CN110312306A - Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information - Google Patents
Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information Download PDFInfo
- Publication number
- CN110312306A CN110312306A CN201910537244.6A CN201910537244A CN110312306A CN 110312306 A CN110312306 A CN 110312306A CN 201910537244 A CN201910537244 A CN 201910537244A CN 110312306 A CN110312306 A CN 110312306A
- Authority
- CN
- China
- Prior art keywords
- local
- anchor node
- database
- information
- probability
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/18502—Airborne stations
- H04B7/18506—Communications with or from aircraft, i.e. aeronautical mobile service
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
- H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Astronomy & Astrophysics (AREA)
- Aviation & Aerospace Engineering (AREA)
- General Physics & Mathematics (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses an unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information, wherein the method comprises the following steps: s1, correspondingly configuring a database for storing information required by node positioning for each non-anchor node in an unmanned aerial vehicle cluster; s2, when the target non-anchor node measures or receives the state of the anchor node or receives messages sent by other nodes, updating the database corresponding to the target non-anchor node, and executing the step S3 after the updating is finished; and S3, the target non-anchor node generates broadcast information according to the information in the corresponding database so as to broadcast the broadcast information to the adjacent nodes for updating the state information, and updates the position information of the target non-anchor node, so that the positioning of the unmanned aerial vehicle node is completed. The method is suitable for large-scale unmanned aerial vehicle clusters, and has the advantages of simple implementation method, small communication bandwidth and calculation overhead, high positioning accuracy and the like.
Description
Technical field
The present invention relates to unmanned plane cluster field of locating technology more particularly to a kind of unmanned plane clusters based on asynchronous information
Cooperative localization method and device.
Background technique
Unmanned plane cluster receives unprecedented attention with its exclusive advantage in the world, has in military and civilian field
Huge application and economic value.The positioning of unmanned plane cluster cooperation formula can make cluster better adapt to complex environment, therefore
It has great application prospect.Unmanned plane is one kind of mobile node, and is still lacked about the cooperative positioning of mobile node at present
There is a series of problem of such as Network Synchronization and delay problem among these in weary effective scheme.
BP (Belief Propagation, confidence are mainly used for the localization method of mobile node in the prior art
Probability propagation) localization method, such method is existed based on factor graph (Factor Graph), and using Sum-Product algorithm
Fiducial probability is propagated in entire factor graph, when not having ring (Cycle) in factor graph, Sum-Product algorithm can be accurately
Obtain marginal probability, but for location algorithm, factor graph is to time and the extension of space both direction, even if in Spatial Dimension not
Comprising ring, time dimension also will necessarily include ring, otherwise just also just cannot achieve cooperative without the mutual measurement between node
Positioning, and when in factor graph including ring, Sum-Product method will be unable to obtain accurate as a result, can only be by repeatedly again
Transmitting information is iterated, and the number of iteration is more, as a result more accurate.
At present the BP localization method of mainstream can not unmanned plane cluster cooperation formula positioning in effective performance, due to have with
Lower defect:
(1) huge communication cost and computation complexity.BP localization method needs to carry out interactive iteration repeatedly, hands over each time
Mutual iteration all corresponds to the expense of communication and computing resource, and this expense can swash with the increase of positioning network size
Increase.Even if only considering bring time delay in communication framework (ignoring communication bandwidth constraint and the constraint of computing resource), BP is fixed
Position method is also difficult to support the real-time positioning of 10 or more mobile nodes, if it is considered that the limitation of the communication resource of real system,
Such case will be worse.However, the quantity of unmanned plane node will be far longer than 10 for unmanned plane cluster.Cause
Traditional BP algorithm can not solve slightly to have the unmanned plane cluster cooperation formula orientation problem of scale at present for this.
(2) dependence synchronous for clock.BP localization method needs all nodes in network to have synchronous clock,
Essential reason is that its movement node discretization model is to propagate clock based on information and establish.Once clock appearance is asynchronous,
By the ambiguousness in generation time between node, this ambiguousness can gradually add up with the increase of network size, to lead
Cause the mistake of motion information.
In conclusion existing unmanned plane cluster localization method, due to the filter of use be based on discrete time so that
It needs the cycle accurate of network to be aligned, and the communication in entire unmanned plane cluster needs stringent synchronization, thus needs to rely on
The synchronization of network communication, simultaneously because the filter used is based on discrete time, it is therefore desirable to the synchronization of measurement, different nothings
Man-machine time of measuring deviation will cause the inaccurate of positioning result, so that the synchronization for the measurement that also needs to rely on, and mainstream
BP localization method needs iterate, and being in communication in the big therefore current cooperative localization method of calculating cost can be only applied to pole
In small-scale unmanned plane cluster, it can not be adapted to slightly to have in the unmanned plane cluster cooperation formula positioning of scale.
Summary of the invention
The technical problem to be solved in the present invention is that, for technical problem of the existing technology, the present invention provides one
Kind can be suitable for extensive unmanned plane cluster, and implementation method is simple, communication bandwidth and computing cost are small, positioning accuracy
The high unmanned plane cluster cooperation formula localization method and device based on asynchronous information.
