure BDA0003871683780000092
third step, AP b Generating a protector for each model size l
Figure BDA0003871683780000093
At the same time, AP b Each model parameter will be decrypted using the public key in the SHE cryptosystem
Figure BDA0003871683780000094
Finally AP b Transmitting information via multiple retransmission transmissions between satellites
Figure BDA0003871683780000095
To GS.
Further, in an embodiment of the present invention, after transmitting the encrypted model update parameters back to the ground station through the satellite network, the method further includes:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (short-range encryption) cryptosystem through a disease model server, and obtaining an aggregation result;
generating a new message by the ground station:
Figure BDA0003871683780000096
and transmitting the message to a disease model server;
decrypting the received one-time ciphertext H (x | | | task) by the disease model server j ||t)·f(id b )；
Then verifying the correctness of the decrypted disposable ciphertext through a disease model server, and if the decrypted disposable ciphertext is correct, calculating the disposable ciphertext H (x | | | task) through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure BDA0003871683780000097
Finally pass through
Figure BDA0003871683780000098
To decrypt the ciphertext to obtain
Figure BDA0003871683780000099
I.e. the updated model parameters.
Specifically, when all model parameters are received
Figure BDA00038716837800000910
Then, the ground receiving station GS integrates the ciphertext of each data dimension l, that is:
Figure BDA00038716837800000911
the GS will then pass the private key of the SHE cryptosystem
Figure BDA00038716837800000912
To decrypt and send the model integration result
Figure BDA00038716837800000913
At the same time, GS formulates and delivers
Figure BDA00038716837800000914
To the medical model server. The medical model server first retrieves each access point AP b The one-time ciphertext H (x | | | task) j ||t)·f(id b )，b∈A j Namely:
Figure BDA0003871683780000101
he then verifies the correctness of this recovered one-time ciphertext as follows:
Figure BDA0003871683780000102
in addition, the disease model server recovers the one-time ciphertext H (x | task) by the Lagrange interpolation polynomial method j ||t)·a 0 Calculating
Figure BDA0003871683780000103
Finally obtaining a polymerization model
Figure BDA0003871683780000104
The above is a complete privacy protection diagnosis method flow of the low-orbit satellite diagnosis model based on the joint learning, and fig. 2 is a schematic diagram of an updating process of the low-orbit satellite diagnosis model, aiming at an environment perception scene where an access point of a malaria patient is located, the problems of data security collection, privacy model aggregation and the like are solved at one time in a data security and privacy protection framework of the low-orbit satellite internet of things.
Compared with the prior art, the privacy protection diagnosis method of the low-orbit satellite diagnosis model based on the joint learning, provided by the embodiment of the invention, has the advantages that on one hand, an efficient real-time safety data collection scheme is summarized from the heterogeneous characteristics of a low-orbit satellite network, and on the premise of ensuring the privacy of collected data, the legality of the data collection process and the authenticity of the data collection are ensured; on the other hand, a safety model training and user participation authentication mechanism suitable for the high dynamic network topology is summarized and designed, and on the premise of reducing the calculation complexity, the privacy protection and authentication of the training model parameters are realized.
The scheme of the invention evaluates the privacy protection diagnosis model in an actual malaria database, and is compared with the traditional model protection scheme using a paillier password system, and the scheme is far superior to the traditional scheme in the aspect of computational complexity. As shown in particular in figure 3.
In order to implement the above embodiments, the present invention further provides a privacy protection diagnosis apparatus based on a joint learning low-earth orbit satellite diagnosis model.
Fig. 4 is a schematic structural diagram of a privacy protection diagnosis apparatus based on a joint learning low-earth orbit satellite diagnosis model according to an embodiment of the present invention.
As shown in fig. 4, the privacy-preserving diagnosis apparatus based on the joint learning low-earth orbit satellite diagnosis model includes: an acquisition module 100, a construction module 200, a training module 300, a transmission module 400, wherein,
the acquisition module is used for acquiring medical diagnosis data of a target area;
the construction module is used for constructing a medical diagnosis model through a disease model server of the ground station and transmitting the medical diagnosis model to a ground data collection point of a target area through a satellite network;
the training module is used for training the medical diagnosis model according to the medical diagnosis data to obtain model updating parameters;
and the transmission module is used for carrying out encryption protection on the model updating parameters and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
Further, in an embodiment of the invention, the building module is further configured to:
the parameters of the medical diagnosis model are protected by a symmetric homomorphic encryption cryptosystem and a verifiable secret sharing scheme.
Further, in an embodiment of the present invention, the transmission module is further configured to:
generation of secret shared secret c by ground data collection points b Secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k through a ground data collection point b For secret sharing value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model via ground data collection points
Figure BDA0003871683780000111
Encrypting parameters of a medical diagnosis model by using a public key in the symmetric homomorphic encryption cryptosystem through a ground data collection point
Figure BDA0003871683780000112
Information to be computed by using terrestrial data collection point satellite nodes
Figure BDA0003871683780000113
And sent back to the ground station.
Further, in an embodiment of the present invention, the transmission module further includes an aggregation unit, configured to:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (short-range encryption) cryptosystem through a disease model server, and obtaining an aggregation result;
generating a new message by the ground station:
Figure BDA0003871683780000114
and transmitting the message to a disease model server;
decrypting the received one-time ciphertext H (x | | task) through the disease model server j ||t)·f(id b )；
Then verifying the correctness of the decrypted disposable ciphertext through a disease model server, and if the decrypted disposable ciphertext is correct, calculating the disposable ciphertext H (x | | | task) through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure BDA0003871683780000115
Finally pass through
Figure BDA0003871683780000116
To decrypt the ciphertext to obtain
Figure BDA0003871683780000117
I.e. the updated model parameters.
To achieve the above object, a third aspect of the present invention provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the privacy-preserving diagnosis method based on the jointly-learned low-orbit satellite diagnosis model.
To achieve the above object, a fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the privacy-preserving diagnosis method based on a jointly-learned low-earth orbit satellite diagnosis model as described above.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless explicitly specified otherwise.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims (10)

