CN112015556A - Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism - Google Patents

Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism Download PDF

Info

Publication number
CN112015556A
CN112015556A CN202010897900.6A CN202010897900A CN112015556A CN 112015556 A CN112015556 A CN 112015556A CN 202010897900 A CN202010897900 A CN 202010897900A CN 112015556 A CN112015556 A CN 112015556A
Authority
CN
China
Prior art keywords
sensing
data
cluster
node
clustering
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.)
Granted
Application number
CN202010897900.6A
Other languages
Chinese (zh)
Other versions
CN112015556B (en
Inventor
岑健
刘溪
宋海鹰
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.)
Guangdong Polytechnic Normal University
Original Assignee
Guangdong Polytechnic Normal University
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 Guangdong Polytechnic Normal University filed Critical Guangdong Polytechnic Normal University
Priority to CN202010897900.6A priority Critical patent/CN112015556B/en
Publication of CN112015556A publication Critical patent/CN112015556A/en
Application granted granted Critical
Publication of CN112015556B publication Critical patent/CN112015556B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a mobile crowd sensing data balancing method based on a block chain equity certification mechanism in the field of mobile crowd sensing networks, which comprises the following steps: s1, in the time slot period, the sensing platform carries out data clustering based on the pyramid tree algorithm on the received sensing data and calculates the quantity of the sensing data contained in each cluster; s2, judging whether the quantity of the perception data contained in the added clusters is less than a quantity threshold value or not, and outputting a quantity judgment parameter value; s3, calculating the utility function value of each sensing node according to the judging parameter value by each sensing node; and S4, determining the activity state of the sensing node according to the utility function. In the whole mobile crowd sensing task execution flow of the method, the sensing platform selects as few participants as possible to complete the sensing task, the quality requirement of spatial coverage of sensing nodes in a designated sensing area is met, and data balance is realized in spatial dimension.