In order to solve the above technical problems, technical solution proposed by the present invention are as follows:
A kind of unmanned plane cluster cooperation formula localization method based on asynchronous information, step include:
S1. for the corresponding configuration of the non-anchor node of each in unmanned plane cluster one for memory node locating desired information
Database, the non-anchor node are the unmanned plane node positioned to itself Location-Unknown, needs;
S2. when the message for measuring or receiving anchor node state there are the non-anchor node of target or the other nodes of reception are sent
When, the corresponding database of the non-anchor node of target is updated, is transferred to after the completion of updating and executes step S3;
S3. the non-anchor node of target generates broadcast message according to the information in the corresponding database to be broadcast to adjacent segments
Point carries out state information updating, and updates the location information of itself, completes the positioning of unmanned plane node.
As the further improvement of the method for the present invention, the data packet of each moment record in the database of non-anchor node
Include local prior probability, local posterior probability, local measurement collection and local four partial data of anchor node state set, the local
Prior probability includes the priori joint probability density information of node relevant to positioning and constantly updates, the local posterior probability
Posteriority joint probability density information comprising node relevant to positioning is simultaneously constantly updated, and the local measurement collection includes and positioning
Relevant all anchor node status informations are simultaneously constantly accepted or rejected and are updated, and the local anchor node state set includes relevant to positioning
All anchor node status informations are simultaneously constantly accepted or rejected and are updated, and the broadcast message includes the local prior probability, the local
Measurement collection and the local anchor node state set three parts information.
As the further improvement of the method for the present invention, the database of non-anchor node specifically:
Wherein,For the corresponding database of first of non-anchor node,For updated database,For data
The renewal time in library gathers,It is updated for database kth time;
It willIt is defined asAnd:
Wherein,Exist for databaseThe record at moment,For it is all record the moment set,For
Local prior probability,For local posterior probability,For local measurement collection;For local anchor node
State set;T is current time,For s-th of record moment in kth time update, wherein k and s is respectively positive integer;
The broadcast message specifically:
Wherein,The broadcast message generated when being updated for first of non-anchor node kth time,Be so thatCollection at the time of establishment.
As the further improvement of the method for the present invention, in the step S2, first local zone time axis is updated, based on more
Time shaft after new using distributed priori dividing method to local measurement collection in the database, local anchor node state set,
Local prior probability and local posterior probability are updated, and specific steps include:
S21. whether judgement is currently the database moment, and the database moment is the newest moment in local zone time axis,
Step S22 is executed if it is being transferred to, is otherwise exited;
S22. local measurement collection and local anchor node state set update: judge renewable time whether the correspondence database moment,
If so, the local measurement collection after the measurement that local measurement concentrates addition new collects and updated, and in local anchor node state
It concentrates and local anchor node state set after new anchor node state set is updated is added, otherwise concentrated in local measurement and interaction is added
Measurement in information collects updated after local measurement collection, and in local anchor node state set be added interactive information in anchor
Node state collection local anchor node state set after being updated;
S23. Prior Fusion: respectively to each node l ' for including in the database of node l, prior variance is found out most
The prior distribution of all nodes found out is stitched together by small neighbor node j, forms joint prior distribution;
S24. priori is divided: the subset based on local measurement collectionLocal prior state setWith
And the subset of local anchor node state setTriple SNCAMST is defined, choosing has the described of smallest partition difference
Triple SNCAMST simultaneously calculates marginal probability;
S25. posteriority updates and priori prediction: the local posterior probability is updated using the marginal probability calculated, and
The local prior probability is predicted using the updated local posterior probability in the database last moment.
It is described when being updated to local zone time axis as the further improvement of the method for the present invention, it usesTo limit number
According to the maximum moment number for including in library, wherein(such as maximal memory constraintMean in the database of node l most
Multipotency includes the information of 3 last time), i.e., updated local zone time axisIt is sum time axisSubset, it is described
Sum time axisIt is old database time shaftWith the union of new arrival information time axis.
As the further improvement of the method for the present invention, the step of step S23, is specifically included: defining local prior varianceTo reflect node l ' in local prior probabilityUnder uncertainty, whereinFor node l '
State,For kth time update in s-th of record moment, j for neighbor node serial number, stored from database other
In node l ', corresponding minimum local prior variance is chosenAsProbability, it is the smallest local prior variance
Serial number specifically:
WhereinFor the corresponding optimal node ID of l ' of return, i.e., the serial number of the smallest local prior variance,
To include local prior variance in databaseAll nodes set;
Priori splicing is carried out according still further to following formula:
Wherein,For the set of the corresponding optimal node ID composition of all l '.
