CN109377218A - A kind of method, server and the mobile terminal of the false perception attack of containment - Google Patents
A kind of method, server and the mobile terminal of the false perception attack of containment Download PDFInfo
- Publication number
- CN109377218A CN109377218A CN201811101427.5A CN201811101427A CN109377218A CN 109377218 A CN109377218 A CN 109377218A CN 201811101427 A CN201811101427 A CN 201811101427A CN 109377218 A CN109377218 A CN 109377218A
- Authority
- CN
- China
- Prior art keywords
- perception
- task
- data
- payment rule
- payment
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/382—Payment protocols; Details thereof insuring higher security of transaction
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Computer Security & Cryptography (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Telephonic Communication Services (AREA)
Abstract
The embodiment of the invention provides method, server and the mobile terminals of a kind of false perception attack of containment, this method comprises: obtaining the first perception task, formulate the first payment rule according to the first perception task；The first perception task and the first payment rule are sent to multiple mobile terminals, so that mobile terminal chooses whether to participate in the first perception task according to the first payment rule；It obtains and participates in the perception data that multiple mobile terminals of the first perception task are sent, EM algorithm evaluation is carried out to perception data, obtains the corresponding perception accuracy of each perception data；According to the first payment rule, it is based on the corresponding perception accuracy of each perception data, to the corresponding remuneration of each mobile terminal payment；The second payment rule is obtained, the second payment rule is obtained according to the perception accuracy of the first perception task based on Q-learning algorithm or DQN algorithm, and the second payment rule is for perception task next time.The embodiment of the present invention is provided with the effect that containment user sends false perception attack.
Description
Technical field
The present embodiments relate to the methods of intelligent perception field more particularly to a kind of false perception attack of containment, service
Device and mobile terminal.
Background technique
It is more and more with the swift and violent growth of the mobile devices such as smart phone, tablet computer, smartwatch and Intelligent bracelet
Mobile device be equipped with the sensors of various functions, such as accelerometer, gyroscope, global positioning system and thermometer, with
These mobile devices be perception basic unit, gradually formed mobile gunz sensing network (Mobile Crowdsesning, with
Lower abbreviation MCS), by cooperating with mobile Internet, distributed awareness task simultaneously collects the perception data of mobile device upload, complete
At large-scale perception task.Therefore, environment, network and in terms of, MCS platform or server are moved by recruiting
The situation that family is monitored ambient enviroment is employed, to provide numerous services.It is quick with Intelligent programmable wireless device
Development, user can be to their wireless device be controlled, for example, user is by manipulating some distinctive built-in sensings freely
Device can accurately determine the effort to be paid for completing perception task, further will affect the quality of data.As one
Private smart phone user can select perception to make great efforts to maximize personal income, and intelligent perception system must stimulate user to mention
For accurately sensing report, and inhibit to forge the attack of sensing data.Otherwise, if user learns sends falseness in MCS task
Perception data will not pay for or even certain smart phone users upload false perception data by excitation and attacked, it will
MCS server is caused to receive the perception report of a large amount of low forgery.
To solve the above-mentioned problems, game theory is the important means for formulating MCS process, such as auctions, is based on price or base
In mechanism such as prestige, it is that MCS task is made contributions that user is motivated using these mechanism.Wherein, the MCS based on auction is proposed
The price in user's auction that solution payment is bid minimum is to save cost.It was noted that effectiveness MCS server not
The payment user to service is depended only on, their position is additionally depended on, incudes the quality of dynamics and sensor.Therefore,
MCS server, which can improve its sensing capabilities by assessment sensing quality and only recruit, provides accurate smart phone report.It moves
Innervation answers server application data mining and learning algorithm to assess false sensing report and can inhibit the motivation of deception.But by
In estimation error, the accuracy that server excites user to provide report in the case where not knowing the sensing model of user is still obtained
Less than guarantee.
Due to the mistake that assessment aspect occurs, how server excites use in the unwitting situation of user's sensing model
Family, which provides accurately report, becomes current institute's facing challenges.Therefore, a kind of method of false perception attack of containment is needed now.
Summary of the invention
The embodiment of the present invention is to solve that perception that the user in the prior art in MCS provides is reported lower to be lacked
It falls into, provides method, server and the mobile terminal of a kind of false perception attack of containment.
In a first aspect, the embodiment of the invention provides a kind of methods of the false perception attack of containment, comprising:
101, the first perception task is obtained, the first payment rule is formulated according to first perception task；
102, first perception task and first payment rule are sent to multiple mobile terminals, for mobile terminal
It is chosen whether to participate in first perception task according to first payment rule；
103, it obtains and participates in the perception data that multiple mobile terminals of first perception task are sent, to the perception number
According to EM algorithm evaluation is carried out, the corresponding perception accuracy of each perception data is obtained；
104, according to first payment rule, it is based on the corresponding perception accuracy of each perception data, to each movement
Terminal pays corresponding remuneration；
105, the second payment rule is obtained, second payment rule is accurate according to the perception of first perception task
Degree, is obtained based on Q-learning algorithm or DQN algorithm, and second payment rule is for perception task next time.
Second aspect, the embodiment of the invention provides a kind of methods of the false perception attack of containment, comprising:
Obtain perception task and payment rule；
Perception task remuneration is estimated in acquisition, described to estimate perception task remuneration according to the perception task and payment rule
Then, it estimates to obtain based on perceived quality；
It estimates task remuneration according to described and chooses whether to receive the perception task, if receiving, complete the perception
After task, server is sent by the perception data of the perception task；
Receive the corresponding remuneration of the perception task, the corresponding remuneration of the perception task is quasi- according to the perception of the perception data
Exactness and the payment rule obtain.
