CN100384115C - Optimization service allocated channel adaptive routing method - Google Patents

Abstract

The present invention provides a shortest path route realizing method based on the channel self-adaptation of a receiver in a mobile communication system. The method comprises steps that a mobile station collects the information of the mobile communication network in real time; the mobile station carries out quality estimation on channels between the mobile station and other mobile stations and determines the modulation speed rate between the mobile station and other mobile stations; the mobile station calculates the channel quality factor of chains i with other mobile stations through a formula; the link cost of one chain i between the mobile station X and the mobile station Y is determined on the basis of the formula; each mobile station maintains a chain state matrix CSM; the shorted path and the secondly shortest path route from the mobile station to a destination node are determined on the basis of the link state matrix; the optimization probability of the service distribution of the shorted path and the secondly shortest path are calculated, and services are distributed to the shortest path and the secondly shortest path respectively according to the probability.

Description

Optimize the channel self-adapting method for routing of traffic assignments

Technical field

The present invention relates to a kind of wireless system that adopts adaptive modulation technology, relate in particular to the method for routing in wireless route network (" Mesh " network), self-organizing (Ad hoc) mobile network and the relay system.

Background technology

Self-organizing (Ad hoc) mobile network is convenient, fast, be not subjected to the characteristics of when and where restriction, it not only can be used separately, satisfy the demand of people's direct communication, and can be superimposed upon on the network of existing infrastructure, to strengthen covering and to activate business, for the user provides adaptive, access service flexibly.

Wireless route network (" Mesh " network) is a kind of effective scheme that broadband services is expanded to a large amount of low sides user market, it has reduced the requirement to wireless line-of-sight transmission, and network construction form is flexible, and cost is cheap, is that future communications has one of ten big technology of development prospect.

But one of subject matter of these network faces is Route Selection.The method for routing of optimizing should be based on present network state, promptly should be dynamic or adaptive.For " Mesh " network, Ad hoc mobile network and relay system, because the restriction of its dynamic topological structure and Radio Resource makes the research of routing algorithm more be rich in challenge.

The performance of wireless communication system is subjected to the restriction of mobile radio channel, and wireless channel is extremely complicated.The multipath transmission of path loss, signal and the mobile Doppler frequency shift that causes etc. all will make signal produce decline, rolling up and ever-increasing demand to the high-speed multimedia business of wireless user requires following wireless network that higher data rate and bigger bandwidth can be provided simultaneously.This demand is not to be unreachable to, and optimizes the adaptive modulation and coding technology of channel conditions, can greatly improve bandwidth availability ratio.If adaptive modulation technology is introduced in " Mesh " network, mobile Ad hoc network and the relay system, and, then might obtain source node to minimum transfer time delay and maximum throughput between the destination node as the criterion of selecting route.

Consider in the Routing Protocol of Link State that based on the on-demand routing protocol (ABR) of association, its main target is the path that provides the most stable between node can reduce the frequency of rebuilding route like this, thereby reduce expense, saves bandwidth resources.Signal stabilization route (SSR) will be set up and optimal path on strong signal link all the shortest, and it has considered the stability problem that connects equally.Based on the multipath Routing Protocol of diversity code, each packet is encoded, and the information after will encode with the method for optimizing distributes on mulitpath, to minimize packet loss rate, balance business, the end-to-end time delay performance of improvement system.

But, if only consider channel conditions, height with modulation rate is that criterion is selected route, might make all nodes all select the link of high modulation speed, will certainly cause congested at some receiving node place like this, cause big time delay and packet loss, thus the valuable Radio Resource of waste.

Summary of the invention

Consider the unreliable and destabilizing factor of wireless channel, the present invention proposes the method for routing that a kind of being applicable to " Mesh " network, self-organizing move Ad hoc network and relay system, promptly optimizes the channel self-adapting method for routing (OSD-CASPR) of traffic assignments.If exist mulitpath between source node and the destination node, the method according to this invention can be grouped in the route on these paths according to different path bandwidth optimizations, to realize the target of whole network average packet time delay minimum.