As the further improvement of the method for the present invention, the step of step S24 is specifically included: definition is based on local measurement
CollectionLocal measurement figure are as follows:
WhereinFor local measurement figureCorresponding vertex set, ε are local measurement figureCorresponding line set;
Defining the triple SNCAMST based on the local measurement figure is
WhereinWithIt is respectivelyWithSubset,For local prior probability,For local posterior probability,For local measurement collection,For local anchor node state set;
Divide difference are as follows:
Wherein,For segmentation after prior probability used for positioning,For the elder generation for being not used in positioning remaining after segmentation
Test probability;
It enablesIt is for maximum rating numberThe triple SNCAMST, be denoted asAll segmentation difference
Gather constituting are as follows:
Obtain the SNCAMST of smallest partition difference are as follows:
The SNCAMST of the smallest partition difference is calculated into marginal probability using following formula
Wherein,For local prior state set subset, andProbability density function be
As the further improvement of the method for the present invention, the step S25 specific steps are as follows:
S251. judge whether the local a prior set, local anchor node state set are sky, if the local a prior set or institute
Stating local anchor node state set is not sky, is transferred to and executes step S252, is otherwise transferred to and executes step S253;
S252. the local posterior probability is updated using Bayesian Method using the marginal probability of calculating, is turned
Enter to execute step S253;
S253. judge whether the current database moment is the last one moment in database, if so, using it is current this
Ground posterior probability predicts local posterior probability using Kolmogorov equation.
It is described that the local posterior probability is updated using Bayesian Method as the further improvement of the method for the present invention
Formula are as follows:
Wherein,For local prior state set,For marginal probability,It is
Subset,For local measurement collection;
The formula that local posterior probability is predicted using Kolmogorov equation are as follows:
Wherein,For local prior state setCorresponding supported collection (i.e. the set of probability density non-zero).
A kind of unmanned plane cluster cooperation formula positioning device based on asynchronous information, including computer equipment, the computer
Equipment is programmed to perform the step of above-mentioned unmanned plane cluster cooperation formula localization method based on asynchronous information.
Compared with the prior art, the advantages of the present invention are as follows:
1, the present invention is by configuring a database for the unmanned plane node that each needs positions in unmanned plane cluster, often
Database update is carried out when non-anchor node measures or receive anchor node state or receives the message of other nodes transmissions, according to
It is broadcast to adjacent node status information according to updated database, while updating the position of itself, it can be real based on asynchronous information
Existing multiple no-manned plane cluster cooperation formula positioning, needs not rely on clock and synchronizes, can be when network asynchronous communication is with asynchronous measurement, still
Correct joint movements information so described, and the traffic and calculation amount be independently of the scale of network, may be implemented it is extensive nobody
The positioning of machine cluster cooperation formula.
2, present invention may apply to asynchronous communication and asynchronous measurements in extensive unmanned plane cluster, are occupying a small amount of communication
Under the premise of bandwidth, accurate self poisoning can be carried out to unmanned plane, solve traditional mainstream based on the localization method of BP without
Method handles the problem of asynchronous communication and asynchronous measurement information, and is not limited by network size, carries out in extensive dynamic network
Communication and computing cost are small when positioning, efficiently solve tradition and carry out the huge communication of positioning needs to extensive dynamic network
And the problem of computing cost.
3, the present invention is based further on distributed priori dividing method and realizes database update, based on priori segmentation difference with
The continuity of location estimation difference, it can be ensured that the correct choice of the partial status of local priori, compared to traditional based on complete
The method of office's information, can effectively reduce position disparity, while improving location efficiency.
Detailed description of the invention
Fig. 1 is the implementation process schematic diagram of unmanned plane cluster cooperation formula localization method of the present embodiment based on asynchronous information.
Fig. 2 is the instrumentation plan in the present embodiment between anchor node and non-anchor node.
Fig. 3 is non-anchor node communication broadcast time diagram in the present embodiment.
Fig. 4 is data packet receiving time schematic diagram in the present embodiment.
Fig. 5 is the specific implementation flow schematic diagram of non-anchor node locating in the present embodiment.
Fig. 6 is the detailed implementation process signal for realizing database update in the present embodiment based on distributed priori dividing method
Figure.
Fig. 7 is the detailed implementation process schematic diagram that local posteriority updates with priori prediction in the present embodiment.
Specific embodiment
Below in conjunction with Figure of description and specific preferred embodiment, the invention will be further described, but not therefore and
It limits the scope of the invention.
As shown in Figure 1, the present embodiment based on asynchronous information unmanned plane cluster cooperation formula localization method the step of include:
S1. for the corresponding configuration of the non-anchor node of each in unmanned plane cluster one for memory node locating desired information
Database, non-anchor node are the unmanned plane node positioned to itself Location-Unknown, needs;
S2. when the message for measuring or receiving anchor node state there are the non-anchor node of target or the other nodes of reception are sent
When, the corresponding database of the non-anchor node of target is updated, is transferred to after the completion of updating and executes step S3;
S3. the non-anchor node of target generates broadcast message according to the information in the corresponding database to be broadcast to adjacent segments
Point carries out state information updating, and updates the location information of itself, completes the positioning of unmanned plane node.
The present embodiment is by configuring one for the unmanned plane node (non-anchor node) that each needs positions in unmanned plane cluster
A database is counted when non-anchor node measures or receive anchor node state or receives the message of other nodes transmissions
It is updated according to library, is broadcast to adjacent node status information according to updated database, while updating the position of itself, can be based on
Asynchronous information realizes the positioning of multiple no-manned plane cluster cooperation formula, and it is synchronous to need not rely on clock, can network asynchronous communication with it is different
When pacing amount, correct joint movements information is still described, and the traffic and calculation amount may be implemented independently of the scale of network
Extensive unmanned plane cluster cooperation formula positioning, i.e., when only a small number of unmanned plane known position informations, all unmanned planes pass through phase
Mutually cooperation can estimate the real time position of itself.
