CN103581974A - Link quality assessment method and system - Google Patents
Link quality assessment method and system Download PDFInfo
- Publication number
- CN103581974A CN103581974A CN201210575763.XA CN201210575763A CN103581974A CN 103581974 A CN103581974 A CN 103581974A CN 201210575763 A CN201210575763 A CN 201210575763A CN 103581974 A CN103581974 A CN 103581974A
- Authority
- CN
- China
- Prior art keywords
- link
- quality
- prr
- neural network
- attribute
- 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
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a link quality assessment method and system based on a dynamic fuzzy neural network. The link quality assessment method comprises the following steps: a converging type wireless sensor network is established, and a network topology structure is initialized; data packets which are from a sensing node or uploaded by a routing node and received by a Sink node in a wireless mode are subjected to statistics, and then information of the data packets is extracted and link attributes of the data packets are acquired; the link attributes are fuzzily processed to establish a link attribute fuzzy set corresponding to the link attributes; a dynamic fuzzy neural network model is established according to the link attribute fuzzy set to obtain measurement output representing link quality. Compared with the prior art, the link quality assessment method and system improve the learning capability and self-adapting capability of a link quality assessment by means of establishing the dynamic fuzzy neural network, thereby improving accuracy and robustness in the link quality assessment and being suitable for various applications of a large-scale wireless sensor network.
Description
Technical field
The present invention relates to wireless sensor network basic agreement field, particularly a kind of link-quality-evaluating method and system thereof that is applied to wireless sensor network.
Background technology
Wireless sensor network (Wireless Sensor Network, WSN) be a kind of Full-Distributed System, numerous sensor node random placements are in each guarded region, between sensor node, with wireless, from the mode of knitting, form network, what intercommunication mutually was therebetween used is that lower powered radio frequency signal is electromagnetic wave.The various characteristics of radio wave is wave reflection for example, diffraction, and owing to reflecting caused multipath effect etc., add that it transmits the existence of the very complicated and changeable and interference signal of residing environment, make internodal communication quality be subject to serious impact and restriction, thereby cause internodal link to there is randomness and unsteadiness.For the operation that guarantees that network-efficient is stable, the overall performance that reduces this impact being caused by the various factors of link and improve network service, this just make to link-quality assess just become necessary.
In wireless sensor network, link-quality assessment (Link Quality Evaluation, LQE) is subject to many factors: link disturbs, links asymmetric and hardware differences in radio frequency transitional region, radio-frequency channel time variation matter, multipath.1), the finite energy of node conventional wireless Mesh network and the larger challenge of Ad-Hoc network faces are compared in the link-quality assessment of wireless sensor network:, major applications does not provide energy supply and requires long-term work, and how improving energy efficiency is that link-quality is assessed one of main challenge facing; 2), existing wireless sensor network belongs to narrow band communication, inferior link quality estimation will cause being greater than 200% network slowdown, link-quality assessment need to be paid the utmost attention to and reduce network service load; 3), link-quality is subject to node self and external influence factor complexity.
At present, from obtaining link-quality evaluate parameter angle, link-quality-evaluating method can be divided three classes: hardware based link-quality assessment (Hardware-Based LQE), the assessment of the link-quality based on software (Software-Based LQE) and comprehensive link-quality assessment (Hybrid Approach LQE).
Based on hardware link, quality evaluation is the link information that utilizes hardware layer to provide, these information are that hardware circuit directly obtains by level conversion a bit, as RSSI(Received Signal Strength Indication, received signal strength indicator), SNR(Signal-to-Noise Ratio, signal to noise ratio), some is the statistics to collection signal coding by hardware layer, as CCI(Co-Channel Interference, co-channel interference).Their common feature is not need to increase extra calculating, but these information are subject to such environmental effects obvious, and Related Experimental Study shows by hardware based appraisal procedure accuracy not high, this be because:
The bag that 1. only can receive from success, obtain; 2. based on local bit sampling in bag, obtain; 3. only link-quality can coarsely be described qualitatively, scale that cannot accurate expression link-quality.
Based on software link, quality evaluation is that quantity or the probability that is transmitted or retransmitted by receiving node statistics bag in link obtains, and can be divided into two classes: the 1. statistics based on transmission success probability, as bag acceptance rate PRR; 2. based on the quantitative statistics of bag number, as ETX, RNP.By continue to send the broadcast probe bag (Boardcast Active Packet, BAP) of some in WSN, by the quality condition of statistics packet delivery rate PDR reflection link, increased communication overhead and to moment link-quality reflection sensitive not.
