CN106358226A - QoS (Quality of Service) optimization method based on cognitive network - Google Patents

Abstract

The invention aims at a problem that QoS (Quality of Service) dynamic optimization is difficult to realize by a cognitive network and realizes QoS elastic optimization management of the cognitive network through establishing a delayed optimization model based on a path, a network throughput capacity model based on cognition, a network fault detection probability and an error alarming probability.

Description

A kind of qos optimization method based on cognition network

Technical field

The present invention relates to communication network field, more particularly to queueing theory, and optimum theory.

Background technology

Cognitive radio, is to further expand on the basis of the software radio that technical college of imperial family of Sweden defines at it, it Break the radio spectrum resources management of traditional " barrier between different departments " and used, by having the Wireless Telecom Equipment sense of cognitive function Know the service condition of surrounding spectrum, and system parameter setting independently can be adjusted according to the change of frequency spectrum service condition, and then Effectively improve radio spectrum resources utilization rate over time and space, this improves frequency spectrum money under the conditions of solving limited resources This communication difficult problem of source utilization rate has irreplaceable progradations, is also a pass promoting mobile Internet to flourish Key technology.On the other hand, from broadly analyzing, cognitive radio technology refers to that all kinds of wireless terminal devices have enough intelligence Or cognitive competence, by the perception of wireless environment around is gathered simultaneously current state and carry out detecting with reference to historical information, Analysis, study and planning, by the adaptive adjustment configured transmission of oneself, are reached and (are included using most suitable Radio Resource Frequency, modulation system, transmission power etc.) complete to be wirelessly transferred.This will be the support technology of Internet of Things development.With Internet of Things Fast development, it is contemplated that the development of intelligent terminal's quantitative indicator type, is generalized to whole perception net simple individual optimal resource allocation Network will be the developing direction of future mobile communications, and the reference configuration of cognition network is as shown in Figure 1.

Primary user's network and time user network is typically co-exist in cognitive radio networks.Primary user's network mainly includes leading User and primary user base station pu.Wherein pu also can become referred to as authorized user, refers to that those have proprietary mandate to certain section of frequency, This mandate Duan Kewei cellular network frequency range, television broadcast band etc..And pbs also can become authorized base station, it is fixing basis Unit, has to the management function authorizing frequency range, and its Main Function is if controlling access and the resource allocation policy of pu.Secondary use Family network, alternatively referred to as dynamic spectrum access network, dynamic spectrum share network or unlicensed networks.

Therefore, for lifting the resource management capacity of cognition wireless network it is necessary to design efficient qos support method.

Content of the invention

The technical problem to be solved is: by set up based on path postpone Optimized model, based on cognitive Network throughput model and Network Fault Detection probability and false alarm probability, realize the qos elasticity optimum management of cognition network.

The present invention comprises the following steps by solving the technical scheme that above-mentioned technical problem is adopted, as shown in Figure 2:

A, the delay Optimized model based on path for the foundation；

B, foundation are based on cognitive network throughput model；

C, set up Network Fault Detection probability and false alarm probability.

In described step a, business sends optimization probabilistic model and is WhereinTransmission collision probability for node v, τ_wSend the probability of Business Stream for node w, cw^vFor the competition window of node v, k is the business number with priority of node w, and m keeps out of the way thresholding for conflict, and n is network section Points, ξ is business sending probability expected value.

In described step a, particularly as follows: based on the delay Optimized model in path being: wherein h is the set of the cluster in network, d For end-to-end requirements set, d is end-to-end demand mark, r_dFor the road with business demand d between transmitting terminal and receiving terminal Footpath is gathered, θ_idJumping figure between for cluster, if the business demand of cluster i is d, θ_id=2, on the contrary then θ_id=1,Produced by business demand d Raw offered load, w_iFor the packet average latency in cluster i,For producing the probability of congestion in cluster i,For decision-making Variable, if the data transfer on the path r of business demand d is through cluster i,Otherwise thenc_clusFor in network The average size of cluster,

$\begin{array}{cc}min& \underset{i\&element;h}{\σ}{w}_i\end{array}$

$\begin{array}{cc}s.t.& w_i=\frac{1}{c_{clus}}\·\frac{{\mathrm{\ρ}}_i^{clus}}{1-{\mathrm{\ρ}}_i^{clus}}\end{array}$

$\∀i\&element;h:{\mathrm{\ρ}}_i^{clus}=\underset{d\&element;d}{\σ}\underset{r\&element;r_d}{\σ}{\mathrm{\ρ}}_d^{dem}{\mathrm{\θ}}_{id}{\mathrm{\δ}}_{id}^r\≤1.$

In described step b, set up based on cognitive network throughput model, particularly as follows: when primary user does not exist When, the handling capacity of networkIn the presence of primary user, the handling capacity of networkh_|m×n|For having the channel square of m transmission antenna and n reception antenna Battle array,WithIt is respectively the signal to noise ratio from user and primary user,For Gauss White noise variance, y is number of users, i_nFor unit battle array.

In described step c, Network Fault Detection probabilityNetwork error Alarm probabilitiesWhereinFor detection threshold, η₁For entering line frequency The decision-making thresholding of spectrum cavity detection, l is frame length, h_eThe channel matrix using for user e, h is expectation channel matrix,Be from The signal variance at family,Signal variance for primary user.