Different from discrete time joint movements model used in traditional BP localization method, the movement of the present embodiment interior joint
Model, so that metrical information and communication interaction information be made to be no longer influenced by the limitation of slot synchronization, was convenient for based on continuous time
Subsequent asynchronous information processing.Following specific model (includes specifically I node in system, with setTo indicate):
dxi(t)=fi(xi(t),t)dt+dβi(t), (1)
yi(t)=gi(xi(t)), (2)
Wherein, formula (1) is system equation, formula (2) is position equation, describes the state x of node i respectivelyi(t) drill
Law and position yi(t) changing rule, fiAnd giRespectively system function and position function, βiIt (t) is Brownian movement.
Metrical information used in the present embodiment is distance measurement information, can specifically pass through (the arrival time ranging side TOA
Method), TDOA (reaching time-difference distance measuring method) or RSS (received signal strength distance measuring method) obtain.When node i exists to node jWhen moment measures, equation is measured are as follows:
WhereinFor noisy measured value,To measure noise.
In concrete application embodiment between unmanned plane node metrical information as shown in Fig. 2, wherein a1,a2,a3For anchor node,
The state of anchor node is accurately known, l1,...,l5For non-anchor node, the state of non-anchor node obeys certain random distribution, in figure
Indicate that non-anchor node may be with the region of high probability appearance using dotted line as the dash area on boundary.By measuring, non-anchor node
The trend that high probability region generally journey reduces.
The transmitting of the communication information occurs only between non-anchor node in the present embodiment, content include distance measurement information and
The location-prior information of node collection.The communication information is transmitted in the form of broadcast data packet, and broadcast time is as shown in figure 3, with node 1
For,First data packet of the broadcast of node 1 is represented, wherein initial time isTerminate the time beThe present embodiment number
According to the coating received time as shown in figure 4, wherein broadcasting dwell time and being received between the time has one section of processing delay.For
Anchor node, the anchor node status information that need to be only issued by the non-anchor node of dynamic response send request.
Each specific non-anchor node of the present embodimentInclude a databaseL pairs of node is stored in database
Information needed for self poisoning;It is non-after measuring, receiving anchor node state or receive the message of other nodes transmissions
Anchor node l is updated the database to its own accordingly;Once database update is completed, non-anchor node l will be according to number
According to information design broadcast message existing in library, to help its neighbor node to update corresponding state information;Once database update
It completes, non-anchor node l will update the location information of itself, all unmanned planes in unmanned plane cluster using the information in database
The real time position of itself can be estimated by cooperating with each other.
As shown in figure 5, the present embodiment realizes non-anchor section based on the unmanned plane cluster cooperation formula location algorithm of asynchronous information
The detailed step of the positioning of point are as follows:
Step 1: initializing the state prior distribution of non-anchor node l
Step 2: if other nodes are measured or had received with the status information of anchor node or is had received other
The information that node is sent is transferred to and executes step 3;
Step 3: more new database
Step 4: databaseAfter update, broadcast message is designed, by formulaCalculating state point
Cloth;
Step 5: usingUpdate location estimation information.
The present embodiment is non-anchor to realize based on the unmanned plane cluster cooperation formula location algorithm of asynchronous information especially by configuring
The positioning of node, the detailed process of algorithm is as shown in following algorithms 1:
In the present embodiment, the data of each moment record include local prior probability, local in the database of non-anchor node
Posterior probability, local measurement collection and local four partial data of anchor node state set, local prior probability areComprising with
It positions the priori joint probability density information of relevant node and constantly updates, local posterior probability isComprising with it is fixed
The posteriority joint probability density information of the relevant node in position is simultaneously constantly updated, and local measurement collection isComprising with positioning phase
All measurements of pass are simultaneously constantly accepted or rejected and are updated, and local anchor node state set isComprising relevant to positioning all
Anchor node status information is simultaneously constantly accepted or rejected and is updated, and t is current time,S-th of record moment in being updated for kth is secondary,
Middle k and s is respectively positive integer;Broadcast message includes local prior probability, the local measurement collection and local anchor node state
Collect three parts information.
In the present embodiment, the database of non-anchor nodeOnly is measuring, receives anchor node state or receiving other
It is updated after the message that node is sent, thereforeIt is changed at the time of some discrete, it may be assumed that
Wherein,For the corresponding database of first of non-anchor node,For updated database,For data
The renewal time in library gathers,It is updated for database kth time.
To simplify the description, willIt is defined asIt is with following form:
Wherein,Exist for databaseThe record at moment,For it is all record the moment set,For
Local prior probability,For local posterior probability,For local measurement collection,For local anchor node
State set, t are current time,For s-th of record moment in kth time update (wherein k and s is respectively positive integer).For
RecordIt is made of four parts, i.e., local prioriLocal posteriorityLocal measurement collectionWith local anchor node state set
Only local priori in the present embodimentLocal measurement collectionWith local anchor node state setParticipate in the information exchange between node, local posteriorityIt is not involved in interaction, i.e., only comprising above-mentioned in broadcast message
Three kinds of information, concrete form are as follows:
Wherein,Be so thatEstablishment when
Collection is carved, i.e.,Contain only the current database part different from database before.