Comprehensive link evaluation is that the cheap and sensitive channel information providing by hardware layer can obtain link-quality more accurately by statistics PDR again, and wherein more famous have Four-bit and a F-LQE.Four-bit utilizes channel information as the foundation of link discovery, then choose a small amount of link preferably and carry out PRR calculating and reduced to a certain extent communication overhead, but it does not really fully utilize much information in conjunction with consider other characteristics of link comprehensively.F-LQE has fully utilized accuracy, stability, the asymmetry that the many measuring methods such as RSSI, PRR can more comprehensively reflect link-quality, but further do not point out the contribution degree to link-quality between each attribute, the mean value of each attribute can not adapt to the situation that wireless sensor network is complicated and changeable.
In sum, link quality method is being obtained in current research, reduces measure error, and the aspects such as raising method efficiency have made some progress, 1), research concentrates on the raising of single performance aspect mainly have the following disadvantages:, also do not relate to and in each performance, seek a dynamic equilibrium; 2), in extensive WSN application, the evaluate parameter obtaining by periodic broadcast Beacons can not reflect moment link-quality, exists energy consumption high and easily cause the drawbacks such as network congestion; 3), attempt under the scene limiting, to obtain link-quality by accurate model, cannot avoid the factor that affect the complicated and variation of link-quality, the adaptive ability of active link quality evaluation mechanism to different scenes not yet proposed.
Summary of the invention
The object of the present invention is to provide a kind of link-quality-evaluating method and system thereof based on dynamic fuzzy neural network, for solving, prior art complexity is high, accuracy is low, a little less than adaptive stress or the problem such as algorithm complex height.
For addressing the above problem and other problems, the present invention provides a kind of link-quality-evaluating method based on dynamic fuzzy neural network on the one hand, is applied in wireless sensor network; Described link-quality-evaluating method comprises: set up convergent type wireless sensor network, and initialization network topology structure; In described convergent type wireless sensor network, comprising: as at least one Sink node of gateway node, the one or more sensing nodes and the routing node that are connected with described Sink node; The packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives, to described packet information extraction and obtain link attribute; Described link attribute is carried out to Fuzzy processing to set up the link attribute fuzzy set corresponding with described link attribute; According to described link attribute fuzzy set, set up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.
Alternatively, described link-quality-evaluating method also comprises the step of schedule link quality detection bag, comprising: not destroying on the packet content basis of application layer submission, packet is encapsulated, add link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.
Alternatively, in to described packet information extraction, described information comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.
Alternatively, described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ; The described link attribute that obtains comprises: described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data: WMEWMA=α WMEWMA+ (1-α) PRR '
iα ∈ [0,1], wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights; The difference of line link and reverse link: ASL (w) before described link asymmetry represents in tested link=| PRR
up-PRR
down|, wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate; Described link stability is used for reflecting link change level: SF
n=COV (PRR0, PRR1 ... PRRn) wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, and COV represents the coefficient of variation; Described channel quality CQ represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
Alternatively, described link attribute being carried out to Fuzzy processing comprises to set up the link attribute fuzzy set corresponding with described link attribute: for packet delivery rate, packet delivery rate finite aggregate PRR is:
wherein, H=is high to be paid, and pays the low payment of L=, μ in M=
prrfor membership function; For link asymmetry, link asymmetry finite aggregate is:
wherein, N=is symmetrical, and A=is asymmetric, μ
aslfor membership function; For link stability, link stability finite aggregate is:
wherein, S=is stable, V=mutation, μ
sffor membership function; For channel quality CQ, channel quality CQ finite aggregate is:
wherein, G=is strong, and a little less than W=, C=is without, μ
cqfor membership function.
Alternatively, described dynamic fuzzy neural network model comprises: link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each; Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process; General several layers of T-, for reflecting the fuzzy rule of neural net; Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link; Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.
Alternatively, in described dynamic fuzzy neural network model: the regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k; According to TSK model, have:
k=1,2,3 ..., u; Output link quality is:
wherein, LQE represents link-quality assessed value.
The present invention provides a kind of link-quality evaluating system based on dynamic fuzzy neural network on the other hand, is applied in wireless sensor network; Described link-quality evaluating system comprises: wireless sensor network initialization unit, for setting up convergent type wireless sensor network initialization network topology structure; In described convergent type wireless sensor network, comprising: as at least one Sink node of gateway node, the one or more sensing nodes and the routing node that are connected with described Sink node; Link attribute acquiring unit, for the packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives and to described packet information extraction and obtain link attribute; Fuzzy processing unit, for carrying out Fuzzy processing to set up the link attribute fuzzy set corresponding with described link attribute to described link attribute; Fuzzy neural network model is set up unit, for according to described link attribute fuzzy set, sets up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.
Alternatively, described link-quality evaluating system also comprises data packet dispatching unit, for: the packet content basis not destroying application layer submission, packet is encapsulated, add link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.
Alternatively, described convergent type wireless sensor network adopts the media access control protocol based on IEEE802.15.4.
Alternatively, the information of described packet being extracted comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.