The third aspect, the embodiment of the invention provides a kind of servers of the false perception attack of containment, comprising:
First processing module is formulated the first payment according to first perception task and is advised for obtaining the first perception task
Then；
First sending module, it is whole to multiple movements for sending first perception task and first payment rule
End, so that mobile terminal chooses whether to participate in first perception task according to first payment rule；
Second processing module, for obtaining the perception number for participating in multiple mobile terminals of first perception task and sending
According to, to the perception data carry out EM algorithm evaluation, obtain the corresponding perception accuracy of each perception data；
First payment module, for it is accurate to be based on the corresponding perception of each perception data according to first payment rule
Degree, to the corresponding remuneration of each mobile terminal payment；
Third processing module, for obtaining the second payment rule, second payment rule is appointed according to first perception
The perception accuracy of business, is obtained based on Q-learning algorithm or DQN algorithm, and second payment rule is for sense next time
Know task.
Fourth aspect, the embodiment of the invention provides a kind of mobile terminals of the false perception attack of containment, comprising:
First obtains module, for obtaining perception task and payment rule；
Second obtains module, estimates perception task remuneration for obtaining, described to estimate perception task remuneration according to the sense
Know task and the payment rule, estimates to obtain based on perceived quality；
Selecting module chooses whether to receive the perception task for estimating task remuneration according to, if receiving,
After completing the perception task, server is sent by the perception data of the perception task；
Remuneration receiving module, for receiving the perception task remuneration, the perception task remuneration is according to the perception number
According to perception accuracy and the payment rule obtain.
5th aspect the embodiment of the invention provides a kind of computer equipment, including memory, processor and is stored in
On reservoir and the computer program that can run on a processor, the processor realize when executing described program such as first aspect or
The method of the false perception attack of containment described in second aspect.
6th aspect, the embodiment of the invention provides a kind of non-transient computer readable storage medium, the non-transient meter
Calculation machine readable storage medium storing program for executing stores computer instruction, and the computer instruction makes the computer execute such as first aspect or second
The method of the false number perception attack of containment described in aspect.
The method and device of the false perception attack of a kind of containment provided in an embodiment of the present invention, by EM algorithm to each sense
Know that the perception data of task is assessed, and according to perception data, using Q-learning algorithm or the study of DQN algorithm to most
Good payment rule, excitation user send most accurate perception data, achieve the effect that contain that user sends false perception data.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is the method flow schematic diagram of the false perception attack of containment provided in an embodiment of the present invention；
Fig. 2 is the method flow schematic diagram of the false perception attack of another containment provided in an embodiment of the present invention；
Fig. 3 is the server architecture schematic diagram of the false perception attack of containment provided in an embodiment of the present invention；
Fig. 4 is the mobile terminal structure schematic diagram of the false perception attack of containment provided in an embodiment of the present invention；
Fig. 5 is computer equipment structural schematic diagram provided in an embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
MCS is a kind of using mobile devices such as the smart phones, tablet computer or wearable device of user as basic sense
Know node, the wireless network of perception task and collection perception data is issued by moving internet.Game theory is current research gunz
The important means of sensing network, such as Game Theory, Option Pricing and credit system are applied in intelligent perception network,
For motivating user to participate in perception task.However, interests of some malicious users for itself, false perception data is sent
To MCS, on the one hand reduce network trap, on the other hand, but also the quality decline of network aware report, or even cause network
Blocking.
Fig. 1 is the method flow schematic diagram of the false perception attack of containment provided in an embodiment of the present invention, as shown in Figure 1, this
Inventive embodiments provide a kind of method of false perception attack of containment, comprising:
Step 101 obtains the first perception task, formulates the first payment rule according to first perception task；
MCS server is responsible for data collection, processing and application, and is made of multiple aware platform servers, and server is first
First perception task is classified, currently, with intelligent perception network research development and deeply, perception task is divided into environment
Monitoring, infrastructure monitoring, Social behaviors and social medical information etc., while according to sorted mission requirements, it collects specific
The perception information of user, in order to which the perception behavior for paying user obtains reasonable return, server formulates branch according to perception task
Pay rule.It should be noted that the first perception task and the first payment rule in the embodiment of the present invention are intelligent perception networks
Establish initial stage setting, wherein the first payment rule according in perception task and network historical data, artificial experience or
Classified reference data is formulated in internet, and the second payment rule and later described after the embodiment of the present invention
It is to learn to obtain by algorithm iteration, therefore, the first payment rule is at the beginning of MCS by the new payment rule for learning to obtain
Phase may show distribution principle it is unreasonable or distribution the lower defect of accuracy rate, pass through the iteration of payment rule new below
Study, new payment rule will gradually tend to be optimal.
Step 102 sends first perception task and first payment rule to multiple mobile terminals, for movement
Terminal chooses whether to participate in first perception task according to first payment rule；
Firstly, MCS server broadcast recruit information, recruit information in include the first perception task and the first payment rule,
Wherein, the remuneration information for completing corresponding first perception task is carried in the first payment rule, so that user be motivated to play an active part in
Perception task.After user receives recruitment information by mobile terminal, the resource paid according to oneself, for example, when perception consumption
Between, terminal power consumption or CPU occupancy situation, consider after paying these resources, whether the remuneration acquired meets expectation, from
And decides whether selection and participate in perception task.