According to the present invention, a kind of channel self-adapting method for routing of optimizing traffic assignments in mobile communication system is provided, described method comprises step:

(1) information of the described mobile communications network of travelling carriage real-time collecting;

(2) described travelling carriage for and other travelling carriages between channel carry out quality and estimate, and determine and other travelling carriages between modulation rate;

(3) described travelling carriage passes through the channel quality factor of link i between following formula calculating and other travelling carriages: $Q_i\left(t\right)=\frac{C_i\left(t\right)}{C_0},$

I=0 wherein, 1,2 ... k-1, C ₀Represent the peak transfer rate under the best channel conditions condition, C _iRepresent the actual modulated speed of link i between two travelling carriage X and the Y;

(4) determine the link cost of link i between travelling carriage X and the travelling carriage Y based on following formula:

$C_{\mathrm{XY}}=\frac{1}{Q_{\mathrm{XY}}}$

Q wherein _XyIt is the channel quality factor of link i between travelling carriage X and the travelling carriage Y;

(5) each described travelling carriage is kept a Link State Matrix C SM, promptly

CSM＝{E(x，y) _N×N|1≤x，y≤N}

$E(x,y)=E_x^y\left(t\right)=\frac{1}{Q_{\mathrm{XY}}}$ $E_x^y\left(t\right)\∈\{\∞,E_0,E_{1,},...E_i...E_{k-1}\};$

(6), determine shortest path and time minor axis route of described travelling carriage to destination node based on described Link State matrix;

(7) by the optimization probability P of following formula calculating to the shortest path distribution service ₁ ^*: ${\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}=\frac{\sqrt{V_1}}{\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1}+\sqrt{{\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">V</figure-callout>}}_2}}+\frac{\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1\&times;{\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">V</figure-callout>}}_2}}{\mathrm{a\&lambda;}\&times;(\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1}+\sqrt{V_2})}\&times;(\sqrt{V_1}-\sqrt{V_2}),$ V wherein ₁, V ₂Be the bottleneck capacity of shortest path and second shortest path, α shows the length of grouping, and λ is the packet arrival rate of node.By the optimization probability P of following formula calculating to the second shortest path distribution service ₂ ^*:

${\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2^{*}=1-{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}$

Here, $0\&le;\mathrm{\&lambda;}{P}_i^{*}<{\mathrm{\&mu;}}_i,$ μ wherein _iIt is the link capacity of link i;

(8) according to above-mentioned probability respectively to shortest path and second shortest path distribution service.

Introduce the term that will use in this patent below:

R: candidate's set of routes is produced by route selection algorithm.

δ _r ⁱ: the probability of the selected link as route r of link i.

λ _r: the packet arrival rate on the r path

C _r ⁱ(t): the capacity of i link on the r path, so the service speed of link i is ${\mathrm{\&mu;}}_r^i=\frac{C_r^i\left(t\right)}{a},$ Just with the link capacity of grouping expression.

V _r(t): the bottleneck capacity of path r is a time to time change, r=1, and 2 ...

$V_r\left(t\right)=\min \left\{C_r^i\left(t\right)\right\},i=\mathrm{1,2},...,$

Capacity with the grouping expression is: ${\mathrm{\&mu;}}_r\left(t\right)=\frac{V_r\left(t\right)}{a}$

P _r(t):, select the probability of path r routing packets, P at time t _r(t) ∈ [0,1]

And ∑ P _r(t)=1

P _r ^*(t): corresponding to the P of optimization aim function _r(t).

Time t is through the stable state average packet time delay of path r.

Description of drawings

Also describe the present invention in conjunction with the embodiments in detail below with reference to accompanying drawing, wherein:

Fig. 1 is the schematic diagram of multipath distribution service;

Fig. 2 is the comparison of Dan Jing and two footpath packet delay performances.

Embodiment

As shown in Figure 1, the multipath route just is meant and has many path candidates between source node and the destination node.Below in conjunction with specific embodiment adaptive routing method of the present invention is described.