It is not that all parts in broadcast message can successfully be connect by other nodes in view of practical communication condition
It receives, thus receives information and can be the whole or only part of it of broadcast message.In the present embodimentMoment, non-anchor section
The concrete form that point l receives information from node j is as follows:
WhereinIt is made of three parts, i.e., local prioriLocal measurement collectionThis
Earth anchor node state collectionConcrete form are as follows:
In the present embodiment step S2, first local zone time axis is updated, based on updated time shaft using distributed
Priori dividing method is to local measurement collection, local anchor node state set, local prior probability and local posterior probability in database
It is updated, specific steps include:
Whether it is the database moment that S21. judgement is current, and the database moment is the newest moment in local zone time axis, if
It is to be transferred to execute step S22, otherwise exits;
S22. local measurement collection and local anchor node state set update: judge renewable time whether the correspondence database moment,
If so, the local measurement collection after the measurement that local measurement concentrates addition new collects and updated, and in local anchor node state
It concentrates and local anchor node state set after new anchor node state set is updated is added, otherwise concentrated in local measurement and interaction is added
Measurement in information collects updated after local measurement collection, and in local anchor node state set be added interactive information in anchor
Node state collection local anchor node state set after being updated;
S23. Prior Fusion: respectively to each node l ' for including in the database of node l, prior variance is found out most
The prior distribution of all nodes found out is stitched together by small neighbor node j, forms joint prior distribution;
S24. priori is divided: the subset based on local measurement collectionLocal prior state setWith
And the subset of local anchor node state setTriple SNCAMST is defined, choosing has the described of smallest partition difference
Triple SNCAMST simultaneously calculates marginal probability;
S25. posteriority updates and priori prediction: updating local posterior probability using the marginal probability calculated, and in database
Last moment predicts local prior probability using updated local posterior probability.
The present embodiment realizes database update based on distributed priori dividing method by above-mentioned, divides difference based on priori
With the continuity of location estimation difference, it can be ensured that the correct choice of the partial status of local priori, compared to it is traditional based on
The method of global information can effectively reduce position disparity, while improve location efficiency.
In the present embodiment, at each database update momentNon- anchor node l all by its database fromIt is updated toThe update of database includes five parts: local zone time axis, local measurement collection, local anchor node state set, local priori,
With local posteriority.The update of local zone time axis is provided first, then again based on updated local zone time axis to other each
Part is updated.Since the computing capability and storage capacity of node are limited,The number at all moment in the past can not be included
According to especially when k becomes very large over time, the present embodiment is usedTo limitIn include the maximum moment
Number,(such as maximal memory constraintMean letter of the most multipotency comprising 3 last time in the database of node l
Breath), i.e., updated local zone time axisIt is sum time axisSubset, wherein sum time axisWhen being old database
Between axisWith the union of new arrival information time axis.
In the present embodiment, local measurement collectionWith local anchor node state setWhen update, if
WithCorresponding (the data moment updated corresponds to the newest moment in local zone time axis), then updated local measurement collection
New measurement collection is just added on the basis of originalLocal anchor node state set updated at this time is just on original basis
It is upper that new anchor node state set is addedIfWithIt is unequal, then updated local measurement collection is just original
On the basis of be added interactive information kind included new measurement collectAt this point, updated local anchor node state set is just
The new anchor node state set that interactive information kind is included is added on the basis of original
Due to each node l ' for including in the database for node l, prior distribution is multiple, i.e. node l
Each neighbours j different node l ' prior informations can be provided, however due to the limitation of information, priori letter that neighbours provide
Breath alternates betwwen good and bad.The present embodiment Prior Fusion is to find out it for each l ' to provide minimum (the i.e. estimation effect of prior variance
It is best) neighbor node j, be denoted asFinally all l ' prior distributions are stitched together, form joint prior distribution.It spells
When connecing/merging, ignore the coupling between different neighbor informations, that is, the prior information for assuming that different j are provided is mutually indepedent, joint point
Cloth is equal to the product that different j provide distribution.
In the present embodiment, step S23 returns through fused prioriSpecific steps are as follows: definition is local first
Prior varianceTo reflect node l ' in local prioriUnder uncertainty, whereinFor node
The state of l ',For s-th of record moment in kth time update, j is the serial number of neighbor node.Local prior varianceIt is smaller, thenWhat is had is uncertain just smaller, and the present embodiment chooses corresponding minimum local prior variance
'sAnd asProbability, according toCarry out the spelling of priori
It connects, whereinTo include local prior variance in databaseAll nodes set.
In the present embodiment, the specific steps of step S24 include: the measuring state three that definition first is constrained with number of states
Tuple SNCAMST.SNCAMST depends on local measurement figure, is based on local measurement collectionLocal measurement figure is defined as:
WhereinFor local measurement figureCorresponding vertex set, ε are local measurement figureCorresponding line set;
Definition status constraint measuring state triple (SNCAMST) on this basis
WhereinWithIt is respectivelyWithSubset.
Estimation difference (estimation gap) is segmentation difference (cut gap)It is continuous
Function, whereinFor segmentation after be used for location algorithm priori,Remaining priori (is not positioned after indicating segmentation
The priori that algorithm uses), the present embodiment can obtain better positioning by choosing the priori with smallest partition difference
Effect.