1. A privacy protection diagnosis method of a low-orbit satellite diagnosis model based on joint learning is characterized by comprising the following steps:
acquiring medical diagnosis data of a target area;
constructing a medical diagnosis model through a disease model server of a ground station, and transmitting the medical diagnosis model to a ground data collection point of the target area through a satellite network;
training the medical diagnosis model according to the medical diagnosis data to obtain a model updating parameter;
and carrying out encryption protection on the model updating parameters, and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
2. The method of claim 1, further comprising, prior to building the medical diagnostic model by the disease model server of the ground station:
the parameters of the medical diagnostic model are protected by a symmetric homomorphic cryptographic cryptosystem and a verifiable secret sharing scheme.
3. The method of claim 1 or 2, wherein said cryptographically protecting said model update parameters and transmitting the encrypted model update parameters back to said ground station via said satellite network comprises:
generating a secret shared secret c through the ground data collection points b The secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k by the surface data collection point b For the secret shared value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model from the ground data collection points
Figure FDA0003871683770000011
Encrypting the parameters of the medical diagnosis model by the ground data collection point by using the public key in the symmetric homomorphic encryption cryptosystem
Figure FDA0003871683770000012
Information to be calculated by using the ground data collection point satellite node
Figure FDA0003871683770000013
And sent back to the ground station.
4. The method of claim 3, further comprising, after transmitting the encrypted model update parameters back to the ground station via the satellite network:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (short-range encryption) cryptosystem through the disease model server, and obtaining an aggregation result;
generating a new message by the ground station:
Figure FDA0003871683770000021
and transmitting the message to the disease model server;
decrypting the received one-time ciphertext H (x | | | task) by the disease model server j ||t)·f(id b )；
Then the correctness of the decrypted disposable ciphertext is verified through the disease model server, if the correctness is right, the disposable ciphertext H (x | | task) is calculated through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure FDA0003871683770000022
Finally pass through
Figure FDA0003871683770000023
To decrypt the ciphertext to obtain
Figure FDA0003871683770000024
I.e. the updated model parameters.
5. A privacy protection diagnosis device based on a low-orbit satellite diagnosis model of joint learning is characterized by comprising the following modules:
the acquisition module is used for acquiring medical diagnosis data of a target area;
the construction module is used for constructing a medical diagnosis model through a disease model server of a ground station and transmitting the medical diagnosis model to a ground data collection point of the target area through a satellite network;
the training module is used for training the medical diagnosis model according to the medical diagnosis data to obtain a model updating parameter;
and the transmission module is used for carrying out encryption protection on the model updating parameters and transmitting the encrypted model updating parameters back to the ground station through the satellite network.
6. The apparatus of claim 5, wherein the build module is further configured to:
parameters of the medical diagnostic model are protected by a symmetric homomorphic cryptographic cryptosystem and a verifiable secret sharing scheme.
7. The apparatus of claim 5 or 6, wherein the transmission module is further configured to:
generating a secret shared secret c by the ground data collection point b The secret shared secret c b For protecting secret shared values f (id) b )；
Selecting a random value k by the surface data collection point b For the secret sharing value f (id) b ) Signing to obtain R b ，s b ；
Generating protection parameters for a model from the ground data collection points
Figure FDA0003871683770000031
Encrypting the medical diagnosis by the ground data collection point using the public key in the symmetric homomorphic encryption cryptosystemParameters of the model
Figure FDA0003871683770000032
Information to be calculated by using the terrestrial data collection point satellite node
Figure FDA0003871683770000033
And sent back to the ground station.
8. The apparatus of claim 7, wherein the transmission module further comprises an aggregation unit configured to:
aggregating, by the disease model server, the ciphertext for each data dimension;
decrypting the ciphertext by using a private key of an SHE (secure Shell) password system through the disease model server to obtain an aggregation result;
generating a new message by the ground station:
Figure FDA0003871683770000034
and transmitting the message to the disease model server;
decrypting the received one-time ciphertext H (x | | | task) by the disease model server j ||t)·f(id b )；
Then the correctness of the decrypted disposable ciphertext is verified through the disease model server, if the correctness is right, the disposable ciphertext H (x | | task) is calculated through a Lagrange interpolation polynomial method j ||t)·a 0 And calculating
Figure FDA0003871683770000035
Finally pass through
Figure FDA0003871683770000036
To decrypt the ciphertext to obtain
Figure FDA0003871683770000037
I.e. updated model parameters。
9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the method of privacy preserving diagnosis based on jointly learned low-orbit satellite diagnostic model according to any one of claims 1 to 4 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the method for privacy-preserving diagnosis based on a jointly learned low-earth orbit satellite diagnosis model according to any one of claims 1 to 4.
CN202211212353.9A 2022-09-29 2022-09-29 Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning Pending CN115604314A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211212353.9A CN115604314A (en) 2022-09-29 2022-09-29 Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211212353.9A CN115604314A (en) 2022-09-29 2022-09-29 Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning

Publications (1)

Publication Number Publication Date
CN115604314A true CN115604314A (en) 2023-01-13

Family

ID=84845054

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211212353.9A Pending CN115604314A (en) 2022-09-29 2022-09-29 Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning

Country Status (1)

Country Link
CN (1) CN115604314A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116709303A (en) * 2023-08-03 2023-09-05 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) Satellite edge calculation method and device for remote monitoring
CN117411683A (en) * 2023-10-17 2024-01-16 中国人民解放军国防科技大学 Method and device for identifying low orbit satellite network attack based on distributed federal learning

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116709303A (en) * 2023-08-03 2023-09-05 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) Satellite edge calculation method and device for remote monitoring
CN116709303B (en) * 2023-08-03 2024-01-16 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) Satellite edge calculation method and device for remote monitoring
CN117411683A (en) * 2023-10-17 2024-01-16 中国人民解放军国防科技大学 Method and device for identifying low orbit satellite network attack based on distributed federal learning

Similar Documents

Publication Publication Date Title
Zhang et al. Reliable and privacy-preserving truth discovery for mobile crowdsensing systems
CN111372243B (en) Security distributed aggregation and access system and method based on fog alliance chain
Khan et al. An efficient and provably secure certificateless key-encapsulated signcryption scheme for flying ad-hoc network
Zhang et al. Privacy-preserving profile matching for proximity-based mobile social networking
CN115604314A (en) Privacy protection diagnosis method of low-orbit satellite diagnosis model based on joint learning
CN112583575A (en) Homomorphic encryption-based federated learning privacy protection method in Internet of vehicles
Touati et al. Collaborative kp-abe for cloud-based internet of things applications
Qiu et al. An efficient multi-message and multi-receiver signcryption scheme for heterogeneous smart mobile IoT
CN110730064B (en) Data fusion method based on privacy protection in crowd sensing network
Dang et al. Resource-efficient secure data sharing for information centric e-health system using fog computing
CN113098686A (en) Group key management method for low-earth-orbit satellite network
Shen et al. Secure emergent data protection scheme for a space-terrestrial integrated network
CN115906172A (en) Method for protecting federated learning data
Zhao et al. Attribute-based encryption with non-monotonic access structures supporting fine-grained attribute revocation in m-healthcare.
Boussada et al. A lightweight privacy-preserving solution for IoT: The case of E-health
Han et al. A secure trust-based key distribution with self-healing for internet of things
Li et al. Identity-based multi-recipient public key encryption scheme and its application in IoT
Xue DSAS: a secure data sharing and authorized searchable framework for e-healthcare system
Kanchan et al. An efficient and privacy-preserving federated learning scheme for flying ad hoc networks
CN116599653A (en) Dynamic security management method, system and storage medium for satellite communication network
Wang et al. Cloud-assisted key distribution in batch for secure real-time mobile services
Subramani et al. Blockchain-enabled secure data collection scheme for fog-based WBAN
Zhou et al. Secure and efficient fine-grained multiple file sharing in cloud-assisted crowd sensing networks
Drosatos et al. An efficient privacy-preserving solution for finding the nearest doctor
Chatterjee et al. An efficient fine-grained access control scheme for hierarchical wireless sensor networks

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