Alternatively, described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ; The described link attribute that obtains comprises: described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data: WMEWMA=α WMEWMA+ (1-α) PRR '
i, α ∈ [0,1], wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights; The difference of line link and reverse link: ASL (w) before described link asymmetry represents in tested link=| PRR
up-PRR
down|, wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate; Described link stability is used for reflecting link change level: SF
n=COV (PRR0, PRR1 ... PRRn), wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, the COV coefficient of variation; Described channel quality represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
Alternatively, described link attribute being carried out to Fuzzy processing comprises to set up the link attribute fuzzy set corresponding with described link attribute: for packet delivery rate, packet delivery rate finite aggregate PRR is:
wherein, H=is high to be paid, and pays the low payment of L=, μ in M=
prrfor membership function; For link asymmetry, link asymmetry finite aggregate is:
wherein, N=is symmetrical, and A=is asymmetric, μ
aslfor membership function; For link stability, link stability finite aggregate is:
wherein, S=is stable, V=mutation, μ
sffor membership function; For channel quality, channel quality finite aggregate is:
wherein, G=is strong, and a little less than W=, C=is without, μ
cqfor membership function.
Alternatively, described dynamic fuzzy neural network model comprises: link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each; Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process; General several layers of T-, for reflecting the fuzzy rule of neural net; Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link; Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.
Alternatively, in described dynamic fuzzy neural network model: the regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k; According to TSK model, have:
k=1,2,3 ..., u, output link quality is:
wherein, LQE represents link-quality assessed value.
Link-quality-evaluating method and system thereof based on dynamic fuzzy neural network provided by the invention, has following effect:
1, by the packet to wireless receiving analyze, information extraction obtain multiple link attribute, adopt multiple link attribute to carry out link-quality assessment, to improve accuracy, the robustness of link-quality assessment, to be applicable to the multiple application of massive wireless sensor.
2, described link attribute is carried out to Fuzzy processing and obtain link attribute fuzzy set, and set up dynamic fuzzy neural network model according to link attribute fuzzy set, link-quality assessment based on fuzzy logic neural network, solve in link quality evaluation and cannot with regard to multilink attribute, set up the puzzlement of clear and definite expert system, reach the learning ability, the adaptive ability that improve link-quality assessment.
Accompanying drawing explanation
The block architecture diagram of the link-quality evaluating system based on dynamic fuzzy neural network that Fig. 1 has been the present invention.
Fig. 2 is the applied convergent type wireless sensor network of the present invention network diagram in one embodiment.
Fig. 3 is the schematic flow sheet of the link-quality-evaluating method based on dynamic fuzzy neural network.
Fig. 4 is the dynamic fuzzy neural network model set up of the present invention block architecture diagram in one embodiment.
Embodiment
The present inventor finds: in existing link-quality assessment technology, still have that complexity is high, accuracy is low, adaptive stress weak or the problem such as algorithm complex height.
Therefore, the present inventor improves prior art, a kind of link-quality-evaluating method and system thereof based on dynamic fuzzy neural network proposed, by the packet to wireless receiving in wireless sensor network analyze, information extraction obtain multiple link attribute, described link attribute is carried out to Fuzzy processing and obtain link attribute fuzzy set, and set up dynamic fuzzy neural network model according to link attribute fuzzy set, for solving, prior art complexity is high, accuracy is low, a little less than adaptive stress or the problem such as algorithm complex height.
Below, by specific instantiation explanation embodiments of the present invention, those skilled in the art can understand other advantages of the present invention and effect easily by the disclosed content of this specification.The present invention can also be implemented or be applied by other different embodiment, and the every details in this specification also can be based on different viewpoints and application, carries out various modifications or change not deviating under spirit of the present invention.
It should be noted that, the diagram providing in the present embodiment only illustrates basic conception of the present invention in a schematic way, satisfy and only show with assembly relevant in the present invention in graphic but not component count, shape and size drafting while implementing according to reality, during its actual enforcement, kenel, quantity and the ratio of each assembly can be a kind of random change, and its assembly layout kenel also may be more complicated.
One aspect of the present invention provides a kind of link-quality evaluating system based on dynamic fuzzy neural network.Refer to Fig. 1, shown the block architecture diagram of the link-quality evaluating system based on dynamic fuzzy neural network.As shown in Figure 1, described link-quality evaluating system comprises: wireless sensor network initialization unit 101, data packet dispatching unit 103, link attribute acquiring unit 105, Fuzzy processing unit 107 and fuzzy neural network model are set up unit 109.