The first step, channel is carried out analysis modeling:

In the adaptive modulation system, the signal to interference and noise ratio of channel (SINR) has been quantified as some rank, and corresponding one by one with certain modulation system, and therefore, the channel transition of adaptive modulation system are shifted and can be represented with the multimode Markov chain.

Under Rayleigh channel, make A represent the signal to noise ratio of acknowledge(ment) signal, then A and signal envelope square is directly proportional.The probability density distribution of A is meant the number form formula, can be write as

$p_A\left(a\right)=\frac{1}{\mathrm{\&rho;}}\exp \{-\frac{a}{\mathrm{\&rho;}}\}$ Here, ρ=E[A], α 〉=0 (1)

Make f _mRepresent maximum doppler frequency, in cellular mobile communication, $f_m=\frac{v}{\mathrm{\&lambda;}},$ Wherein, v is the speed of travelling carriage, and λ is a wavelength.And in mobile Ad hoc network, Hu Tongxin both sides have double mobility, maximum doppler frequency at this moment in motion mutually $f_m={\mathrm{<figure-callout\; id="1"\; label="f\; m"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">f</figure-callout>}}_m=\frac{\mathrm{<figure-callout\; id="1"\; label="v"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">v</figure-callout>}1+v2}{c}{f}_c,$ Wherein, v ₁, v ₂Be the speed of node, f _cBe carrier frequency.Then in the unit interval, signal to noise ratio is faded to number of times (be similar to level crossing rate (lcr), the can be described as the signal to noise ratio percent of pass) N of a certain given level α _aBe one by f _mAnd p _A(α) function of common influence then has:

$N_a=\sqrt{\frac{2\mathrm{\&pi;a}}{\mathrm{\&rho;}}}{f}_m\exp \{-\frac{a}{\mathrm{\&rho;}}\}---\left(2\right)$

Make 0=A ₀＜A ₁＜A ₂＜...＜A _K=∞ is the threshold value of SNR.The signal to noise ratio of Rayleigh channel can be quantified as a K state.If α ∈ is [A _k, A _K+1), then claim a to belong to state S _k, k=0,1 ..., K-1, and these states, in adaptive modulation system, can be corresponding with different modulating-coding compound modes.

According to formula (1), can write out signal to noise ratio and be in the probability of stability of k state:

$p_k={\&Integral;}_{A_k}^{A_{k+1}}\frac{1}{\mathrm{\&rho;}}\exp \{-\frac{x}{\mathrm{\&rho;}}\}\mathrm{dx}=\exp \{-\frac{A_k}{\mathrm{\&rho;}}\}-\exp \{-\frac{A_{k+1}}{\mathrm{\&rho;}}\}---\left(3\right)$

In our research, suppose that the decline of channel is enough slow, make the change of state only may occur between the adjacent states.This hypothesis meaning in actual applications is the prediction of channel and feeds back the variation that can get caught up in channel.Make p _{I, j}Expression state i, the transition probability between the j then has p _{I, j}=0, $\&ForAll;|i-j|>1.$

Based on such hypothesis, ask transition probability p below _{I, j}

Because shifting can only be between adjacent states.So, incident: should be that threshold value is A from state k to the transition probability of state k-1 _kThe signal to noise ratio percent of pass divided by the number of times that the k state takes place in the per second.If the character rate of link is Rsymbols/s, then in the per second, the number of times that the k state takes place can be used the number of symbols R of the inherent k state transmission of per second ^kRepresent, and have

R ^k＝R×p _k。(4)

SINR drops to thresholding A in the note per second _kUnder number of times be N _k, k=1,2,3 ..., K-1.

