CN106912010A - Bluetooth assist wireless network alignment system based on recurrent neural networks - Google Patents

Bluetooth assist wireless network alignment system based on recurrent neural networks Download PDF

Info

Publication number
CN106912010A
CN106912010A CN201510970254.0A CN201510970254A CN106912010A CN 106912010 A CN106912010 A CN 106912010A CN 201510970254 A CN201510970254 A CN 201510970254A CN 106912010 A CN106912010 A CN 106912010A
Authority
CN
China
Prior art keywords
mobile phone
neural networks
recurrent neural
bluetooth
alignment system
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201510970254.0A
Other languages
Chinese (zh)
Inventor
吕姗
陈三风
黄焕波
胡涛
郭森
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Information Technology
Original Assignee
Shenzhen Institute of Information Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Information Technology filed Critical Shenzhen Institute of Information Technology
Priority to CN201510970254.0A priority Critical patent/CN106912010A/en
Publication of CN106912010A publication Critical patent/CN106912010A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/023—Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)

Abstract

A kind of bluetooth assist wireless network alignment system based on recurrent neural networks, including multiple mobile phones, these mobile phones possess function of Bluetooth communication；These mobile phone zonings are：Anchor node mobile phone, it is that possess the independent ability for obtaining positional information；With blind node mobile phone, it is that do not possess the independent ability for obtaining positional information；Wherein, each blind node mobile phone is associated with a dynamic neuron, all dynamic neurons constitute a recurrent neural networks by the bluetooth communications link of blind node mobile phone, the positional information that the recurrent neural networks can be provided according to anchor node mobile phone, estimates the positional information of each blind node mobile phone.Blind area can be greatly reduced, the accuracy of positioning is improved.