Wireless sensor network initialization unit 101, for setting up convergent type wireless sensor network initialization network topology structure.In the present embodiment, as shown in Figure 2, in described convergent type wireless sensor network, comprise: at least one Sink node 20, the one or more sensing node 22 being connected with Sink node 20 and routing node 21, each node has a father node object converging as information and the source of ordering distribution.Sensing node 22 is laid in each sensing place, for obtain relevant data by sensing, and described data are converted to packet upload.Routing node 21, between sensing node 22 and Sink node 20, is realized routing function, for the packet of sensing node 22 sensings is turned and reaches Sink node 20.Sink node 20 for: collect the packet of wireless sensor network sensing node 22 sensings or via routing node 21, turn the packet of biography, and carry out exchanges data as the centre data server of gateway node and far-end.Especially, in the present invention, described convergent type wireless sensor network adopts the media access control protocol based on IEEE802.15.4.
Data packet dispatching unit 103, for packet content basis not destroying application layer and submitting to, encapsulates packet, adds link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.Especially, due to poly-type wireless sensor network employing of the present invention is the media access control protocol based on IEEE802.15.4, therefore, can adopt and link-quality is assessed to detection information join in normal data flow, reduce the quantity of broadcast probe beacon (Beacons), reduce because link-quality is assessed the burden of bringing to wireless sensor network.Because data flow is to send with clean culture form radio frequency, in wireless sensor network, node M AC layer filters unicast packet, the TinyOS operating system of the Crossbow-Telosb node simultaneously adopting based on increasing income, making to upgrade mac-layer protocol becomes possibility.
Link attribute acquiring unit 105 is for the packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives and to described packet information extraction and obtain link attribute.In the present embodiment, link attribute acquiring unit 105 will extract the information of unicast data stream and broadcast probe bag simultaneously.The described information of extracting comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.Described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ; The described link attribute that obtains comprises: described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data: WMEWMA=α WMEWMA+ (1-α) PRR '
i, α ∈ [0,1], wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights; The difference of line link and reverse link: ASL (w) before described link asymmetry represents in tested link=| PRR
up-PRR
down|, wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate; Described link stability is used for reflecting link change level: SF
n=COV (PRR0, PRR1 ... PRRn), wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, and COV represents the coefficient of variation; Described channel quality CQ represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
Fuzzy neural network model is set up unit 109 for according to described link attribute fuzzy set, sets up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.In the present embodiment, described dynamic fuzzy neural network model (as shown in Figure 4) comprising: link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each; Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process; General several layers of T-, for reflecting the fuzzy rule of neural net; Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link; Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.In described dynamic fuzzy neural network model: the regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k; According to TSK model, have:
k=1,2,3 ..., u; Output link quality is:
wherein, LQE represents link-quality assessed value.
The embodiment of the present invention provides a kind of link-quality-evaluating method based on dynamic fuzzy neural network on the other hand.As shown in Figure 3, the schematic flow sheet that has shown the link-quality-evaluating method based on dynamic fuzzy neural network.
As shown in Figure 3, described link-quality-evaluating method comprises the steps:
Step S301, sets up convergent type wireless sensor network, and initialization network topology structure.In the present embodiment, as shown in Figure 2, in described convergent type wireless sensor network, comprise: at least one Sink node 20, the one or more sensing node 22 being connected with Sink node 20 and routing node 21, each node has a father node object converging as information and the source of ordering distribution.Sensing node 22 is laid in each sensing place, for obtain relevant data by sensing, and described data are converted to packet upload.Routing node 21, between sensing node 22 and Sink node 20, is realized routing function, for the packet of sensing node 22 sensings is turned and reaches Sink node 20.Sink node 20 for: collect the packet of wireless sensor network sensing node 22 sensings or via routing node 21, turn the packet of biography, and carry out exchanges data as the centre data server of gateway node and far-end.Especially, in the present invention, described convergent type wireless sensor network adopts the media access control protocol based on IEEE802.15.4.
Step S303, schedule link quality detection bag.In the present embodiment, not destroying on the packet content basis of application layer submission, packet is encapsulated, add link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.Especially, due to poly-type wireless sensor network employing of the present invention is the media access control protocol based on IEEE802.15.4, therefore, can adopt and link-quality is assessed to detection information join in normal data flow, reduce the quantity of broadcast probe beacon (Beacons), reduce because link-quality is assessed the burden of bringing to wireless sensor network.Because data flow is to send with clean culture form radio frequency, in wireless sensor network, node M AC layer filters unicast packet, the TinyOS operating system of the Crossbow-Telosb node simultaneously adopting based on increasing income, making to upgrade mac-layer protocol becomes possibility.