By (2), have

$N_k=\sqrt{\frac{{2\mathrm{\&pi;A}}_k}{\mathrm{\&rho;}}}{f}_m\exp \{-\frac{A_k}{\mathrm{\&rho;}}\}---\left(5\right)$

Then, the transition probability that can get Markov chain is:

$p_{k,k+1}=\frac{N_{k+1}}{R^k},k=\mathrm{0,1,2,3},\&CenterDot;\&CenterDot;\&CenterDot;,K-2---\left(6\right)$

$p_{k,k-1}=\frac{N_k}{R^k},k=\mathrm{1,2,3},\&CenterDot;\&CenterDot;\&CenterDot;,K-2.---\left(7\right)$

p _k，k＝1-p _k，k+1-p _k，k-1，k＝1，2，3，…，K-2 (8)

p _0，0＝1-p _0，1；p _K-1，K-1＝1-p _K-1，K-2 (9)

Can verify above-mentioned derivation by the equilibrium response of Markov chain.

For example, at state S ₀,, have according to equilibrium equation

p ₀×p _0，1＝p ₁×p _1，0 (10)

With (4)-(6) formula substitutions (10), equation the right and left all equals N ₁/ R is same, can proofing state S _kUnder equilibrium equation.

p _k×(p _k，k+1+p _k，k-1)＝p _k-1×p _k-1，k+p _k+1×p _k+1，k (11)

With (4)-(6) formula substitutions (11), can see that the equation left side is again

$\frac{R^k}{R}\&times;(\frac{N_{k+1}}{R^k}+\frac{N_k}{R^k})=\frac{1}{R}(N_{k+1}+N_k)$

Equation the right equals

$\frac{R^{k-1}}{R}\&times;\frac{N_k}{R^{k-1}}+\frac{R^{k+1}}{R}\&times;\frac{N_{k+1}}{R^{k+1}}=\frac{1}{R}(N_{k+1}+N_k)$

Equation the right and left equates, the homeostasis characteristic that transition probability that this proof is above-mentioned and probability of stability acquiring method can satisfy Markov chain.

In discussing above, the A in (3) formula ₀, A ₁, A ₂... A _KEquivalence all is predefined.Therefore, the channel signal to noise ratio (snr) is in state S _kProbability distribution only relevant with ρ.If transmitting power is certain, receiver structure is identical, and then the size of ρ mainly is subjected to the influence of path loss.Different positions, the ρ value of travelling carriage is different, makes that the probability of stability of their pairing Markov chains and the transition probability between each state are also different.In addition, if can measure the translational speed of node, according to formula $f_m={\mathrm{<figure-callout\; id="1"\; label="f\; m"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">f</figure-callout>}}_m=\frac{\mathrm{<figure-callout\; id="1"\; label="v"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">v</figure-callout>}1+v2}{c}{f}_c$ And formula (4)-(6) formula can estimate state transition probability.If node motion speed can not be measured, then need the SNR situation of change of node according to actual measurement, directly carry out statistical analysis, obtain state transition probability.

The state of SNR be with certain order of modulation and channel coding schemes one to one, and order of modulation and channel coding schemes corresponding certain link transmission speed.Therefore, we can be corresponding with the corresponding link transmission rate with each state of above-mentioned Markov chain, just obtained the markov link model of rate variation.

The channel self-adapting improved shortest path by research in because utilization finite state markov channel model carries out modeling to fading channel, and adopted adaptive channel encoding and modulation technique.When also being, the channel throughput that time varying channel produced between the portable terminal becomes.So just can be in the research of route, the ins and outs of physical layer are shielded, make research be easy to carry out.

Second step, set up routing table, write down the instant link circuit condition between every pair of node and transfer to the probability of NextState with matrix.

The 3rd step, the calculating channel factor of quality.If use S={C ₀, C ₁... C _K-1The a series of limited channel statuss of representative, C ₀Represent the peak transfer rate (just corresponding to the highest order of modulation) under the optimum channel status condition, C _iRepresent the actual modulated speed of link i between two terminals, then the channel quality factor of link i is between two terminals: $Q_i\left(t\right)=\frac{C_i\left(t\right)}{C_0}$ I=0,1,2 ... k-1, by the quality of its reflection channel quality.So preferably the factor of quality of link is 1, and the factor of quality of other link is along with C _iReduce and less than 1.