Description

Bluetooth assist wireless network alignment system based on recurrent neural networks
Technical field
The present invention relates to location technology, more particularly to the location technology realized by mobile phone.
Background technology
For the cell phone equipped with GPS functions, existing is usually that the GPS functions by cell phone in itself come real It is existing.For the cell phone without GPS functions, existing is usually that the signal monitoring of the cell phone is come by mobile network Realize., be present very big blind area in this location technology realized by mobile phone, for example, do not covered in wireless network Place, the cell phone without GPS functions will lose positioning function, even if can to a certain extent infer according to historical data The current location of cell phone, also lacking accuracy can say.It can be seen that, it is necessary in fact to the existing positioning skill realized by mobile phone Art is improved
The content of the invention
The technical problem to be solved in the present invention is to overcome the shortcomings of that above-mentioned prior art is present, and proposes one kind and be based on back Return the bluetooth assist wireless network alignment system of neutral net, blind area can be greatly reduced, improve the accuracy of positioning.
For above-mentioned technical problem, the technical scheme that proposes includes the present invention, proposes a kind of based on recurrent neural networks Bluetooth assist wireless network alignment system, including multiple mobile phones, wherein, these mobile phones possess function of Bluetooth communication；These Mobile phone zoning is：Anchor node mobile phone, it is that possess the independent ability for obtaining positional information；With blind node mobile phone, it is that do not have The standby independent ability for obtaining positional information；Wherein, each blind node mobile phone is associated with a dynamic neuron, all dynamic god A recurrent neural networks are constituted by the bluetooth communications link of blind node mobile phone through unit, the recurrent neural networks can be according to anchor The positional information that node mobile phone is provided, estimates the positional information of each blind node mobile phone.
In certain embodiments, the algorithm for estimating the positional information of each blind node mobile phone is：
Wherein, xiIt is i-th estimated location of blind node, ε is factor of influence, ωijIt is a positive weights, IijIt is a mark Value, it is defined as follows：
Wherein, xiAnd xjRepresent i-th and j-th coordinate of cell phone respectively, R for Bluetooth of mobile phone maximum communication away from From.
In certain embodiments, the recurrent neural networks are a distributed recurrent neural networks, are faced between node only Only directly carry out Bluetooth communication.
In certain embodiments, x in the recurrent neural networksiDynamic variation characteristic be decided by it face node xj, wherein j∈N(i),Face node xjTo xiActive force be:-εIij(xi-xj)。
In certain embodiments, each dynamic neuron can be by the software of corresponding blind node mobile phone and/or hardware reality It is existing.
In certain embodiments, the anchor node mobile phone is the cell phone for possessing GPS functions；The blind node mobile phone is that do not have The cell phone of standby GPS functions.
Compared with prior art, the bluetooth assist wireless network alignment system based on recurrent neural networks of the invention, leads to Cross and dexterously make each blind node mobile phone and associated with a dynamic neuron, all dynamic neurons by blind node mobile phone indigo plant Tooth communication constitutes a recurrent neural networks, and the positional information that the recurrent neural networks can be provided according to anchor node mobile phone is estimated The positional information of each blind node mobile phone is counted out, blind area can be greatly reduced, improve the accuracy of positioning.
Brief description of the drawings
Fig. 1 is the structural representation of bluetooth assist wireless network alignment system of the present invention based on recurrent neural networks.
Fig. 2 is the schematic diagram of the neuron models interacted with blind node.
Fig. 3 is the location simulation signal in the first application scenarios.
Fig. 4 is the positioning precision signal in the first application scenarios.
Fig. 5 is illustrated in second topological structure of the alignment system of application scenarios.
Fig. 6 is illustrated in second location simulation of application scenarios.
Fig. 7 is illustrated in second positioning precision of application scenarios.
Wherein, description of reference numerals is as follows：The blind bluetooth of node mobile phone 103 of 100 alignment system, 101 anchor node mobile phone 102 is led to The letter emulation emulation emulation final value of final value 620 of initial value 320 of link 310.
Specific embodiment
Below in conjunction with accompanying drawing, give the present invention further elaboration.
Referring to Fig. 1, Fig. 1 is that the structure of bluetooth assist wireless network alignment system of the present invention based on recurrent neural networks is shown Meaning.The present invention proposes a kind of bluetooth assist wireless network alignment system 100 based on recurrent neural networks, including multiple mobile phones, Wherein, these mobile phones possess function of Bluetooth communication；These mobile phone zonings are：Anchor node mobile phone 101, it is that possess independently to obtain Obtain the ability of positional information；With blind node mobile phone 102, it is that do not possess the independent ability for obtaining positional information；Wherein, often Individual blind node mobile phone is associated with a dynamic neuron, all dynamic neurons by blind node mobile phone bluetooth communications link 103 constitute a recurrent neural networks, and the positional information that the recurrent neural networks can be provided according to anchor node mobile phone is estimated The positional information of each blind node mobile phone.In the present embodiment, each dynamic neuron can be by corresponding blind node mobile phone 102 Software and/or hardware realize.The anchor node mobile phone 101 is the cell phone for possessing GPS functions；The blind node mobile phone 102 is Do not possess the cell phone of GPS functions.
The positional information of anchor node mobile phone 101 is obtained by GPS.Each bluetooth connection be to one of cell phone about Beam, specific formula is as follows：
(xi-xj)T(xi-xj)≤R2 i∈N(j) (1a)