Step 103 obtains the perception data for participating in multiple mobile terminals transmission of first perception task, to the sense
Primary data carries out EM algorithm evaluation, obtains the corresponding perception accuracy of each perception data；
In MCS, any user for carrying mobile terminal can receive perception task, the perception for causing user to upload
The accuracy of data cannot be guaranteed, and especially upload as user to mobile terminal misoperation or certain malicious users false
Perception attack, therefore, the reliability of perception data is assessed, so that the accuracy of entire MCS could be improved.In this hair
In bright embodiment, user submit perception data accuracy be it is unknown, need to carry out Accuracy evaluation, since there are hidden changes
Amount, i.e., the exact value (section) of each perception task, so cannot directly be estimated using maximum likelihood, so calculated using EM
Method.
It in embodiments of the present invention, is each user akSet an effort matrix ek, e herekIt is the square of a m*m
Battle array, each element of the insideWherein, i=1,2 ... m, j=1,2...m indicate user akThe perception of submission
Data are in section djIn, but accurately perception data is in section diIn, specifically,Include
The possible situation of the m kind for the correct perception data that user submits, wherein the value in matrix meets following formula:
Then, the interval probability collection that defining each perception task may be distributed is combined intoAccording to receipts
The data set S of collection initializes P, then executes the E step in EM algorithm, the i.e. value of estimation effort square matrix E, then executes in EM algorithm
M step, the correct section P of estimation task is gone according to the value of obtained square matrix E in turn, later constantly execute E step and M step, until
Convergence, wherein t indicates t-th of perception task.
Step 104, according to first payment rule, the corresponding perception accuracy of each perception data is based on, to each
The corresponding remuneration of mobile terminal payment；
It is set as according to the effort that the user j that EM algorithm is estimated is paid in period kSo server is available
The scale of precision of estimation is that the total number of the report of i can be indicated with following formula:
Wherein, I function is a Knowledge Function, whenWhen equation is set up, I=1, otherwise I=0.
The benefit of server end refers to that the obtained interests of server subtract expense to be paid, formula are as follows:
Wherein, G(i)Indicate the income that server obtains the data that a grade is i, in embodiments of the present invention, due to
The factors such as sensing location and submission time, the data that user submits are affected to the contribution of server, can set in the formula
Setting impact factor is λ, then the data of grade i is submitted for user j, the income that server obtains is λjG(i).Y indicates payment rule
Then, in embodiments of the present invention, the corresponding perception accuracy of payment rule y is divided into H rank, be expressed as
Wherein PiInclude H different payment rules.In this way, server is in period k according to the accuracy and payment rule that are estimated to
It can be indicated to the remuneration of user are as follows:
The remuneration paid to all users of the task of participation can indicate are as follows:
By above-mentioned formula, according to the perception accuracy for the perception data that EM algorithm evaluation arrives, in conjunction with payment rule, to
The user for participating in perception task pays corresponding remuneration, while obtaining the total benefit of intelligent perception network under current state.
Step 105 obtains the second payment rule, and second payment rule is quasi- according to the perception of first perception task
Exactness is obtained based on Q-learning algorithm or DQN algorithm, and second payment rule is for perception task next time.
It is exerted since the payment rule in MCS will affect mobile terminal user what the perception task that will be participated in be paid
Power, therefore, in embodiments of the present invention, payment process can be described as limited markov decision process (MDP).Due in work
In the dynamic environment that the sensing model of jump user hardly results in, MCS can apply Q-learning algorithm, a kind of model-free reinforcing
Learning method tests to obtain optimal payment strategy by limited Markovian decision.More specifically statement is based on Q-
The intelligent perception network of learning can arrive according to the observation before perception report quality, payment strategy and mass function
(the Q function for having discount to long-term reward) to determine payment strategy for each perception task.For example, in Q-learning algorithm
In, the ambient condition that period k as movement, can be arranged in payment rule is s(k), each accurate etc. comprising preceding state
Number of stages set and payment rule, formula are as follows:
Q (s, y) is set by the Q equation that the MCS payment strategy based on Q-learning is wherein relied on, then state is dynamic
(s, y) long-term expected utility of opposing is updated according to Bellman equation, formula are as follows:
Q(s,y)←(1-α)Q(s,y)+α(us(s,y)+γV(s'))；
Wherein s' is next state after state s implementation strategy y, and value equation V provides the maximum value of Q equation, γ
For discount factor, indicate that the time is longer, the reward that future obtains is lower, and α ∈ [0,1] indicates the learning efficiency of s-y-s'.
According to the state value s of current system(k)The value for the Q equation being calculated with action value, MCS server use ε-
Greedy algorithm removes selection action value, in this way can be to avoid resting on local optimum.It is specifically exactly to be acted in selection
When, the optimal policy predicted under current state is selected with the high probability of 1- ε:
Other strategies are randomly choosed with the probability of ε.
When state space reaches certain radix, due to needing a large amount of calculating, the operation effect of the MCS based on Q-learning
Rate will be very low, and this problem can pass through depth Q network (DQN) very good solution.More specifically, using applying instantly
The depth convolutional neural networks (CNN) and Q-learning algorithm of every field combine, and have both reached the safety of perception task,
Study statespace is had compressed again, improves operational efficiency.