Description

Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism
Technical Field
The invention relates to the field of mobile swarm intelligence perception networks, in particular to a mobile swarm intelligence perception data balancing method based on a block chain equity certification mechanism.
Background
In the mobile swarm intelligence perception network, the distribution of perception nodes (participants) in a perception area is non-uniform and is characterized by power-law distribution. This property causes 1) part of the sensing area to be oversampled. 2) The problem of data loss exists in a part of sensing areas, and the whole sensing quality is further influenced. 3) The randomness of the arrival time of the sensing nodes enables a considerable part of sensing areas to exceed the number k of the nodes required by the coverage model, but long sensing time is required, and the convergence speed of sensing data is influenced.
Therefore, for problems 1) -3), for large-scale crowd sensing applications, when the number of nodes in a sensing area exceeds the requirement, the sensing state of a redundant node needs to be controlled, and more sensing nodes need to be added to a "sparse" sensing area. In general, it is difficult for mobile crowd sensing data to reach balance in space and time dimensions simultaneously, and the problem of mobile crowd sensing data imbalance needs to be solved.
Disclosure of Invention
Based on the analysis, the problem of imbalance of the mobile crowd sensing data is actually the problem of the fewest participant groups which need to meet the task coverage quality requirement in space and the problem of how to realize the rapid convergence of the mobile crowd sensing data in time. In order to solve the defects, a block chain rights and interests certification mechanism is introduced, and a sensing node has a certain amount of rights and interests when sending sensing data, so that the problems needing to be solved in time and space are solved on the basis, and the mobile group intelligence sensing data balancing method based on the block chain rights and interests certification mechanism is provided.
In order to achieve the above purpose, the invention provides the following technical scheme:
a mobile swarm intelligence perception data balancing method based on a block chain rights and interests certification mechanism comprises the following steps:
s1, in the time slot period, the sensing platform carries out data clustering based on the pyramid tree algorithm on the received sensing data and calculates the quantity of the sensing data contained in each cluster;
s2, when adding new sensing data to the cluster, the sensing platform judges whether the quantity of the sensing data contained in the added cluster is less than a quantity threshold value, and outputs a quantity judgment parameter value;
s3, each sensing node calculates the utility function value of each sensing node according to the quantity evaluation parameter value returned by each cluster of the sensing platform;
s4, if the utility function value of the sensing node is 0 during two continuous time slots, the sensing node does not participate in the sensing activity; and if the quantity of the sensing data contained in the clusters is less than the quantity threshold value in the period of two continuous time slots, an active sensing program of the sensing platform sends the clustering result to all the sensing nodes in the next time slot, and the clustering result is the result of the clustering with the quantity of the sensing data less than the quantity threshold value.
Further, the method comprises the following steps:
and S5, if the quantity of the sensing data contained in the cluster is less than the quantity threshold value after the next time slot period is finished, the active sensing program sends the cluster result to all newly participating sensing nodes.
Further, the method comprises the following steps:
and S6, after the sensing node collects any cluster data according to the data acquisition requirement, similarity comparison is carried out on the cluster data and the cluster result sent by the sensing platform, if the cluster result meets the similarity clustering requirement, the sensing data of the sensing node is marked as data to be sent preferentially, and the data is uploaded preferentially in the nearest sending window.
As a preferred scheme of the invention, the calculation formula of the quantity judgment parameter value is
Figure BDA0002659052090000031
Among them, CountR (Hotspots)j) Is the amount of perceptual data contained in the cluster, k is the number threshold, NR (Hotspots)j,uiAnd t) is a quantity evaluation parameter value.
As a preferred scheme of the invention, the calculation formula of the utility function value of the sensing node is
Figure BDA0002659052090000032
Wherein UT (u)i,t) Is the value of the utility function of the sensing node, NR (Hotspots)j,uiT) is the value of the quantity judgment parameter, C (Hotspots)j,uiAnd t) represents the sensing period t cohesive results HotspotsjWhether or not to be perceived node uiAnd (6) voting.
As a preferred scheme of the present invention, the specific steps of performing data clustering based on the pyramid tree algorithm on the received perception data include:
dividing received perception data into data of f characteristics;
setting the pyramid tree as an f +2 layer;
with the increase of the received sensing data, the pyramid tree grows, and the number of nodes at the nth layer of the pyramid tree is not less than the number of nodes at the n-1 layer;
each leaf node of the (f +2) th layer is a data set, and the data set is a clustering result.
Based on the same conception, the invention also provides a mobile group intelligence perception data equalization system based on a block chain equity certification mechanism, which is characterized by comprising at least one processor and a memory which is in communication connection with the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the above.
Compared with the prior art, the invention has the beneficial effects that:
in the whole mobile crowd sensing task execution flow of the method, a sensing platform selects as few participants as possible to complete a sensing task, the quality requirement of spatial coverage of sensing nodes in a designated sensing area is met, and data balance is realized in spatial dimension; meanwhile, an active sensing program is applied to a sensing area with sparse sensing node distribution, the convergence speed of the whole sensing data is accelerated, and data balance is realized on a time dimension.
Description of the drawings:
FIG. 1 is a flow chart of a mobile crowd sensing data equalization method based on a block chain equity certification mechanism according to the present invention;
fig. 2 is a flowchart of a mobile crowd sensing data balancing method based on a blockchain equity certification mechanism according to an embodiment 1 of the present invention;
fig. 3 is a schematic diagram of a data aggregation algorithm of a pyramid tree in embodiment 1 of the present invention.
Detailed Description
The present invention will be described in further detail with reference to test examples and specific embodiments. It should be understood that the scope of the above-described subject matter is not limited to the following examples, and any techniques implemented based on the disclosure of the present invention are within the scope of the present invention.