Brief description

The reference configuration schematic diagram of Fig. 1 cognition network

The qos Optimizing Flow schematic diagram based on cognition network for the Fig. 2

Specific embodiment

For reaching above-mentioned purpose, technical scheme is as follows:

The first step, sets up the delay Optimized model based on path, particularly as follows: business sends optimization probabilistic model beingWhereinSending out for node v Send collision probability, τ_wSend the probability of Business Stream, cw for node w^vFor the competition window of node v, k be node w there is priority Business number, m keeps out of the way thresholding for conflict, and n is number of network node, and ξ is business sending probability expected value.

Second step, based on the delay Optimized model in path be: wherein h is the set of the cluster in network, d is end-to-end demand Set, d is end-to-end demand mark, r_dFor the set of paths with business demand d between transmitting terminal and receiving terminal, θ_idFor cluster Between jumping figure, if the business demand of cluster i be d, θ_id=2, on the contrary then θ_id=1,Offered load produced by business demand d, w_iFor the packet average latency in cluster i,For producing the probability of congestion in cluster i,For decision variable, if business needs Seek data transfer on the path r of d through cluster i thenOtherwise thenc_clusFor the average size of the cluster in network,

$\begin{array}{cc}min& \underset{i\&element;h}{\σ}{w}_i\end{array}$

$\begin{array}{cc}s.t.& w_i=\frac{1}{c_{clus}}\·\frac{{\mathrm{\ρ}}_i^{clus}}{1-{\mathrm{\ρ}}_i^{clus}}\end{array}$

$\∀i\&element;h:{\mathrm{\ρ}}_i^{clus}=\underset{d\&element;d}{\σ}\underset{r\&element;r_d}{\σ}{\mathrm{\ρ}}_d^{dem}{\mathrm{\θ}}_{id}{\mathrm{\δ}}_{id}^r\≤1.$

3rd step, sets up based on cognitive network throughput model, particularly as follows: when primary user does not exist, The handling capacity of networkIn the presence of primary user, the handling capacity of networkh_m×nFor having the channel matrix of m transmission antenna and n reception antenna,WithIt is respectively the signal to noise ratio from user and primary user,For white Gaussian Noise variance, y is number of users, i_nFor unit battle array.

4th step, sets up Network Fault Detection probability and false alarm probability, particularly as follows: Network Fault Detection probabilityNetwork error alarm probabilitiesIts InFor detection threshold, η₁For carrying out the decision-making thresholding of frequency spectrum cavity-pocket detection, l is frame length, h_eFor The channel matrix that user e uses, h is expectation channel matrix,It is the signal variance from user,Signal side for primary user Difference.

The present invention proposes a kind of qos optimization method based on cognition network, by dynamic regulation laminated network capacity and Set up qos Optimized model, realize the qos elasticity optimum management of laminated network.

Claims (5)

1. a kind of qos optimization method based on cognition network, optimizes mould by dynamic regulation laminated network capacity with setting up qos Type, realizes the qos elasticity optimum management of laminated network, comprises the steps:

A, the delay Optimized model based on path for the foundation；

B, foundation are based on cognitive network throughput model；

C, set up Network Fault Detection probability and false alarm probability.

2. method according to claim 1, for described step a it is characterized in that: business send optimize probabilistic model beWhereinSending out for node v Send collision probability, τ_wSend the probability of Business Stream, cw for node w^vFor the competition window of node v, k be node w there is priority Business number, m keeps out of the way thresholding for conflict, and n is number of network node, and ξ is business sending probability expected value.

3. method according to claim 1, for described step a it is characterized in that: based on the delay Optimized model in path be: its Middle h is the set of the cluster in network, and d is end-to-end requirements set, and d is end-to-end demand mark, r_dFor transmitting terminal and receiving terminal Between the set of paths with business demand d, θ_idJumping figure between for cluster, if the business demand of cluster i is d, θ_id=2, otherwise then θ_id=1,Offered load produced by business demand d, w_iFor the packet average latency in cluster i,For in cluster i Produce the probability of congestion,For decision variable, if the data transfer on the path r of business demand d is through cluster i,Instead Thenc_clusFor the average size of the cluster in network,

$\begin{array}{cc}min& \underset{i\&element;h}{\σ}{w}_i\end{array}$

$\begin{array}{cc}s.t.& w_i=\frac{1}{c_{clus}}\·\frac{{\mathrm{\ρ}}_i^{clus}}{1-{\mathrm{\ρ}}_i^{clus}}\end{array}$

$\begin{array}{cc}\∀i\&element;h:& {\mathrm{\ρ}}_i^{clus}=\underset{d\&element;d}{\σ}\underset{r\&element;r_d}{\σ}{\mathrm{\ρ}}_d^{dem}{\mathrm{\θ}}_{id}{\mathrm{\δ}}_{id}^r\≤1\end{array}.$

4. method according to claim 1, for described step b it is characterized in that: set up based on cognitive network throughput mould Type, particularly as follows: when primary user does not exist, the handling capacity of networkWhen primary In the presence of family, the handling capacity of networkh_|m×n|For having m transmission antenna With the channel matrix of n reception antenna,WithIt is respectively from user and master The signal to noise ratio of user,For white Gaussian noise variance, y is number of users, i_nFor unit battle array.

5. method according to claim 1, for described step c it is characterized in that: Network Fault Detection probabilityNetwork error alarm probabilitiesWhereinFor detection threshold, η₁For carrying out the decision-making thresholding of frequency spectrum cavity-pocket detection, l is frame length, h_eIt is use The channel matrix that family e uses, h is expectation channel matrix,It is the signal variance from user,Signal variance for primary user.