Wherein, xiAnd xjRepresent i-th and j-th coordinate of cell phone respectively, R for Bluetooth of mobile phone maximum communication away from From.N (j) is j-th Lin Yuji of cell phone, and it includes all cell phone collection for passing through Bluetooth communication with it.B is anchor section Point mobile phone collection,For k-th anchor node mobile phone position information by GPS acquisitions.
Formula (1) does not only wait constraint and equated constraint without explicit object function.Such problem solving is general all It is not unique.In the present invention, Real-time solution is only needed to obtain the feasible solution of formula (1) not all solutions.In this regard, the present invention will A dual regression neutral net is designed to solve this problem.
The solution of formula (1) is converted into the following common optimum problem (2) with apparent objective function.
Also,
N is all mobile phone numbers in the wireless network, ωijIt is the connection weight between i-th and j mobile phone.It is worth One is mentioned that, due to containing max () function in problem (2), so this is a non-smoothing problasm.Problem (2) is to xi Carrying out partial differential can obtain：Or 0 (now and if only if (xi-xj)T(xi-xj)-R2=0, to Regression Neutral net carries out double optimization, finds the feasible solution of problem (2)：
Wherein, xiIt is i-th estimated location of blind node, ε is factor of influence, ωijIt is a positive weights, IijIt is a mark Knowledge value, it is defined as follows：
Recurrent neural network model proposed by the present invention has distributed nature, embodies as follows：
1st, recurrent neural networks (3) are a distributed recurrent neural networks.In this recurrent neural networks, face section Bluetooth communication is only directly carried out between point, it is not necessary to route or other cross-communications.Recurrent nerve proposed by the present invention The distributed nature of network thoroughly reduces the communications burden between node, and real-time is good, is highly suitable for large-scale wireless network The orientation problem of network.
2nd, x in recurrent neural networks (3)iDynamic variation characteristic be decided by it face node xj, wherein j ∈ N (i) faces section Point xjTo xiActive force be:-εIij(xi-xj).This formula similar to physics in " spring particle " model, i.e. particle xi To xjActive force be ε or be 0, now | | xi-xj| |≤R or | | xi-xj||>R.This mechanism ensure that determining for blind node Position problem is only decided by the comprehensive function power of other nodes in bluetooth communications range R.
3rd, the recurrent neural networks are stable.
Recurrent neural networks scheme proposed by the present invention can both have been estimated to obtain each blind section with the mode of dispersed problem (3) The positional information of point, it is also possible to completed with parallel analog circuit scheme.In the present embodiment, using parallel analog circuit To realize.In neutral net (3), each blind mobile node is associated with a dynamic neuron.Using this neuron as Individual department pattern, its as whole neutral net a part.This model interacts with the model for closing on, all of model Realize positioning and solving orientation problem (1) together.It is different from conventional iterative scheme that (common iterative scheme can only use serial side Formula is realized), recurrent neural networks scheme proposed by the present invention can be realized with parallel analog circuit, and can be solved in real time Problem.
Referring to Fig. 2, Fig. 2 is the schematic diagram of the neuron models 200 interacted with blind node 102.Wherein j1,j2,..., jkTo face mobile node, it both can be blind node that node is faced in movement, it is also possible to used as anchor node.The input of neuron models be with The active force faced between mobile node, is output as the estimated location of the blind node.Using this scheme, each blind section can be obtained The rough location information of point.
It is the location simulation signal in the first application scenarios referring to Fig. 3 and Fig. 4, Fig. 3.Fig. 4 is in the first applied field The positioning precision of scape is illustrated.It is that alignment system of the invention is applied to tunnel vehicle location, the wherein unit of abscissa is rice, Point position 310 is emulation initial value, and point position 320 is emulation final value.The abscissa unit of Fig. 4 is：The second of 10-4 powers；Ordinate For：10 5 powers.Simulation accuracy result is 0 from the 5 power rapid drawdowns of 3.9*10 in -4 powers second of 1.2*10.It can be seen that, this hair Bright alignment system has preferable positioning performance, real-time performance and stability.
Referring to Fig. 5, Fig. 6 and Fig. 7, Fig. 5 is illustrated in second topological structure of the alignment system of application scenarios.Fig. 6 is Illustrate in second location simulation of application scenarios.Fig. 7 is illustrated in second positioning precision of application scenarios.It is by the present invention Alignment system be applied to large supermarket's positioning, the unit of wherein abscissa and ordinate is rice, this emulation experiment supermarket Area is 60*60 square metres, and 121 shoppers with bluetooth cellular phone are randomly dispersed in supermarket, in supermarket circumferentially and the center of circle 9 anchor nodes are distributed, the relative coordinate of this 9 anchor nodes is：[0,0],[30,0],[60,0],[60,30],[60,60], [30,60], [0,60], [0,30], [30,30], bluetooth cellular phone works in pattern 2, i.e. maximum transmission distance for 10 meters.Point position 620 is emulation final value.The abscissa unit of Fig. 7 is：The second of 10-5 powers；The multiple that ordinate is is 1, is exactly normally. Simulation result precision is reduced to 0 in 4*10-5 powers second from 1365.7.It can be seen that, alignment system of the invention has preferably positioning Performance, real-time performance and stability.
Compared with prior art, the bluetooth assist wireless network alignment system based on recurrent neural networks of the invention, leads to Cross and dexterously make each blind node mobile phone 102 and associated with a dynamic neuron, all dynamic neurons are by blind node mobile phone 102 Bluetooth communication constitutes a recurrent neural networks, the position that the recurrent neural networks can be provided according to anchor node mobile phone 101 Confidence ceases, and estimates the positional information of each blind node mobile phone 102, can greatly reduce blind area, improves the accuracy of positioning.
The above, only presently preferred embodiments of the present invention, are not intended to limit embodiment of the present invention, and this area is general Logical technical staff's central scope of the invention and spirit, can very easily carry out corresponding flexible or modification, therefore originally The protection domain of invention should be defined by the protection domain required by claims.