Step S305, the packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives, to described packet information extraction and obtain link attribute.In the present embodiment, extract the information of unicast data stream and broadcast probe bag simultaneously.The described information of extracting comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.Described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ; The described link attribute that obtains comprises: described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data: WMEWMA=α WMEWMA+ (1-α) PRR '
i, α ∈ [0,1], wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights; The difference of line link and reverse link: ASL (w) before described link asymmetry represents in tested link=| PRR
up-PRR
down|, wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate; Described link stability is used for reflecting link change level: SF
n=COV (PRR0, PRR1 ... PRRn), wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, and COV represents the coefficient of variation; Described channel quality CQ represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
Step S307, carries out Fuzzy processing to set up the link attribute fuzzy set corresponding with described link attribute to described link attribute.In the present embodiment, described link attribute being carried out to Fuzzy processing comprises to set up the link attribute fuzzy set corresponding with described link attribute: for packet delivery rate, packet delivery rate finite aggregate PRR is:
wherein, H=is high to be paid, and pays the low payment of L=, μ in M=
prrfor membership function; For link asymmetry, link asymmetry finite aggregate is:
wherein, N=is symmetrical, and A=is asymmetric, μ
aslfor membership function; For link stability, link stability finite aggregate is:
wherein, S=is stable, V=mutation, μ
sffor membership function; For channel quality CQ, channel quality CQ finite aggregate is:
wherein, G=is strong, and a little less than W=, C=is without, μ
cqfor membership function.
Step S309, according to described link attribute fuzzy set, sets up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.In the present embodiment, described dynamic fuzzy neural network model comprises: link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each; Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process; General several layers of T-, for reflecting the fuzzy rule of neural net; Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link; Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.In described dynamic fuzzy neural network model: the regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k; According to TSK model, have:
k=1,2,3 ..., u; Output link quality is:
wherein, LQE represents link-quality assessed value.
Link-quality-evaluating method and system thereof based on dynamic fuzzy neural network provided by the invention, has following effect:
1, by the packet to wireless receiving analyze, information extraction obtain multiple link attribute, adopt multiple link attribute to carry out link-quality assessment, to improve accuracy, the robustness of link-quality assessment, to be applicable to the multiple application of massive wireless sensor.
2, described link attribute is carried out to Fuzzy processing and obtain link attribute fuzzy set, and set up dynamic fuzzy neural network model according to link attribute fuzzy set, link-quality assessment based on fuzzy logic neural network, solve in link quality evaluation and cannot with regard to multilink attribute, set up the puzzlement of clear and definite expert system, reach the learning ability, the adaptive ability that improve link-quality assessment.
Above-described embodiment just lists expressivity principle of the present invention and effect is described, but not for limiting the present invention.Any person skilled in the art person all can without departing from the spirit and scope of the present invention, modify to above-described embodiment.Therefore, the scope of the present invention, should be as listed in claims.
Claims (15)
1. the link-quality-evaluating method based on dynamic fuzzy neural network, is applied to, in wireless sensor network, it is characterized in that, described link-quality-evaluating method comprises:
Set up convergent type wireless sensor network, and initialization network topology structure; In described convergent type wireless sensor network, comprising: as at least one Sink node of gateway node, the one or more sensing nodes and the routing node that are connected with described Sink node;
The packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives, to described packet information extraction and obtain link attribute;
Described link attribute is carried out to Fuzzy processing to set up the link attribute fuzzy set corresponding with described link attribute;
According to described link attribute fuzzy set, set up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.
2. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 1, it is characterized in that, the step that also comprises schedule link quality detection bag, comprise: do not destroying on the packet content basis of application layer submission, packet is encapsulated, add link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.
3. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 1, it is characterized in that, in to described packet information extraction, described information comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.
4. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 3, is characterized in that:
Described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ;
The described link attribute that obtains comprises:
Described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data:
WMEWMA=α·WMEWMA+(1-α)·PRR′
i,α∈[0,1]
Wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights;
The difference of line link and reverse link before described link asymmetry represents in tested link:
ASL(w)=PRR
up-PRR
down|
Wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate;
Described link stability is used for reflecting link change level:
SF
n=COV(PRR0,PRR1,……PRRn)
Wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, and COV represents the coefficient of variation;
Described channel quality CQ represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
5. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 4, is characterized in that, described link attribute is carried out to Fuzzy processing and to set up the link attribute fuzzy set corresponding with described link attribute, comprise:
For packet delivery rate, packet delivery rate finite aggregate PRR is:
Wherein, H=is high to be paid, and pays the low payment of L=, μ in M=
prrfor membership function;
For link asymmetry, link asymmetry finite aggregate is:
Wherein, N=is symmetrical, and A=is asymmetric, μ
aslfor membership function;
For link stability, link stability finite aggregate is:
Wherein, S=is stable, V=mutation, μ
sffor membership function;
For channel quality CQ, channel quality CQ finite aggregate is:
Wherein, G=is strong, and a little less than W=, C=is without, μ
cqfor membership function.
6. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 5, is characterized in that, described dynamic fuzzy neural network model comprises:
Link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each;
Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process;
General several layers of T-, for reflecting the fuzzy rule of neural net;
Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link;
Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.
7. the link-quality-evaluating method based on dynamic fuzzy neural network as claimed in claim 6, is characterized in that, in described dynamic fuzzy neural network model:
The regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k;
According to TSK model, have:
Output link quality is:
Wherein, LQE represents link-quality assessed value.