In the 4th step, calculate link cost.Based on the channel self-adapting improved shortest path of receiver by (RB-CASPR) in, link cost is

$C_{\mathrm{XY}}=\frac{1}{Q_{\mathrm{XY}}}$

Wherein,

Be the channel quality part of X to the Y link cost.

In the 5th step, each travelling carriage is kept a Link State matrix (CSM).

CSM＝{E(x，y) _N×N|1≤x，y≤N}

$E(x,y)=E_x^y\left(t\right)=\frac{1}{Q_{\mathrm{XY}}},$ $E_x^y\left(t\right)\&Element;\{\&infin;,E_0,{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,},...E_i...E_{k-1}\}$

If E _x ^y(t) be not equal to infinitely, illustrate that then node x can directly send packet to node y.Otherwise node x and node y then can not direct communications.

The 6th step, according to connection status Matrix C SM, each source travelling carriage utilize based on Bel graceful-the distance vector method of Ford (Bellman-Ford) shortest path first seeks minor axis and time minor axis route.

The 7th goes on foot, and calculates the probability of distribution service on every paths, and according to the probability assignments business.

According to the definition of front, the packet arrival rate on the r of path is λ P _r

At time t, the stable state average packet time delay on the r of path:

${\stackrel{\&OverBar;}{T}}_r\left(t\right)=\frac{1}{{\mathrm{\&mu;}}_r\left(t\right)-\mathrm{\&lambda;}{P}_r\left(t\right)}$

Here, λ P _r(t)≤μ _r(t)

Whole packet arrival rates on the link i are:

So whole network delay: $\stackrel{\&OverBar;}{T}=\frac{1}{\mathrm{\&gamma;}}\underset{i\&Element;A}{\mathrm{\&Sigma;}}\frac{\underset{r\&Element;R}{\mathrm{\&Sigma;}}{\mathrm{\&lambda;}}_r{\mathrm{\&delta;}}_r^i{P}_r}{\frac{C_r^i}{a}-\underset{r\&Element;R}{\mathrm{\&Sigma;}}{\mathrm{\&lambda;}}_r{\mathrm{\&delta;}}_r^i{P}_r}$

The Route Selection problem is to determine the Route Selection probability P _r, make:

$\stackrel{\&OverBar;}{T}=\min \left\{\frac{1}{\mathrm{\&gamma;}}\underset{i\&Element;A}{\mathrm{\&Sigma;}}\frac{\underset{r\&Element;R}{\mathrm{\&Sigma;}}{\mathrm{\&lambda;}}_r{\mathrm{\&delta;}}_r^i{P}_r}{\frac{C_r^i}{a}-\underset{r\&Element;R}{\mathrm{\&Sigma;}}{\mathrm{\&lambda;}}_r{\mathrm{\&delta;}}_r^i{P}_r}\right\}---\left(1\right)$

Constraints:

$a\underset{r\&Element;R}{\mathrm{\&Sigma;}}{\mathrm{\&lambda;}}_r{\mathrm{\&delta;}}_r^i{P}_r\&le;C_r^i,\&ForAll;i\&Element;A---\left(2\right)$

$\underset{r\&Element;R}{\mathrm{\&Sigma;}}{P}_r=1---\left(3\right)$

Restriction relation (2) guarantees that the traffic carrying capacity on each bar link is no more than its capacity, and (3) formula guarantees route of each grouping.Above-mentioned is the optimization problem of more options, multi-constraint condition, promptly is a complete problem of NP.

Consider cost and the expense of seeking route and storage routing table, in the present invention the multipath route is restricted to two paths.Just OSD-CASPR according to the channel quality factor, not only keeps minimal cost path in the pathfinding process, has also preserved source node simultaneously to the inferior little cost path between the destination node.

In the OSD-CASPR route,, at this moment be the CASPR route if traffic carrying capacity hour, then utilizes the single path transmission all professional; When traffic carrying capacity is increased to a certain degree, then utilize shortest path and second shortest path transport service.