Example 1
The problem of mobile crowd sensing data equalization is actually the problem of computing the minimum participant population that meets the task coverage quality requirement, i.e. within the sensing period T ═ l × T, U ═ U { (U) for a given candidate set of participants1,u2,...,unH participants are selected from the set to form a participant set Ui＝{ui1,ui2,...,uihThe sensing area and a plurality of reference points in the sensing area form a target reference point set R ═ R1,r2,...,rm′R, each target location R in the requirement R' in the mobile crowd sensingjThere are at least k participants. Namely, it is
Minimize:|Ui| (1)
Figure BDA0002659052090000051
Figure BDA0002659052090000052
Wherein, formula 1 represents that the employed participant set is minimum, and formula 2 represents that each target location R in the target reference point set RjThere are at least k participants, equation 3 represents that the participant set is a subset of the participant candidate set and the perceptual data of all participants in the period T constitutes a target reference point set R'. The problem is an extension of the classical set coverage problem, which has proven to be an NP-hard problem, and thus the mobile crowd-sourcing perceptual data equalization problem is also at least an NP-hard problem.
Meanwhile, in the time dimension, the uncertainty of the data uploading opportunity of the sensing nodes also increases the difficulty of fast convergence and equalization of the mobile crowd sensing data. In fact, at present, a commonly used NP-hard problem solving algorithm based on a greedy algorithm, a genetic algorithm and the like is complex in calculation on one hand, and on the other hand, it is difficult to achieve the balance of the mobile crowd sensing data in space and time dimensions simultaneously, so that it is difficult to directly apply to the mobile crowd sensing data balance problem.
A block chain right-interest proving mechanism is introduced for this purpose, namely a sensing node needs to have a certain amount of right before sending sensing data, and a sensing node u is defined in the designiThe utility function in the sensing period t is the right and is specifically defined as follows:
Figure BDA0002659052090000053
among them, NR (Hotspots)j，uiAnd t) represents a sensing node uiWhether the jth sensing region data sent in the sensing period time slot t is the first k data received in the region data cluster (Hotspots) or not is defined as C (Hotspots)j,uiAnd t) represents the sensing period t cohesive results HotspotsjWhether or not to be perceived node uiVoting, then:
Figure BDA0002659052090000054
as a specific embodiment, a flow chart of a mobile crowd sensing data balancing method based on a blockchain equity certification mechanism is shown in fig. 1, and a flow chart of the mobile crowd sensing data balancing method based on the blockchain equity certification mechanism is shown in fig. 2. The method comprises the following specific steps:
1) in the x (x ═ 1, 2.., l) perceptionIn a period time slot t, assuming that n mobile sensing nodes upload sensing data recorded in the memory to a sensing platform, wherein the node uiThe uploaded data contains Q pieces of recorded perception data information (F)i s1,Fi s2,...,Fi sQ)。
2) The perception platform carries out data clustering based on the pyramid tree algorithm on all perception data information received in the time slot from the n mobile perception nodes, and calculates the number CountR (Hotspots) of perception data contained in each clusterj)。
For a data aggregation task with f features, the pyramid tree is defined as a (f +2) layer tree, which grows continuously with the addition of data, and all leaf nodes only appear at the bottom layer of the tree, i.e. the number of nodes at the nth layer is not less than that at the (n-1) th layer. The root node is only used as the root node of the whole tree to organize each layer; each leaf node of the lowest layer, i.e., the (f +2) th layer, represents a data set; and the middle 1 st to (f +1) th layers are non-leaf nodes, each layer represents a feature to which the task belongs, and all data are clustered according to the features to obtain the lowest-layer clustering result.
3) The perception platform confirms the number of perception data CountR (Hotspots) in each cluster when adding perception data to the clusterj) And whether the coverage requirement is met or not is judged, and the NR function value (whether the node is the TOP-k node or not) of the sensing node which uploads the sensing data at the moment in the Hotspots is calculated.
In particular when moving the perceiving node uiWhen certain sensing data information in the data uploaded in the x (x ═ 1, 2.. multidot.l) th sensing period time slot t is determined as jth hotspot region data through clustering calculation, the jth hotspot region data is obtained at the moment
Figure BDA0002659052090000071
4) After the end of the x (x ═ 1, 2.., l) th sensing period time slot tAnd each sensing node clusters the returned NR (Hotspots) according to each hotspot of the sensing platformj) Calculating own utility function UT (u) of function value and coverage of Hotspots by the function value and the perception time slot ti)。
5) After two consecutive sensing time slots xt and (x +1) t, the sensing node uiUtility function UT (u)i) If 0, the sensing node ui enters an inactive state and no longer actively participates in the sensing activity.
If passing through two consecutive sensing time slots xt and (x +1) t, a certain hotspot cluster always fails to reach the coverage requirement, namely CountR (Hotspots)j)<And k, activating an active sensing program by the sensing platform, sending the current hotspot clustering results to all sensing nodes (including inactive nodes) by the active sensing program in a sensing time slot (x +2) t, comparing the similarity of the current hotspot clustering results with the clustering results sent by the sensing platform after the sensing nodes collect any hotspot data according to the data acquisition requirement, marking the current hotspot clustering results as the data to be sent preferentially if the current hotspot clustering results meet the similarity clustering requirement, and uploading the data preferentially in the nearest sending window. If the hotspot clustering in (x +2) t in the sensing period still cannot meet the coverage requirement, the active sensing program sends the current clustering result of the Hotspots to all newly participating sensing nodes from the sensing time slot (x +3) t, after the sensing nodes collect any hotspot data according to the data acquisition requirement, similarity comparison is carried out on the clustering result and the clustering result sent by the sensing platform, if the similarity clustering requirement is met, the clustering result is marked as the priority for sending the data, and the data is uploaded preferentially in the nearest sending window.
The sensing platform selects as few participants as possible to complete the sensing task in the whole mobile crowd sensing task execution process, and the quality requirement of sensing node space coverage of a designated sensing area is met; meanwhile, an active sensing program is applied to a sensing area with sparse sensing node distribution, the convergence speed of the whole sensing data is accelerated, and data balance is realized on a time dimension.