8. the link-quality evaluating system based on dynamic fuzzy neural network, is applied to, in wireless sensor network, it is characterized in that, described link-quality evaluating system comprises:
Wireless sensor network initialization unit, for setting up convergent type wireless sensor network initialization network topology structure; In described convergent type wireless sensor network, comprising: as at least one Sink node of gateway node, the one or more sensing nodes and the routing node that are connected with described Sink node;
Link attribute acquiring unit, for the packet of uploading from described sensing node or described routing node of adding up that described Sink node wireless receives and to described packet information extraction and obtain link attribute;
Fuzzy processing unit, for carrying out Fuzzy processing to set up the link attribute fuzzy set corresponding with described link attribute to described link attribute;
Fuzzy neural network model is set up unit, for according to described link attribute fuzzy set, sets up dynamic fuzzy neural network model, to obtain the tolerance output that represents link-quality.
9. the link-quality evaluating system based on dynamic fuzzy neural network as claimed in claim 8, it is characterized in that, also comprise data packet dispatching unit, be used for: on the packet content basis of not destroying application layer submission, packet is encapsulated, add link evaluation detection information, evaluates calculation, evaluation type; According to link-quality assessment real-time parameter, when link idle, generate and insert detection packet.
10. the link-quality evaluating system based on dynamic fuzzy neural network as claimed in claim 8, is characterized in that, described convergent type wireless sensor network adopts the media access control protocol based on IEEE802.15.4.
The 11. link-quality evaluating systems based on dynamic fuzzy neural network as claimed in claim 8, it is characterized in that, the information that described packet is extracted comprises: received signal strength indicator RSSI, link-quality indication LQI, source data packet address and clean culture/broadcast nature.
The 12. link-quality evaluating systems based on dynamic fuzzy neural network as claimed in claim 11, is characterized in that:
Described link attribute comprises: packet delivery rate, link asymmetry, link stability and channel quality CQ;
The described link attribute that obtains comprises:
Described packet delivery rate is to be bundled into merit acceptance rate and to be utilized WMEWMA algorithm process to obtain by measurement data:
WMEWMA=α·WMEWMA+(1-α)·PRR′
i,α∈[0,1]
Wherein, PRR '
ithe current measurement value that represents described packet success acceptance rate, α represents weights;
The difference of line link and reverse link before described link asymmetry represents in tested link:
ASl(w)=PRR
up-PRR
down|
Wherein, ASL represents link asymmetry value, PRR
upline link bag acceptance rate before representing, PRR
downrepresent reverse link bag acceptance rate;
Described link stability is used for reflecting link change level:
SF
n=COV(PRR0,PRR1,……PRRn)
Wherein, SF represents stability, PRR0, and PRR1 ... PRRn represents PRR sequence, and COV represents the coefficient of variation;
Described channel quality CQ represents the radiofrequency signal identification definition of link, by obtaining signal to noise ratio snr or the link-quality indication LQI of radio frequency, obtains.
The 13. link-quality evaluating systems based on dynamic fuzzy neural network as claimed in claim 12, is characterized in that, described link attribute is carried out to Fuzzy processing and to set up the link attribute fuzzy set corresponding with described link attribute, comprise:
For packet delivery rate, packet delivery rate finite aggregate PRR is:
Wherein, H=is high to be paid, and pays the low payment of L=, μ in M=
prrfor membership function;
For link asymmetry, link asymmetry finite aggregate is:
Wherein, N=is symmetrical, and A=is asymmetric, μ
aslfor membership function;
For link stability, link stability finite aggregate is:
Wherein, S=is stable, V=mutation, μ
sffor membership function;
For channel quality CQ, channel quality CQ finite aggregate is:
Wherein, G=is strong, and a little less than W=, C=is without, μ
cqfor membership function.
The 14. link-quality evaluating systems based on dynamic fuzzy neural network as claimed in claim 13, is characterized in that, described dynamic fuzzy neural network model comprises:
Link attribute measured value input layer, for receiving described in each link attribute and adopting corresponding filtering algorithm according to the different characteristic of link attribute described in each;
Membership function layer, for reflecting the corresponding membership function of each input variable fuzzification process;
General several layers of T-, for reflecting the fuzzy rule of neural net;
Normalization layer, for weighing each regular result in conjunction with and select activation primitive to carry out normalizing output by link;
Link-quality output layer, for exporting the stack of all signals, represents the tolerance of link-quality.