Traffic assignments problem on two paths is discussed below.

If require the total cost balance of link of this two paths, at this moment should have:

$\frac{\&PartialD;({\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1\&times;\stackrel{\&OverBar;}{{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">T</figure-callout>}}_1})}{\&PartialD;{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1}|{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}=\frac{\&PartialD;({\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2\&times;\stackrel{\&OverBar;}{{\mathrm{<figure-callout\; id="2"\; label="T"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">T</figure-callout>}}_2})}{\&PartialD;{\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2}|{\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2^{*}\&le;\frac{\&PartialD;(P_l\&times;\stackrel{\&OverBar;}{T_l})}{{\&PartialD;P}_l}|P_l^{*}=0,\&ForAll;l\&NotEqual;\mathrm{1,2}$

Just

$\frac{{\mathrm{\&mu;}}_1}{{({\mathrm{\&mu;}}_1-\mathrm{\&lambda;}{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*})}^2}=\frac{{\mathrm{\&mu;}}_2}{{({\mathrm{\&mu;}}_2-\mathrm{\&lambda;}{\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2^{*})}^2}$

Can draw, select path V ₁The optimization probability of distribution service:

${\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}=\frac{\sqrt{{\mathrm{\&mu;}}_1}}{\sqrt{{\mathrm{\&mu;}}_1}+\sqrt{{\mathrm{\&mu;}}_2}}+\frac{\sqrt{{\mathrm{\&mu;}}_1\&times;{\mathrm{\&mu;}}_2}}{\mathrm{\&lambda;}\&times;(\sqrt{{\mathrm{\&mu;}}_1}+\sqrt{{\mathrm{\&mu;}}_2})}\&times;(\sqrt{{\mathrm{\&mu;}}_1}-\sqrt{{\mathrm{\&mu;}}_2})$

Promptly ${\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}=\frac{\sqrt{V_1}}{\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1}+\sqrt{{\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">V</figure-callout>}}_2}}+\frac{\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1\&times;{\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">V</figure-callout>}}_2}}{\mathrm{a\&lambda;}\&times;(\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1}+\sqrt{V_2})}\&times;(\sqrt{V_1}-\sqrt{V_2})$

Select path V ₂The optimization probability of distribution service:

${\mathrm{<figure-callout\; id="2"\; label="P"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">P</figure-callout>}}_2^{*}=1-{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{*}$

Here, $0\&le;\mathrm{\&lambda;}{P}_i^{*}<{\mathrm{\&mu;}}_i,$ μ wherein _iIt is the link capacity of link i.

If the bottleneck capacity of two paths is:

${\mathrm{\&mu;}}_{12}=\underset{l=1}{\overset{2}{\mathrm{\&Sigma;}}}{\mathrm{\&mu;}}_l$

Discussion according to the front can have following service-allocation algorithm:

Distribution service on path 1:

${\mathrm{\&mu;}}_1-\frac{\sqrt{{\mathrm{\&mu;}}_1}\&times;({\mathrm{\&mu;}}_{12}-\mathrm{\&lambda;})}{\underset{l=1}{\overset{2}{\mathrm{\&Sigma;}}}\sqrt{{\mathrm{\&mu;}}_l}},i=\mathrm{1,2}$

Promptly $\frac{V_1}{a}-\frac{\sqrt{{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20041000154000141.png"\; state="\{\{state\}\}">V</figure-callout>}}_1}\&times;(\frac{V_{12}}{a}-\mathrm{\&lambda;})}{\underset{l=1}{\overset{2}{\mathrm{\&Sigma;}}}\sqrt{V_l}},$ I=1,2, wherein α is the length of grouping.