It enablesIt is for maximum rating numberSNCAMST, be denoted asAll segmentation difference, which will be constituted, to be collected
It closes:
Obtain the SNCAMST of smallest partition difference are as follows:
The SNCAMST of smallest partition difference is calculated using following marginal probability formula on this basis
Wherein,For local prior state set subset, andProbability density function be
In the present embodiment, step S25 specific steps are as follows:
S251. judge whether local a prior set, local anchor node state set are sky, if local a prior set or local anchor node
State set is not sky, is transferred to and executes step S252, is otherwise transferred to and executes step S253;
S252. local posterior probability is updated using Bayesian Method using the marginal probability of calculating, is transferred to and executes step
Rapid S253;
S253. judge whether the current database moment is the last one moment in database, if so, using it is current this
Ground posterior probability predicts local posterior probability using Kolmogorov equation.
If the specific local a prior set of the present embodiment or local anchor node state set are not sky, local posteriority uses Bayesian Method
Then it is updated:
IfIt is notIn the last one moment, above-mentioned updated local posteriority will be used as calculating local elder generation
It testsUse Kolmogorov prediction equation local posterior probability:
Wherein,For local prior state setCorresponding supported collection (the i.e. collection of probability density non-zero
It closes).
In concrete application embodiment, it is based on updated time shaftIt is real by configuring distributed priori partitioning algorithm
Now local measurement collection, local anchor node state set, local priori and local posteriority are updated, distributed priori dividing method is specific
Process is as shown in algorithm 2.
In above-mentioned algorithm 2, databaseAccording to time collectionIn time ascending sequence connect by an element
It successively updates to one element, element in databaseInclude when update local measurement collection and local anchor node state set more
Newly, Prior Fusion, priori segmentation, posteriority update and five parts of priori prediction.
As shown in fig. 6, above-mentioned distribution priori partitioning algorithm realizes local measurement collection, local anchor node state set, local
The detailed process that priori and local posteriority update are as follows:
Step 1: whether judgement is currently the database moment, i.e., meetStep 2 is executed if it is being transferred to,
Otherwise terminate;
Step 2: judging whether renewable time is the correspondence database moment, i.e., whether meetsIf so, executing
Step 3, no to then follow the steps 4;
Step 3: pressingNew measurement collection is added in local data base, and pressesNew anchor node state set is added in local data base, executes step 5;
Step 4: pressingThe measurement in interactive information is added in local data base
Collection, and pressThe new anchor node shape in interactive information is added in local data base
State collection executes step 5;
Step 5: for other nodes l ' stored in each database, taking the smallest local prior variance serial number
Step 6: according toCarry out priori splicing;
Step 7: pressingIt chooses and minimizes segmentation difference
SNCAMST;
Step 8: pressing formulaCalculate marginal probability;
Step 9: local posteriority being updated and predicts prior probability.
Above-mentioned only presently preferred embodiments of the present invention, is not intended to limit the present invention in any form.Although of the invention
It has been disclosed in a preferred embodiment above, however, it is not intended to limit the invention.Therefore, all without departing from technical solution of the present invention
Content, technical spirit any simple modifications, equivalents, and modifications made to the above embodiment, should all fall according to the present invention
In the range of technical solution of the present invention protection.
Claims (10)
1. a kind of unmanned plane cluster cooperation formula localization method based on asynchronous information, which is characterized in that step includes:
S1. the data of memory node locating desired information are used for for the corresponding configuration of the non-anchor node of each in unmanned plane cluster one
Library, the non-anchor node are the unmanned plane node positioned to itself Location-Unknown, needs;
S2. when measuring or receiving anchor node state there are the non-anchor node of target or receive the message of other nodes transmissions,
The corresponding database of the non-anchor node of target is updated, is transferred to after the completion of updating and executes step S3;
S3. the non-anchor node of target according in the corresponding database information generate broadcast message be broadcast to adjacent node into
Row state information updating, and the location information of itself is updated, complete the positioning of unmanned plane node.
2. the unmanned plane cluster cooperation formula localization method according to claim 1 based on asynchronous information, which is characterized in that non-
The data of each moment record include local prior probability, local posterior probability, local measurement in the database of anchor node
Collection and local four partial data of anchor node state set, the local prior probability include that the priori of node relevant to positioning joins
It closes probability density information and constantly updates, the local posterior probability includes that the posteriority joint probability of node relevant to positioning is close
Degree information simultaneously constantly update, the local measurement collection include all anchor node status informations relevant to positioning and constantly choice and
It updating, the local anchor node state set includes all anchor node status informations relevant to positioning and constantly accepts or rejects and update,
The broadcast message includes the local prior probability, the local measurement collection and the local anchor node state set three parts
Information.
3. the unmanned plane cluster cooperation formula localization method according to claim 2 based on asynchronous information, which is characterized in that non-
The database of anchor node specifically:
Wherein,For the corresponding database of first of non-anchor node,For updated database,For database
Renewal time set,It is updated for database kth time;
It willIt is defined asAnd:
Wherein,Exist for databaseThe record at moment,For the set at all record moment;For local
Prior probability,For local posterior probability,For local measurement collection,For local anchor node state
Collection;T is current time,For s-th of record moment in kth time update, wherein k and s is respectively positive integer;
The broadcast message specifically:
Wherein,The broadcast message generated when being updated for first of non-anchor node kth time,Be so thatCollection at the time of establishment.