The 15. link-quality evaluating systems based on dynamic fuzzy neural network as claimed in claim 14, is characterized in that, in described dynamic fuzzy neural network model:
The regular output signal that described T-is general several layers is
k is number of regulation, and the link power of k rule is ω
k;
According to TSK model, have:
Output link quality is:
Wherein, LQE represents link-quality assessed value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210575763.XA CN103581974B (en) | 2012-12-26 | 2012-12-26 | Link-quality-evaluating method and system thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210575763.XA CN103581974B (en) | 2012-12-26 | 2012-12-26 | Link-quality-evaluating method and system thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103581974A true CN103581974A (en) | 2014-02-12 |
CN103581974B CN103581974B (en) | 2016-09-07 |
Family
ID=50052685
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210575763.XA Active CN103581974B (en) | 2012-12-26 | 2012-12-26 | Link-quality-evaluating method and system thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103581974B (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106714262A (en) * | 2017-01-09 | 2017-05-24 | 长春工业大学 | Adaptive fuzzy neural topology control algorithm for wireless sensor networks |
WO2017133433A1 (en) * | 2016-02-04 | 2017-08-10 | Jrd Communication Inc. | Estimating success rate of direct data transmissions between mobile devices in wireless network |
CN107302768A (en) * | 2017-07-31 | 2017-10-27 | 南昌航空大学 | A kind of method using the learning machine comprehensive assessment link-quality that transfinites |
CN108966270A (en) * | 2018-07-19 | 2018-12-07 | 南昌航空大学 | Wireless sensor network link-quality-evaluating method |
CN110087206A (en) * | 2019-04-26 | 2019-08-02 | 南昌航空大学 | Using the method for generalized regression nerve networks assessing link quality |
CN110958626A (en) * | 2019-12-06 | 2020-04-03 | 紫光云(南京)数字技术有限公司 | Link selection method based on shortest response time |
CN111526556A (en) * | 2020-04-29 | 2020-08-11 | 黄东 | DTN (delay tolerant network) routing optimization method based on neural network |
CN111935747A (en) * | 2020-08-17 | 2020-11-13 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
CN112187946A (en) * | 2020-09-30 | 2021-01-05 | 国网河北省电力有限公司电力科学研究院 | Internet of things sensing equipment evaluation system and method |
CN117041250A (en) * | 2023-07-17 | 2023-11-10 | 中国人民解放军军事科学院系统工程研究院 | Wireless communication link quality assessment method for intelligent command control |
CN117914829A (en) * | 2023-12-01 | 2024-04-19 | 广州市弘宇科技有限公司 | Intelligent agent fusion communication architecture and method based on neural network model |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101867960A (en) * | 2010-06-08 | 2010-10-20 | 江苏大学 | Comprehensive evaluation method for wireless sensor network performance |
CN101957602A (en) * | 2009-07-15 | 2011-01-26 | 河南天擎机电技术有限公司 | Method and system thereof for monitoring and controlling environments of public place based on Zigbee |
CN102497641A (en) * | 2011-12-20 | 2012-06-13 | 山东大学 | Method for improving communication quality of wireless sensor network |
-
2012
- 2012-12-26 CN CN201210575763.XA patent/CN103581974B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101957602A (en) * | 2009-07-15 | 2011-01-26 | 河南天擎机电技术有限公司 | Method and system thereof for monitoring and controlling environments of public place based on Zigbee |
CN101867960A (en) * | 2010-06-08 | 2010-10-20 | 江苏大学 | Comprehensive evaluation method for wireless sensor network performance |
CN102497641A (en) * | 2011-12-20 | 2012-06-13 | 山东大学 | Method for improving communication quality of wireless sensor network |
Non-Patent Citations (1)
Title |
---|
刘琳岚等: ""一种基于BP神经网络的WSNs链路质量预测方法"", 《计算机研究与发展》, vol. 48, no. 2, 31 December 2011 (2011-12-31) * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2017133433A1 (en) * | 2016-02-04 | 2017-08-10 | Jrd Communication Inc. | Estimating success rate of direct data transmissions between mobile devices in wireless network |
CN109792616A (en) * | 2016-02-04 | 2019-05-21 | 捷开通讯(深圳)有限公司 | Direct message transmission success rate assessment between mobile equipment in wireless network |
CN109792616B (en) * | 2016-02-04 | 2022-05-06 | 捷开通讯(深圳)有限公司 | Direct message transmission success rate assessment between mobile devices in wireless networks |
CN106714262A (en) * | 2017-01-09 | 2017-05-24 | 长春工业大学 | Adaptive fuzzy neural topology control algorithm for wireless sensor networks |
CN106714262B (en) * | 2017-01-09 | 2020-04-07 | 长春工业大学 | Wireless sensor network topology control method based on self-adaptive fuzzy control |
CN107302768A (en) * | 2017-07-31 | 2017-10-27 | 南昌航空大学 | A kind of method using the learning machine comprehensive assessment link-quality that transfinites |
CN107302768B (en) * | 2017-07-31 | 2019-10-15 | 南昌航空大学 | A method of using the learning machine comprehensive assessment link-quality that transfinites |
CN108966270B (en) * | 2018-07-19 | 2021-08-10 | 南昌航空大学 | Wireless sensor network link quality evaluation method |