A kind of method and device of the false perception attack of containment provided in an embodiment of the present invention, issues when MCS server and feels
After knowing task, initial payment rule is formulated first, excitation user uploads perception data, by EM algorithm to each perception task
Perception data assessed, and according to perception data, utilize Q-learning algorithm or the study of DQN algorithm to optimal branch
Rule is paid, further user is motivated to send most accurate perception data, achievees the effect that contain that user sends false perception data.
On the basis of the above embodiments, step 105, in the second payment rule of the acquisition, second payment rule
It according to the perception accuracy of first perception task, is obtained based on Q-learning algorithm or DQN algorithm, second payment
After rule is for perception task next time, comprising:
Based on next time perception task and second payment rule, repeat step 102 to 105, update current state
Benefit value, until MCS total benefit restrain.
Q-learning algorithm or DQN algorithm be one kind in dynamic environment by trial and error experimental learning behavior artificial intelligence
Technology, the action after being allowed to by study automatically selects ideal behavior in certain circumstances, thus the state being optimal.
In embodiments of the present invention, when being in the initial stage due to MCS, to the perception number of external web environment and mobile terminal upload
According to priori knowledge is lacked, the payment rule formulated at this time is simultaneously not perfect, and general mobile terminal user can select according to payment rule
The expected maximum sensed activation of benefit can be obtained by selecting.
However, the mobile terminal user of certain malice may find certain payment rule according to the not perfect of payment rule
Loophole then uploads false perception data with lesser cost and carrys out payt to cheat MCS, therefore, MCS at this time easily by
To falseness perception attack.By Q-learning algorithm or DQN algorithm to MCS iterative learning, server can sense to observing
The last perception data knowing that report quality and payment rule etc. become familiar with grasp, and arriving according to the observation, is gradually increased sense
Know the quotation of task, adjust price catalog, final server obtains optimal payment rule, so that malicious user uploads false perception
The probability of attack is preferably minimized to reach a stationary value.
In embodiments of the present invention, it is iterated based on Q-learning algorithm or DQN algorithm, has obtained payment rule,
On the one hand excitation user avoids uploading false perception attack, on the other hand as far as possible using the perception data for uploading high quality
Reduce the perception data that MCS is user's upload high quality and the remuneration paid.
On the basis of the above embodiments, step 105, the second payment rule of the acquisition, the second payment rule root
According to the perception accuracy of first perception task, obtained based on Q-learning algorithm or DQN algorithm, comprising:
According to the perception accuracy of first perception task, obtained based on Q-learning algorithm or DQN algorithm multiple
Payment rule；
Target payment rule is chosen from multiple payment rules according to ε-greedy algorithm as the second payment rule.
In the learning process of Q-learning algorithm or DQN algorithm, to avoid learning outcome false convergence, according to ε-
Greedy algorithm determines optimal payment rule, and in current perception task, server is gone to select most effective by the probability of ε
Payment rule randomly chooses other payment rules with the probability of 1- ε, wherein 0 < ε < 1, and levels off to 1.
In embodiments of the present invention, optimal payment rule is determined by ε-greedy algorithm, had both guaranteed to select every time in this way
The payment rule taken makes MCS obtain greatest benefit as far as possible, and prevents from falling into local maximum, to obtain global optimum
Payment rule.
On the basis of the above embodiments, step 101, the first perception task of the acquisition is appointed according to first perception
The first payment rule is formulated in business, comprising:
Using classified perception data as initial perception data；
The first perception task is obtained, the first payment is formulated according to first perception task and corresponding initial perception data
Rule.
In embodiments of the present invention, since there are the insufficient defects of priori by the MCS of initialization, to improve study effect
Rate, by having classified existing and perception data that accuracy rate has been assessed is input in MCS, these classified perception data categories
It in history perception data, is acquired according to the prior art, details are not described herein again for concrete mode.When server receives first
When perception task, the first perception task is classified, it should be noted that classified perception data is only used as payment rule
The reference sample of formulation establishes initial stage in MCS, initial payment rule is formulated by existing perception data and perception task,
Namely the first payment rule.In addition, also including that malice perception attack or perceived quality are lower in classified perception data
And undesirable data information.
By inputting existing initial perception data in the MCS of initialization, the efficiency of e-learning is improved, to make
It obtains entire MCS and reaches benefit convergence in a relatively short period of time, achieved the effect that the false perception of containment is attacked.
On the basis of the above embodiments, step 104, described according to first payment rule, it is based on each perception number
According to corresponding perception accuracy, to the corresponding remuneration of each mobile terminal payment, comprising:
If the perception accuracy is less than or equal to first threshold, determine that the perception data is false data attack；
If the perception accuracy is greater than first threshold and is less than or equal to second threshold, determine that the perception data is target
Perception data；
If the perception data is greater than the second threshold, determine that the perception data is excess perception data；
The corresponding remuneration of mobile terminal payment to the target apperception data and the excess perception data are determined as；
Wherein, the second threshold is greater than the first threshold.
In embodiments of the present invention, after the perception data of this perception task being assessed, to perception data by accurate
Grade is classified, it is assumed that each perception data can be attributed to one of them in scale of precision, setting first threshold and
Second threshold, it should be noted that second threshold schedule be this perception task best quality and satisfactory perception data,
That is, if the quality for the perception data that certain perception task needs need to only reach medium level, if it is more than to appoint that user, which submits,
Business requires the perception data of quality, also judges not meet perception task requirement.It should be noted that can be according to actual needs
The numerical value of first threshold and second threshold is set, and the present invention is not especially limit this.