Claims (7)

1. A mobile swarm intelligence perception data equalization method based on a block chain equity certification mechanism is characterized by comprising the following steps:
s1, in the time slot period, the sensing platform carries out data clustering based on the pyramid tree algorithm on the received sensing data and calculates the quantity of the sensing data contained in each cluster;
s2, when adding new sensing data to the cluster, the sensing platform judges whether the quantity of the sensing data contained in the added cluster is less than a quantity threshold value, and outputs a quantity judgment parameter value;
s3, each perception node calculates the utility function value of each perception node according to the quantity evaluation parameter value returned by each cluster of the perception platform;
s4, if the utility function value of the sensing node is 0 during two continuous time slots, the sensing node does not participate in the sensing activity; and if the quantity of the sensing data contained in the clusters is less than the quantity threshold value in the period of two continuous time slots, an active sensing program of the sensing platform sends the clustering result to all sensing nodes in the period of the next time slot, wherein the clustering result is the result of clustering with the quantity of the sensing data less than the quantity threshold value.
2. The method according to claim 1, wherein the step of equalizing the mobile group-wisdom-aware data based on the blockchain equity certification mechanism further comprises:
s5, if the amount of sensing data included in the cluster is less than the threshold value after the next time slot period ends, the active sensing program sends the cluster result to all newly participating sensing nodes.
3. The method according to claim 2, wherein the step of equalizing the mobile group-wisdom-aware data based on the blockchain equity certification mechanism further comprises:
s6, after the sensing node collects any cluster data according to the data collection requirement, similarity comparison is carried out on the cluster data and the cluster result sent by the sensing platform, if the cluster result meets the similarity clustering requirement, the sensing data of the sensing node is marked as priority sending data, and the priority uploading is carried out in the nearest sending window.
4. The method of claim 1, wherein the quantity evaluation parameter value is calculated by the following formula
Figure FDA0002659052080000021
Among them, CountR (Hotspots)j) Is the amount of perceptual data contained in the cluster, k is the number threshold, NR (Hotspots)j,uiAnd t) is a quantity evaluation parameter value.
5. The method of claim 1, wherein the utility function value of the sensing node is calculated by the following formula
Figure FDA0002659052080000022
Wherein UT (u)i,t) Is the value of the utility function of the sensing node, NR (Hotspots)j,uiT) is the value of the quantity judgment parameter, C (Hotspots)j,uiAnd t) represents the sensing period t cohesive results HotspotsjWhether or not to be perceived node uiAnd (6) voting.
6. The method for mobile crowd-sourcing perceptual data equalization based on blockchain equity certification mechanism as claimed in any one of claims 1 to 5 wherein said step of performing pyramid-tree algorithm based data clustering on the received perceptual data comprises:
dividing the received perception data into data of f characteristics;
setting the pyramid tree as an f +2 layer;
with the increase of the received sensing data, the pyramid tree grows, and the number of nodes at the nth layer of the pyramid tree is not less than the number of nodes at the n-1 layer;
each leaf node of the (f +2) th layer is a data set, and the data set is a clustering result.
7. A mobile swarm intelligence aware data equalization system based on a blockchain equity certification mechanism is characterized by comprising at least one processor and a memory which is in communication connection with the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.
CN202010897900.6A 2020-08-31 2020-08-31 Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism Active CN112015556B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010897900.6A CN112015556B (en) 2020-08-31 2020-08-31 Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010897900.6A CN112015556B (en) 2020-08-31 2020-08-31 Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism

Publications (2)

Publication Number Publication Date
CN112015556A true CN112015556A (en) 2020-12-01
CN112015556B CN112015556B (en) 2021-05-04

Family

ID=73502781

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010897900.6A Active CN112015556B (en) 2020-08-31 2020-08-31 Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism

Country Status (1)