Claims (6)

1. a kind of bluetooth assist wireless network alignment system based on recurrent neural networks, including multiple mobile phones, it is characterised in that These mobile phones possess function of Bluetooth communication；These mobile phone zonings are：Anchor node mobile phone, it is that possess the independent position that obtains to believe The ability of breath；With blind node mobile phone, it is that do not possess the independent ability for obtaining positional information；Wherein, each blind node hand Machine is associated with a dynamic neuron, and all dynamic neurons constitute a recurrence by the bluetooth communications link of blind node mobile phone Neutral net, the positional information that the recurrent neural networks can be provided according to anchor node mobile phone, estimates each blind node mobile phone Positional information.
2., according to the bluetooth assist wireless network alignment system based on recurrent neural networks described in claim 1, its feature exists In the algorithm for estimating the positional information of each blind node mobile phone is：
x · i = - ϵ Σ j ∈ N ( i ) ω i j I i j ( x i - x j )
Wherein, xiIt is i-th estimated location of blind node, ε is factor of influence, ωijIt is a positive weights, IijIt is a mark Value, it is defined as follows：
I i j = 1 ( x i - x j ) T ( x i - x j ) - R 2 > 0 0 ( x i - x j ) T ( x i - x j ) - R 2 ≤ 0
Wherein, xiAnd xjI-th and j-th coordinate of cell phone is represented respectively, and R is the maximum communication distance of Bluetooth of mobile phone.
3., according to the bluetooth assist wireless network alignment system based on recurrent neural networks described in claim 2, its feature exists In the recurrent neural networks are a distributed recurrent neural networks, and facing between node only directly carries out Bluetooth communication.
4., according to the bluetooth assist wireless network alignment system based on recurrent neural networks described in claim 2, its feature exists In x in the recurrent neural networksiDynamic variation characteristic be decided by it face node xj, wherein j ∈ N (i) faces node xjTo xi Active force be:-εIij(xi-xj)。
5., according to the bluetooth assist wireless network alignment system based on recurrent neural networks described in claim 1, its feature exists In each dynamic neuron can be realized by the software of corresponding blind node mobile phone and/or hardware.
6., according to the bluetooth assist wireless network alignment system based on recurrent neural networks described in claim 1, its feature exists In the anchor node mobile phone is the cell phone for possessing GPS functions；The blind node mobile phone is the cell phone for not possessing GPS functions.
CN201510970254.0A 2015-12-22 2015-12-22 Bluetooth assist wireless network alignment system based on recurrent neural networks Pending CN106912010A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510970254.0A CN106912010A (en) 2015-12-22 2015-12-22 Bluetooth assist wireless network alignment system based on recurrent neural networks

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510970254.0A CN106912010A (en) 2015-12-22 2015-12-22 Bluetooth assist wireless network alignment system based on recurrent neural networks