For example, setting 0 for first threshold, second threshold is set as 1, and the perception by perception data less than or equal to 0 is reported
Be defined as false perception attack, wherein in actual assessment, can also the perception Report Definition by perception data less than 0 be to use
Family receives an assignment, but the case where cannot participate in perception task, is still defined as false perception at this point, perception data is equal to 0 and attacks
It hits, any remuneration will not be sent to the mobile terminal of such perception data at this time.When perception accuracy is greater than 0 and is less than or equal to
When 1, then defining perception data is target apperception data, pays corresponding remuneration to mobile terminal according to payment rule.Equally, work as sense
When knowing that accuracy is greater than 1, corresponding remuneration can be also paid according to payment rule at this time to mobile terminal, the difference is that, at this time
Perception data quality is higher, also more deviates the requirement of perception task, and the corresponding remuneration of payment rule is typically provided to perception number
It is fewer according to the higher payt of quality, so that the user of higher perceived equipment weighs the resource consumption situation of itself, to select
Whether selection receives perception task.
The embodiment of the present invention has contained that user initiates the general of false perception attack by setting first threshold and second threshold
Rate, user higher for perceived quality, makes such user keep silent under budget limit, reduces the redundancy of perception data
And the loss of transimission power.
Fig. 2 is the method for the false perception attack of another containment provided in an embodiment of the present invention, as shown in Fig. 2, the present invention is real
It applies example and provides a kind of method of false perception attack of containment, comprising:
Step 201 obtains perception task and payment rule；
Perception task remuneration is estimated in step 202, acquisition, described to estimate perception task remuneration according to the perception task and institute
Payment rule is stated, estimates to obtain based on perceived quality；
Step 203 estimates task remuneration according to and chooses whether to receive the perception task, if receiving, completes
After the perception task, server is sent by the perception data of the perception task；
Step 204 receives the corresponding remuneration of the perception task, and the corresponding remuneration of the perception task is according to the perception data
Perception accuracy and the payment rule obtain.
In embodiments of the present invention, firstly, MCS server broadcast recruits information, recruiting information includes perception task and branch
Rule is paid, wherein payment rule can motivate user to play an active part in perception task, when each user is received by mobile terminal
After recruiting information, the perceptual strategy of oneself is determined.Such as, if agreement receives perception task, if agreeing to, can think deeply this for sense
Know that task distribution how many resource are handled, since the quality of perception data depends on the perception of the sensor of mobile terminal
Degree, such as time, the electricity of perception consumption, mobile terminal can be according to the estimated distribution resource combination mobile terminal sheet of user at this time
The perception dynamics of body estimates a perception task remuneration, if the perception task remuneration estimated has reached the ideal expectation value of user,
Then user will receive this perception task, and send MCS server by mobile terminal for perception data after the completion,
Server assesses the perception accuracy to perception data, and according to payment rule, sends corresponding report to mobile terminal
Reward.
In embodiments of the present invention, mobile terminal is according to payment rule, in conjunction with the perception dynamics of itself, estimates out this sense
Know task remuneration, user can analyze the ideal expectation value of itself early stage receiving perception task, select next
Corresponding actions have contained the probability of false perception attack to improve the accuracy of perception data.
Fig. 3 is the server architecture schematic diagram of the false perception attack of containment provided in an embodiment of the present invention, as shown in figure 3,
The embodiment of the invention provides a kind of servers of the false perception attack of containment, comprising: first processing module 301, sending module
302, Second processing module 303, payment module 304 and third processing module 305, wherein first processing module 301 is for obtaining
First perception task formulates the first payment rule according to first perception task；Sending module 302 is for sending described first
Perception task and first payment rule are to multiple mobile terminals, so that mobile terminal is selected according to first payment rule
Whether first perception task is participated in；Second processing module 303 is for obtaining the multiple shiftings for participating in first perception task
The perception data that dynamic terminal is sent carries out EM algorithm evaluation to the perception data, obtains the corresponding perception of each perception data
Accuracy；Payment module 304 is used to be based on the corresponding perception accuracy of each perception data according to first payment rule,
To the corresponding remuneration of each mobile terminal payment；Third processing module 305 is for obtaining the second payment rule, second payment
Rule is obtained according to the perception accuracy of first perception task based on Q-learning algorithm or DQN algorithm, and described second
Payment rule is for perception task next time.
In embodiments of the present invention, third processing module 305 is based on Q-learning algorithm or DQN algorithm to this perception
The perception data of task is iterated, and is obtained the payment rule for perception task next time, is on the one hand motivated user using upper
The perception data of biography high quality avoids uploading falseness perception attack, and on the other hand reduction MCS as far as possible is uploaded high-quality for user
The perception data of amount and the remuneration paid.
Fig. 4 is the mobile terminal structure schematic diagram of the false perception attack of containment provided in an embodiment of the present invention, such as Fig. 4 institute
Show, the embodiment of the invention provides a kind of mobile terminals of the false perception attack of containment, comprising: first obtains module 401, second
Obtain module 402, selecting module 403 and remuneration receiving module 404, wherein the first acquisition module 401 is for obtaining perception task
And payment rule；Second acquisition module 402 estimates perception task remuneration for obtaining, described to estimate perception task remuneration according to institute
Perception task and the payment rule are stated, estimates to obtain based on perceived quality；Selecting module 403 according to for estimating task
Remuneration chooses whether to receive the perception task, if receiving, after completing the perception task, by the sense of the perception task
Primary data is sent to server；Remuneration receiving module 404 is for receiving the perception task remuneration, the perception task remuneration root
It is obtained according to the perception accuracy and the payment rule of the perception data.