4. the unmanned plane cluster cooperation formula localization method according to claim 2 or 3 based on asynchronous information, feature exist
In in the step S2, being first updated to local zone time axis, based on updated time shaft using distributed priori segmentation side
Method carries out more local measurement collection, local anchor node state set, local prior probability and local posterior probability in the database
Newly, specific steps include:
S21. whether judgement is currently the database moment, and the database moment is the newest moment in local zone time axis, if
It is to be transferred to execute step S22, otherwise exits;
S22. local measurement collection and local anchor node state set update: judge renewable time whether the correspondence database moment, if
It is the local measurement collection after the measurement that local measurement concentrates addition new collects and updated, and in local anchor node state set
Local anchor node state set is added after new anchor node state set is updated, is otherwise concentrated in local measurement and interactive information is added
In measurement collect updated after local measurement collection, and in local anchor node state set be added interactive information in anchor node
State set local anchor node state set after being updated;
S23. respectively to each node l ' for including in the database of node l, it is the smallest Prior Fusion: to find out prior variance
The prior distribution of all nodes found out is stitched together by neighbor node j, forms joint prior distribution;
S24. priori is divided: the subset based on local measurement collectionLocal prior state setAnd this
The subset of earth anchor node state collectionTriple SNCAMST is defined, the ternary with smallest partition difference is chosen
Group SNCAMST simultaneously calculates marginal probability;
S25. posteriority updates and priori prediction: updating the local posterior probability using the marginal probability calculated, and in number
The local prior probability is predicted using the updated local posterior probability according to the library last moment.
5. the unmanned plane cluster cooperation formula localization method according to claim 4 based on asynchronous information, which is characterized in that institute
It states when being updated to local zone time axis, usesLimit the maximum moment number for including in database, whereinFor maximum
Memory Constrained, i.e., updated local zone time axisIt is sum time axisSubset, the sum time axisIt is old number
According to library time shaftWith the union of new arrival information time axis.
6. the unmanned plane cluster cooperation formula localization method according to claim 4 based on asynchronous information, which is characterized in that institute
The step of stating step S23 specifically includes: defining local prior varianceTo reflect that node l ' is general in local priori
RateUnder uncertainty, whereinFor the state of node l ',For kth time update in s-th of record when
It carves, j is the serial number of neighbor node, from other nodes l ' stored in database, chooses corresponding minimum local prior varianceAsProbability, it is the smallest local prior variance serial number specifically:
WhereinFor the corresponding optimal node ID of l ' of return, i.e., the serial number of the smallest local prior variance,For number
According to including local prior variance in libraryAll nodes set;
Priori splicing is carried out according still further to following formula:
Wherein,For the set of the corresponding optimal node ID composition of all l '.
7. the unmanned plane cluster cooperation formula localization method according to claim 4 based on asynchronous information, which is characterized in that institute
The step of stating step S24 specifically includes: definition is based on local measurement collectionLocal measurement figure are as follows:
WhereinFor local measurement figureCorresponding vertex set, ε are local measurement figureCorresponding line set;
Defining the triple SNCAMST based on the local measurement figure isWhereinWithIt is respectivelyWithSubset,For local prior state collection
It closes,For local prior probability,For local posterior probability,For local measurement collection,
For local anchor node state set;
Divide difference are as follows:
Wherein,For segmentation after prior probability used for positioning,It is general for the priori for being not used in positioning remaining after segmentation
Rate;
It enablesIt is for maximum rating numberThe triple SNCAMST, be denoted asAll segmentation difference is by structure
At set are as follows:
Obtain the SNCAMST of smallest partition difference are as follows:
The SNCAMST of the smallest partition difference is calculated into marginal probability using following formula
Wherein,For local prior state set subset, andProbability density function be
8. the unmanned plane cluster cooperation formula localization method according to claim 4 based on asynchronous information, which is characterized in that institute
State step S25 specific steps are as follows:
S251. judge whether the local a prior set, local anchor node state set are sky, if the local a prior set or described
Earth anchor node state collection is not sky, is transferred to and executes step S252, is otherwise transferred to and executes step S253;
S252. the local posterior probability is updated using Bayesian Method using the marginal probability of calculating, is transferred to and holds
Row step S253;
S253. judge whether the current database moment is the last one moment in database, if so, after using current local
Probability is tested to predict local posterior probability using Kolmogorov equation.
9. the unmanned plane cluster cooperation formula localization method according to claim 8 based on asynchronous information, which is characterized in that institute
State the formula being updated using Bayesian Method to the local posterior probability are as follows:
Wherein,For local prior state set,For marginal probability,It isSon
Collection,For local measurement collection;
The formula that local posterior probability is predicted using Kolmogorov equation are as follows:
Wherein,For local prior state setCorresponding supported collection.