Publications (1)

Publication Number Publication Date
CN106912010A true CN106912010A (en) 2017-06-30

Family

ID=59199758

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510970254.0A Pending CN106912010A (en) 2015-12-22 2015-12-22 Bluetooth assist wireless network alignment system based on recurrent neural networks

Country Status (1)

Country Link
CN (1) CN106912010A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102413564A (en) * 2011-11-25 2012-04-11 北京工业大学 Indoor positioning method based on BP neural network and improved centroid algorithm
CN103428663A (en) * 2012-05-25 2013-12-04 深圳信息职业技术学院 Communication method and system realized based on TTS control center
CN103945533A (en) * 2014-05-15 2014-07-23 济南嘉科电子技术有限公司 Big data based wireless real-time position positioning method
CN105163282A (en) * 2015-09-22 2015-12-16 济南东朔微电子有限公司 Indoor positioning system and positioning method based on Bluetooth location fingerprint

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102413564A (en) * 2011-11-25 2012-04-11 北京工业大学 Indoor positioning method based on BP neural network and improved centroid algorithm
CN103428663A (en) * 2012-05-25 2013-12-04 深圳信息职业技术学院 Communication method and system realized based on TTS control center
CN103945533A (en) * 2014-05-15 2014-07-23 济南嘉科电子技术有限公司 Big data based wireless real-time position positioning method
CN105163282A (en) * 2015-09-22 2015-12-16 济南东朔微电子有限公司 Indoor positioning system and positioning method based on Bluetooth location fingerprint

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
SHUAI LI等: "Bluetooth aided mobile phone localization：A nonlinear neural circuit approach", 《ACM TRANSACTIONS ON EMBEDDED COMPUTING SYSTEMS (TECS)》 *
SHUAI LI等: "Neural network based mobile phone localization using Bluetooth connectivity", 《NEURAL COMPUTING AND APPLICATIONS》 *

Similar Documents

Publication Publication Date Title
McNally The four-step model
Wu et al. A context-aware multiarmed bandit incentive mechanism for mobile crowd sensing systems
CN101247650A (en) Node indoor locating method based on wireless signal strength in wireless sensor network
Varma et al. Random forest learning based indoor localization as an IoT service for smart buildings
CN108051779A (en) A kind of positioning node preferred method towards TDOA
CN106793080A (en) It is a kind of based on hotspot can localization method offline
CN107727095A (en) 3D indoor orientation methods based on spectral clustering and weighting reverse transmittance nerve network
CN104125537A (en) CC2530-based multi-mode co-location system and method
CN103885030A (en) Locating method of mobile node in wireless sensor network
Ever et al. Performance evaluation of hybrid disaster recovery framework with D2D communications
CN104978450B (en) A kind of helicopter vibration active control position preferred method
CN104581943A (en) Node locating method for distribution type wireless sensing network
CN105302858A (en) Distributed database system node-spanning check optimization method and system
CN110247719A (en) The playback of 5G time varying channel and emulation mode based on machine learning
CN110221290A (en) Unmanned plane target based on ant group algorithm optimization searches for construction method
CN106960281A (en) A kind of power telecom network risk management system based on neutral net
Roque-Cilia et al. Transport tracking through communication in WDSN for smart cities
CN106131951A (en) RSSI based on equilateral triangle model weights distance-finding method
Lanza et al. Urban traffic forecasting using federated and continual learning
CN106912010A (en) Bluetooth assist wireless network alignment system based on recurrent neural networks
Chin et al. Nonadditive public transit fare pricing under congestion with policy lessons from a case study in Toronto, Ontario, Canada
CN105872987A (en) Method for node location of distributed wireless sensor network based on single mobile robot
Tennekes et al. A Bayesian approach to location estimation of mobile devices from mobile network operator data
Feng et al. Internet of things system of spatial structure sports events health monitoring based on cloud computing
Wu et al. RangingNet: A convolutional deep neural network based ranging model for wireless sensor networks (WSN)

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20170630