CN108966270A (en) * | 2018-07-19 | 2018-12-07 | 南昌航空大学 | Wireless sensor network link-quality-evaluating method |
CN110087206A (en) * | 2019-04-26 | 2019-08-02 | 南昌航空大学 | Using the method for generalized regression nerve networks assessing link quality |
CN110087206B (en) * | 2019-04-26 | 2021-12-21 | 南昌航空大学 | Method for evaluating link quality by adopting generalized regression neural network |
CN110958626A (en) * | 2019-12-06 | 2020-04-03 | 紫光云(南京)数字技术有限公司 | Link selection method based on shortest response time |
CN111526556A (en) * | 2020-04-29 | 2020-08-11 | 黄东 | DTN (delay tolerant network) routing optimization method based on neural network |
CN111935747A (en) * | 2020-08-17 | 2020-11-13 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
CN111935747B (en) * | 2020-08-17 | 2021-04-27 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
CN112187946A (en) * | 2020-09-30 | 2021-01-05 | 国网河北省电力有限公司电力科学研究院 | Internet of things sensing equipment evaluation system and method |
CN112187946B (en) * | 2020-09-30 | 2024-01-19 | 国网河北省电力有限公司电力科学研究院 | System and method for evaluating sensing equipment of Internet of things |
CN117041250A (en) * | 2023-07-17 | 2023-11-10 | 中国人民解放军军事科学院系统工程研究院 | Wireless communication link quality assessment method for intelligent command control |
CN117041250B (en) * | 2023-07-17 | 2024-04-19 | 中国人民解放军军事科学院系统工程研究院 | Wireless communication link quality assessment method for intelligent command control |
CN117914829A (en) * | 2023-12-01 | 2024-04-19 | 广州市弘宇科技有限公司 | Intelligent agent fusion communication architecture and method based on neural network model |
Also Published As
Publication number | Publication date |
---|---|
CN103581974B (en) | 2016-09-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103581974A (en) | Link quality assessment method and system | |
Iyer et al. | Performance analysis of wireless mesh routing protocols for smart utility networks | |
Liu et al. | Deeplora: Learning accurate path loss model for long distance links in lpwan | |
CN104361765B (en) | Parking space detecting system based on magnetoresistive sensor and ZigBee | |
CN108064065A (en) | A kind of long-distance water quality monitoring system based on wireless sensor network | |
CN102938742B (en) | Communication bandwidth forecasting method and device based on power business requirements | |
Xue et al. | RVFL-LQP: RVFL-based link quality prediction of wireless sensor networks in smart grid | |
Luo et al. | Link quality estimation method for wireless sensor networks based on stacked autoencoder | |
CN107302768B (en) | A method of using the learning machine comprehensive assessment link-quality that transfinites | |
JP6136830B2 (en) | Interference source position estimation apparatus, position estimation method, and program | |
CN106843100A (en) | Substation secondary device running status level determination method and device | |
CN107332631A (en) | A kind of method of use multi-attribute group decision making theoretical evaluation link-quality assessment models | |
CN107333294A (en) | A kind of combination AdaBoost and SVMs link quality prediction method | |
Nhan et al. | Improving the performance of mobile data collecting systems for electricity meter reading using wireless sensor network | |
CN103873171A (en) | Cooperative spectrum sensing decision fusion method based on multi-user cooperation | |
CN116761194A (en) | Police affair cooperative communication optimization system and method in wireless communication network | |
CN113220751A (en) | Metering system and evaluation method for multi-source data state quantity | |
CN103139804A (en) | Energy-saving transmission self-adaption recursive least squares (RLS) distributed-type detection method of wireless sensor network | |
Sun et al. | Intelligent spacing selection model under energy-saving constraints for the selection of communication nodes in the Internet of Things | |
CN103428724B (en) | Spectrum resource cooperation cut-in method based on geographical location information and system | |
Cao et al. | Automatic meter reading system based on wireless mesh networks and SOPC technology | |
CN109587070A (en) | There is the data assemblage method of secret protection and load balancing simultaneously in smart grid | |
CN105357707A (en) | Indoor distributed mobile communication signal cover prediction method based on Kriging interpolation algorithm | |
CN113453148A (en) | Indoor position fingerprint positioning method combining deep learning and weighted K-neighbor algorithm | |
CN109143407A (en) | The data of automatic weather station unify collection system and its method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20230428 Address after: No. 18 Xiaoqi Village, Lianwei Town, Xinjian District, Nanchang City, Jiangxi Province, 330000 Patentee after: Jiangxi Pogu Agricultural Technology Co.,Ltd. Address before: 200433 No. 335, National Road, Shanghai, Yangpu District Patentee before: AVCON INFORMATION TECHNOLOGY Co.,Ltd. |
|
TR01 | Transfer of patent right |