In embodiments of the present invention, second module 402 is obtained according to payment rule, in conjunction with the perception dynamics of mobile terminal,
This perception task remuneration is estimated out, user can analyze the ideal expectation value of itself early stage receiving perception task,
Next corresponding actions are selected, to improve the accuracy of perception data, have contained the probability of false perception attack.
Device provided in an embodiment of the present invention is for executing above-mentioned each method embodiment, detailed process and detailed content
Above-described embodiment is please referred to, details are not described herein again.
Fig. 5 is computer equipment structural schematic diagram provided in an embodiment of the present invention, as shown in figure 5, the computer equipment can
To include: processor (processor) 501,502, memory communication interface (Communications Interface)
(memory) 503 and communication bus 504, wherein processor 501, communication interface 502, memory 503 pass through communication bus 504
Complete mutual communication.Processor 501 can call the logical order in memory 503, to execute following method: obtain the
One perception task formulates the first payment rule according to first perception task；Send first perception task and described
One payment rule is to multiple mobile terminals, so that mobile terminal chooses whether to participate in described first according to first payment rule
Perception task；It obtains and participates in the perception data that multiple mobile terminals of first perception task are sent, to the perception data
EM algorithm evaluation is carried out, the corresponding perception accuracy of each perception data is obtained；According to first payment rule, based on each
The corresponding perception accuracy of perception data, to the corresponding remuneration of each mobile terminal payment；Obtain the second payment rule, described
Two payment rules are obtained according to the perception accuracy of first perception task based on Q-learning algorithm or DQN algorithm, institute
The second payment rule is stated for perception task next time；
Or, obtaining perception task and payment rule；Perception task remuneration is estimated in acquisition, described to estimate perception task remuneration root
According to the perception task and the payment rule, estimate to obtain based on perceived quality；According to it is described estimate task remuneration selection be
It is no to receive the perception task, if receiving, after completing the perception task, the perception data of the perception task is sent
To server；Receive the corresponding remuneration of the perception task, the corresponding remuneration of the perception task is according to the perception of the perception data
Accuracy and the payment rule obtain.
In addition, the logical order in above-mentioned memory 503 can be realized by way of SFU software functional unit and conduct
Independent product when selling or using, can store in a computer readable storage medium.Based on this understanding, originally
Substantially the part of the part that contributes to existing technology or the technical solution can be in other words for the technical solution of invention
The form of software product embodies, which is stored in a storage medium, including some instructions to
So that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation of the present invention
The all or part of the steps of example the method.And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM,
Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. it is various
It can store the medium of program code.
The embodiment of the present invention discloses a kind of computer program product, and the computer program product is non-transient including being stored in
Computer program on computer readable storage medium, the computer program include program instruction, when described program instructs quilt
When computer executes, computer is able to carry out method provided by above-mentioned each method embodiment, for example, obtains the first perception
Task formulates the first payment rule according to first perception task；Send first perception task and first payment
Rule arrives multiple mobile terminals, so that mobile terminal chooses whether to participate in first perception times according to first payment rule
Business；It obtains and participates in the perception data that multiple mobile terminals of first perception task are sent, EM is carried out to the perception data
Algorithm evaluation obtains the corresponding perception accuracy of each perception data；According to first payment rule, it is based on each perception number
According to corresponding perception accuracy, to the corresponding remuneration of each mobile terminal payment；Obtain the second payment rule, second payment
Rule is obtained according to the perception accuracy of first perception task based on Q-learning algorithm or DQN algorithm, and described second
Payment rule is for perception task next time；
Or, obtaining perception task and payment rule；Perception task remuneration is estimated in acquisition, described to estimate perception task remuneration root
According to the perception task and the payment rule, estimate to obtain based on perceived quality；According to it is described estimate task remuneration selection be
It is no to receive the perception task, if receiving, after completing the perception task, the perception data of the perception task is sent
To server；Receive the corresponding remuneration of the perception task, the corresponding remuneration of the perception task is according to the perception of the perception data
Accuracy and the payment rule obtain.
The embodiment of the present invention provides a kind of non-transient computer readable storage medium, the non-transient computer readable storage medium
The instruction of matter storage server, the computer instruction make computer execute the false perception attack of containment provided by above-described embodiment
Method, for example, obtain the first perception task, the first payment rule is formulated according to first perception task；Described in transmission
First perception task and first payment rule are to multiple mobile terminals, so that mobile terminal is according to first payment rule
It chooses whether to participate in first perception task；It obtains and participates in the perception that multiple mobile terminals of first perception task are sent
Data carry out EM algorithm evaluation to the perception data, obtain the corresponding perception accuracy of each perception data；According to described
One payment rule is based on the corresponding perception accuracy of each perception data, to the corresponding remuneration of each mobile terminal payment；It obtains
Second payment rule, second payment rule are based on Q-learning according to the perception accuracy of first perception task
Algorithm or DQN algorithm obtain, and second payment rule is for perception task next time；
Or, obtaining perception task and payment rule；Perception task remuneration is estimated in acquisition, described to estimate perception task remuneration root
According to the perception task and the payment rule, estimate to obtain based on perceived quality；According to it is described estimate task remuneration selection be
It is no to receive the perception task, if receiving, after completing the perception task, the perception data of the perception task is sent
To server；Receive the corresponding remuneration of the perception task, the corresponding remuneration of the perception task is according to the perception of the perception data
Accuracy and the payment rule obtain.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of method of the false perception attack of containment characterized by comprising
101, the first perception task is obtained, the first payment rule is formulated according to first perception task；
102, first perception task and first payment rule are sent to multiple mobile terminals, for mobile terminal according to
First payment rule chooses whether to participate in first perception task；
103, obtain and participate in the perception data that multiple mobile terminals of first perception task are sent, to the perception data into
Row EM algorithm evaluation obtains the corresponding perception accuracy of each perception data；
104, according to first payment rule, it is based on the corresponding perception accuracy of each perception data, to each mobile terminal
Pay corresponding remuneration；
105, the second payment rule, perception accuracy of second payment rule according to first perception task, base are obtained
It is obtained in Q-learning algorithm or DQN algorithm, second payment rule is for perception task next time.