Country Link
CN (1) CN112015556B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105022835A (en) * 2015-08-14 2015-11-04 武汉大学 Public safety recognition method and system for crowd sensing big data
CN111209825A (en) * 2019-12-31 2020-05-29 武汉中海庭数据技术有限公司 Method and device for dynamic target 3D detection
CN111222916A (en) * 2020-01-03 2020-06-02 腾讯科技（深圳）有限公司 Method and device for determining delivery area, model training method and storage medium
CN111292122A (en) * 2020-01-16 2020-06-16 支付宝(杭州)信息技术有限公司 Method and apparatus for facilitating user to perform target behavior for target object
US20200220881A1 (en) * 2019-01-09 2020-07-09 Ebay Inc. Weighted verification of entity data blocks on a blockchain
CN111414381A (en) * 2020-03-04 2020-07-14 腾讯科技（深圳）有限公司 Data processing method and device, electronic equipment and storage medium
US10754936B1 (en) * 2013-03-13 2020-08-25 United Services Automobile Associate (USAA) Behavioral profiling method and system to authenticate a user

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10754936B1 (en) * 2013-03-13 2020-08-25 United Services Automobile Associate (USAA) Behavioral profiling method and system to authenticate a user
CN105022835A (en) * 2015-08-14 2015-11-04 武汉大学 Public safety recognition method and system for crowd sensing big data
US20200220881A1 (en) * 2019-01-09 2020-07-09 Ebay Inc. Weighted verification of entity data blocks on a blockchain
CN111209825A (en) * 2019-12-31 2020-05-29 武汉中海庭数据技术有限公司 Method and device for dynamic target 3D detection
CN111222916A (en) * 2020-01-03 2020-06-02 腾讯科技（深圳）有限公司 Method and device for determining delivery area, model training method and storage medium
CN111292122A (en) * 2020-01-16 2020-06-16 支付宝(杭州)信息技术有限公司 Method and apparatus for facilitating user to perform target behavior for target object
CN111414381A (en) * 2020-03-04 2020-07-14 腾讯科技（深圳）有限公司 Data processing method and device, electronic equipment and storage medium

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JUNFENG XIE等: "A Survey of Blockchain Technology Applied to Smart Cities:Research Issues and Challenges", 《IEEE COMMUNICATIONS SURVEYS & TUTORIALS》 *
吴梦宇等: "基于工作量证明和权益证明改进的区块链共识机制", 《计算机应用》 *

Also Published As

Publication number Publication date
CN112015556B (en) 2021-05-04

Similar Documents

Publication Publication Date Title
CN110445866B (en) Task migration and cooperative load balancing method in mobile edge computing environment
CN112020103A (en) Content cache deployment method in mobile edge cloud
CN112148484B (en) Coupling degree-based micro-service online distribution method and system
CN110851429A (en) Edge computing credible cooperative service method based on influence self-adaptive aggregation
CN112637883A (en) Federal learning method with robustness to wireless environment change in power Internet of things
CN115358487A (en) Federal learning aggregation optimization system and method for power data sharing
WO2021008675A1 (en) Dynamic network configuration
Imani et al. iSample: Intelligent client sampling in federated learning
CN115329985A (en) Unmanned cluster intelligent model training method and device and electronic equipment
CN115119233A (en) Clustered wireless communication method and system
CN114785692A (en) Virtual power plant aggregation regulation and control communication network flow balancing method and device
CN112015556B (en) Mobile crowd sensing data balancing method based on block chain rights and interests certification mechanism
CN111542078B (en) Core network control plane elastic resource allocation method in NFV environment
KR101565694B1 (en) Artificial bee colony clustering design method and system in wireless sensor network
CN114997877B (en) IP network topology data analysis method of virtual currency public-link network transaction node, storage medium and electronic equipment
CN104320772A (en) Trust degree and physical distance based D2D (Device to Device) communication node clustering method and device
Jovanovic et al. Comparison of different topologies for island-based multi-colony ant algorithms for the minimum weight vertex cover problem
CN114401192B (en) Multi-SDN controller cooperative training method
CN103096380B (en) Wireless access point load balancing load balancing
Sinthiya et al. Low-cost Task Offloading Scheme for Mobile Edge Cloud and Internet Cloud Using Genetic Algorithm
Cheng et al. An approximate bandwidth allocation algorithm for tradeoff between fairness and throughput in WSN
Wieselthier et al. Standard clock simulation and ordinal optimization applied to admission control in integrated communication networks
CN115345320A (en) Method for realizing personalized model under layered federal learning framework
CN104219679A (en) Improved method for assigning channels in cognitive radio networks
Gao et al. FedIM: An Anti-attack Federated Learning Based on Agent Importance Aggregation

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