10. a kind of unmanned plane cluster cooperation formula positioning device based on asynchronous information, including computer equipment, which is characterized in that
The computer equipment is programmed to perform the unmanned plane cluster described in any one of claim 1~9 based on asynchronous information
The step of cooperative localization method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910537244.6A CN110312306B (en) | 2019-06-20 | 2019-06-20 | Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910537244.6A CN110312306B (en) | 2019-06-20 | 2019-06-20 | Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110312306A true CN110312306A (en) | 2019-10-08 |
CN110312306B CN110312306B (en) | 2020-10-30 |
Family
ID=68076767
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910537244.6A Active CN110312306B (en) | 2019-06-20 | 2019-06-20 | Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110312306B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113311384A (en) * | 2021-04-29 | 2021-08-27 | 中国人民解放军国防科技大学 | Unilateral bidirectional distance measurement method and device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637040A (en) * | 2012-04-23 | 2012-08-15 | 清华大学 | Unmanned aerial vehicle cluster visual navigation task coordination method and system |
CN102749847A (en) * | 2012-06-26 | 2012-10-24 | 清华大学 | Cooperative landing method for multiple unmanned aerial vehicles |
CN108966120A (en) * | 2018-06-09 | 2018-12-07 | 中国电子科技集团公司第五十四研究所 | A kind of three side localization method of combination and system for dynamic cluster network improvement |
CN109460065A (en) * | 2019-01-12 | 2019-03-12 | 中国人民解放军国防科技大学 | Unmanned aerial vehicle cluster formation characteristic identification method and system based on potential function |
-
2019
- 2019-06-20 CN CN201910537244.6A patent/CN110312306B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637040A (en) * | 2012-04-23 | 2012-08-15 | 清华大学 | Unmanned aerial vehicle cluster visual navigation task coordination method and system |
CN102749847A (en) * | 2012-06-26 | 2012-10-24 | 清华大学 | Cooperative landing method for multiple unmanned aerial vehicles |
CN108966120A (en) * | 2018-06-09 | 2018-12-07 | 中国电子科技集团公司第五十四研究所 | A kind of three side localization method of combination and system for dynamic cluster network improvement |
CN109460065A (en) * | 2019-01-12 | 2019-03-12 | 中国人民解放军国防科技大学 | Unmanned aerial vehicle cluster formation characteristic identification method and system based on potential function |
Non-Patent Citations (2)
Title |
---|
潘瑞鸿,徐胜红: "基于几何特性的多无人机协同导航算法", 《兵器装备工程学报》 * |
韩亮,任章,董希旺,李清东: "多无人机协同控制方法及应用研究", 《导航定位与授时》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113311384A (en) * | 2021-04-29 | 2021-08-27 | 中国人民解放军国防科技大学 | Unilateral bidirectional distance measurement method and device |
CN113311384B (en) * | 2021-04-29 | 2023-11-10 | 中国人民解放军国防科技大学 | Unilateral two-way distance measurement method and device |
Also Published As
Publication number | Publication date |
---|---|
CN110312306B (en) | 2020-10-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1988550B (en) | Distributing realizing method for radio sensor network no-anchor point location | |
CN103648164B (en) | A kind of based on the difference time of advent and the wireless-sensor network distribution type localization method of Gossip algorithm | |
CN103428275B (en) | Indoor moving goal activity route tracking method based on WSN | |
CN102291817B (en) | Group positioning method based on location measurement sample in mobile communication network | |
CN105372628A (en) | Wi-Fi-based indoor positioning navigation method | |
CN104363654B (en) | Wireless sensor network tri-dimensional node positioning method based on Tunneling method | |
CN102685677B (en) | A kind of indoor orientation method and device | |
CN112469117B (en) | Improved DV-hop positioning method for irregular wireless sensor network | |
CN109661030A (en) | Unknown object location algorithm in wireless sensor network based on dynamic grid | |
CN107592671A (en) | A kind of autonomous co-located method of networking multiple agent active variable topological | |
CN108737959A (en) | 5G super-intensive method for switching network based on Static Precise Point Positioning | |
CN108882172A (en) | A kind of prediction technique of the indoor moving track data based on HMM model | |
CN108966120A (en) | A kind of three side localization method of combination and system for dynamic cluster network improvement | |
CN115361051B (en) | Frequency sharing analysis system for large-scale space internet constellation | |
CN109547929A (en) | Distributed sensor node positioning method based on conjugate gradient method | |
CN110312306A (en) | Unmanned aerial vehicle cluster cooperative positioning method and device based on asynchronous information | |
CN114513805A (en) | Wireless modeling method and system | |
CN109064750A (en) | Urban road network traffic estimation method and system | |
CN117031519A (en) | Multi-target three-dimensional co-location method and system | |
CN110662167A (en) | Indoor heterogeneous network cooperative positioning method and system and readable storage medium | |
CN109581280A (en) | The adaptive tuning on-line method, system and device of terminal | |
CN116047410A (en) | UWB indoor pedestrian positioning method based on graph convolution neural network | |
CN115397012A (en) | Realization method of UWB positioning tracking system based on TWR-TDOA estimation and MPGA layout optimization | |
CN108924734A (en) | A kind of three-dimension sensor node positioning method and system | |
CN110856251B (en) | Terminal positioning method in ultra-dense network |
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 |