2. the method according to claim 1, wherein described second pays in the second payment rule of the acquisition
Rule is obtained according to the perception accuracy of first perception task based on Q-learning algorithm or DQN algorithm, and described second
After payment rule is for perception task next time, comprising:
Based on next time perception task and second payment rule, repeat step 102 to 105, update the effect of current state
Benefit value, until the total benefit of intelligent perception network is restrained.
3. method according to claim 1 or 2, which is characterized in that the second payment rule of the acquisition, second payment
Rule is obtained according to the perception accuracy of first perception task based on Q-learning algorithm or DQN algorithm, comprising:
According to the perception accuracy of first perception task, multiple payments are obtained based on Q-learning algorithm or DQN algorithm
Rule；
Target payment rule is chosen from multiple payment rules according to ε-greedy algorithm as the second payment rule.
4. the method according to claim 1, wherein the first perception task of the acquisition, feels according to described first
Know that task formulates the first payment rule, comprising:
Using classified perception data as initial perception data；
The first perception task is obtained, the first payment is formulated according to first perception task and corresponding initial perception data and is advised
Then.
5. being based on each sense the method according to claim 1, wherein described according to first payment rule
The corresponding perception accuracy of primary data, to the corresponding remuneration of each mobile terminal payment, comprising:
If the perception accuracy is less than or equal to first threshold, determine that the perception data is false data attack；
If the perception accuracy is greater than first threshold and is less than or equal to second threshold, determine that the perception data is target apperception
Data；
If the perception data is greater than the second threshold, determine that the perception data is excess perception data；
The corresponding remuneration of mobile terminal payment to the target apperception data and the excess perception data are determined as；
Wherein, the second threshold is greater than the first threshold.
6. a kind of method of the false perception attack of containment characterized by comprising
Obtain perception task and payment rule；
Perception task remuneration is estimated in acquisition, described to estimate perception task remuneration according to the perception task and the payment rule,
It estimates to obtain based on perceived quality；
It estimates task remuneration according to described and chooses whether to receive the perception task, if receiving, complete the perception task
Afterwards, server is sent by the perception data of the perception task；
Receive the corresponding remuneration of the perception task, the corresponding remuneration of the perception task is according to the perception accuracy of the perception data
It is obtained with the payment rule.
7. a kind of server of the false perception attack of containment characterized by comprising
First processing module formulates the first payment rule according to first perception task for obtaining the first perception task；
Sending module, for sending first perception task and first payment rule to multiple mobile terminals, for moving
Dynamic terminal chooses whether to participate in first perception task according to first payment rule；
Second processing module is right for obtaining the perception data for participating in multiple mobile terminals of first perception task and sending
The perception data carries out EM algorithm evaluation, obtains the corresponding perception accuracy of each perception data；
Payment module, for being based on the corresponding perception accuracy of each perception data according to first payment rule, to each
The corresponding remuneration of mobile terminal payment；
Third processing module, for obtaining the second payment rule, second payment rule is according to first perception task
Accuracy is perceived, is obtained based on Q-learning algorithm or DQN algorithm, second payment rule is appointed for perception next time
Business.
8. a kind of mobile terminal of the false perception attack of containment characterized by comprising
First obtains module, for obtaining perception task and payment rule；
Second obtains module, estimates perception task remuneration for obtaining, described to estimate perception task remuneration according to the perception times
Business and the payment rule, estimate to obtain based on perceived quality；
Selecting module chooses whether to receive the perception task for estimating task remuneration according to, if receiving, completes
After the perception task, server is sent by the perception data of the perception task；
Remuneration receiving module, for receiving the perception task remuneration, the perception task remuneration is according to the perception data
Perception accuracy and the payment rule obtain.
9. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes the containment as described in any one of claim 1 to 6 when executing described program
The method of false perception data.
10. a kind of non-transient computer readable storage medium, which is characterized in that the non-transient computer readable storage medium is deposited
Computer instruction is stored up, the computer instruction executes the computer and contains false sense as described in any one of claim 1 to 6
The method of primary data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811101427.5A CN109377218B (en) | 2018-09-20 | 2018-09-20 | Method, server and mobile terminal for suppressing false sensing attack |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811101427.5A CN109377218B (en) | 2018-09-20 | 2018-09-20 | Method, server and mobile terminal for suppressing false sensing attack |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109377218A true CN109377218A (en) | 2019-02-22 |
| CN109377218B CN109377218B (en) | 2020-10-27 |
Family
ID=65405711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811101427.5A Active CN109377218B (en) | 2018-09-20 | 2018-09-20 | Method, server and mobile terminal for suppressing false sensing attack |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109377218B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110365647A (en) * | 2019-06-13 | 2019-10-22 | 广东工业大学 | A kind of false data detection method for injection attack based on PCA and BP neural network |
| CN112016047A (en) * | 2020-07-24 | 2020-12-01 | 浙江工业大学 | Heuristic data acquisition method and device based on evolutionary game, computer equipment and application thereof |
| CN112258420A (en) * | 2020-11-02 | 2021-01-22 | 北京航空航天大学杭州创新研究院 | DQN-based image enhancement processing method and device |
| CN113516229A (en) * | 2021-07-09 | 2021-10-19 | 哈尔滨理工大学 | Credible user optimization selection method facing crowd sensing system |
| CN113885330A (en) * | 2021-10-26 | 2022-01-04 | 哈尔滨工业大学 | Information physical system safety control method based on deep reinforcement learning |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103647671A (en) * | 2013-12-20 | 2014-03-19 | 北京理工大学 | Gur Game based crowd sensing network management method and system |
| CN105809477A (en) * | 2016-03-04 | 2016-07-27 | 武汉大学 | Information quality based participation-type perception encouragement method |
| CN108337656A (en) * | 2018-01-16 | 2018-07-27 | 武汉工程大学 | A kind of mobile intelligent perception motivational techniques |
-
2018
- 2018-09-20 CN CN201811101427.5A patent/CN109377218B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103647671A (en) * | 2013-12-20 | 2014-03-19 | 北京理工大学 | Gur Game based crowd sensing network management method and system |
| CN105809477A (en) * | 2016-03-04 | 2016-07-27 | 武汉大学 | Information quality based participation-type perception encouragement method |
| CN108337656A (en) * | 2018-01-16 | 2018-07-27 | 武汉工程大学 | A kind of mobile intelligent perception motivational techniques |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110365647A (en) * | 2019-06-13 | 2019-10-22 | 广东工业大学 | A kind of false data detection method for injection attack based on PCA and BP neural network |
| CN110365647B (en) * | 2019-06-13 | 2021-09-14 | 广东工业大学 | False data injection attack detection method based on PCA and BP neural network |
| CN112016047A (en) * | 2020-07-24 | 2020-12-01 | 浙江工业大学 | Heuristic data acquisition method and device based on evolutionary game, computer equipment and application thereof |
| CN112258420A (en) * | 2020-11-02 | 2021-01-22 | 北京航空航天大学杭州创新研究院 | DQN-based image enhancement processing method and device |
| CN112258420B (en) * | 2020-11-02 | 2022-05-20 | 北京航空航天大学杭州创新研究院 | DQN-based image enhancement processing method and device |
| CN113516229A (en) * | 2021-07-09 | 2021-10-19 | 哈尔滨理工大学 | Credible user optimization selection method facing crowd sensing system |
| CN113885330A (en) * | 2021-10-26 | 2022-01-04 | 哈尔滨工业大学 | Information physical system safety control method based on deep reinforcement learning |
| CN113885330B (en) * | 2021-10-26 | 2022-06-17 | 哈尔滨工业大学 | Information physical system safety control method based on deep reinforcement learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109377218B (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109377218A (en) | A kind of method, server and the mobile terminal of the false perception attack of containment | |
| Du et al. | Learning resource allocation and pricing for cloud profit maximization | |
| CN109767319A (en) | The accrediting amount determines method, apparatus, computer equipment and storage medium | |
| CN111340244B (en) | Prediction method, training method, device, server and medium | |
| CN108550090A (en) | A kind of processing method and system of determining source of houses pricing information | |
| Patel | Optimizing market making using multi-agent reinforcement learning | |
| US20190073244A1 (en) | Computer network-based event management | |
| EP3961507A1 (en) | Optimal policy learning and recommendation for distribution task using deep reinforcement learning model | |
| Salk et al. | Limitations of majority agreement in crowdsourced image interpretation | |
| Baek et al. | Small profits and quick returns: An incentive mechanism design for crowdsourcing under continuous platform competition | |
| CN113689270B (en) | Method for determining black product device, electronic device, storage medium, and program product | |
| Xu et al. | Distributed no-regret learning in multiagent systems: Challenges and recent developments | |
| Reddy et al. | Negotiated learning for smart grid agents: entity selection based on dynamic partially observable features | |
| CN112037063A (en) | Exchange rate prediction model generation method, exchange rate prediction method and related equipment | |
| Parhizkar et al. | Indirect Trust is Simple to Establish. | |
| US20170103341A1 (en) | Continual learning in slowly-varying environments | |
| CN111046156A (en) | Method and device for determining reward data and server | |
| CN113037648B (en) | Data transmission method and device | |
| CN115345663A (en) | Marketing strategy evaluation method and device, electronic equipment and storage medium | |
| CN116339932A (en) | Resource scheduling method, device and server | |
| Liu et al. | Mechanism learning with mechanism induced data | |
| CN110795232A (en) | Data processing method, data processing device, computer readable storage medium and computer equipment | |
| Turocy et al. | Reservation Values and Regret in Laboratory First‐Price Auctions: Context and Bidding Behavior | |
| Lim et al. | Assessing the accuracy of Grey System Theory against Artificial Neural Network in predicting online auction closing price | |
| Kheradmand et al. | Price and QoS competition in cloud market by using cellular learning automata |
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 |