Distribution service on path 2:

${\mathrm{\&mu;}}_2-\frac{\sqrt{{\mathrm{\&mu;}}_2}\&times;({\mathrm{\&mu;}}_{12}-\mathrm{\&lambda;})}{\underset{l=1}{\overset{2}{\mathrm{\&Sigma;}}}\sqrt{{\mathrm{\&mu;}}_l}},i=\mathrm{1,2}$

Promptly $\frac{V_2}{a}-\frac{\sqrt{{\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="C20041000154000142.png"\; state="\{\{state\}\}">V</figure-callout>}}_2}\&times;(\frac{V_{12}}{a}-\mathrm{\&lambda;})}{\underset{l=1}{\overset{2}{\mathrm{\&Sigma;}}}\sqrt{V_l}},i=\mathrm{1,2}$

What Fig. 2 showed is the performance comparison of single path and two-way footpath route.As can be seen, when low traffic (λ≤4.6 packets/second), single path route delay performance is better than two-way footpath route.Because of the channel of inferior little cost path in the route of two-way footpath not necessarily is in good state, thus when traffic carrying capacity than hour, utilize the minimum cost route, the delay performance that can obtain.But along with the increase (λ＞4.6 packets/second) of traffic carrying capacity, CASPR might form congested at some node place, degradation.If at this moment adopt two-way footpath route, just can improve network performance, as shown in FIG..And the OSD-CASPR route makes full use of the good one sides of two kinds of routing performances and (when being low traffic, uses the single path route; During heavy traffic, use two-way footpath route), can obtain less network delay like this.

But the threshold value of traffic carrying capacity in the OSD-CASPR route, promptly the switching point of single path and two-way footpath route will come to determine by experiment.In addition, OSD-CASPR utilizes the bottleneck capacity of every paths to carry out traffic assignments, so might there be the phenomenon of some link bandwidth waste.

Claims (5)

1. channel self-adapting method for routing of in mobile communication system, optimizing traffic assignments, described method comprises step:

(1) information of the described mobile communications network of travelling carriage real-time collecting;

(2) described travelling carriage for and other travelling carriages between channel carry out quality and estimate, and determine and other travelling carriages between modulation rate;

(3) described travelling carriage passes through the channel quality factor of link i between following formula calculating and other travelling carriages: $Q_i\left(t\right)=\frac{C_i\left(t\right)}{C_0},$

I=0 wherein, 1,2 ... k-1, C ₀Represent the peak transfer rate under the best channel conditions condition, C _iRepresent the actual modulated speed of link i between two travelling carriage X and the Y;

(4) determine the link cost of link i between travelling carriage X and the travelling carriage Y based on following formula:

$C_{\mathrm{XY}}=\frac{1}{Q_{\mathrm{XY}}}$

Q wherein _XyIt is the channel quality factor of link i between travelling carriage X and the travelling carriage Y;

(5) each described travelling carriage is kept a Link State Matrix C SM, promptly

CSM＝{E(x，y) _N×N|1≤x，y≤N}

$E(x,y)=E_x^y\left(t\right)=\frac{1}{Q_{\mathrm{XY}}}$ $E_x^y\left(t\right)\&Element;\{\&infin;,E_0,E_{1,},...,E_i...E_{k-1}\};$

(6), determine shortest path and time minor axis route of described travelling carriage to destination node based on described Link State matrix;

(7) by the optimization probability P of following formula calculating to the shortest path distribution service ₁ ^*:

$P_1^{*}=\frac{\sqrt{V_1}}{\sqrt{V_1}+\sqrt{V_2}}+\frac{\sqrt{V_1\&times;V_2}}{\mathrm{a\&lambda;}\&times;(\sqrt{V_1}+\sqrt{V_2})}\&times;(\sqrt{V_1}-\sqrt{V_2}),$ V wherein ₁, V ₂Be the bottleneck capacity of shortest path and second shortest path, α shows the length of grouping, and λ is the packet arrival rate of node.By the optimization probability P of following formula calculating to the second shortest path distribution service ₂ ^*:

$P_2^{*}=1-P_1^{*}$

Here, $0\&le;\mathrm{\&lambda;}{p}_i^{*}\&le;{\mathrm{\&mu;}}_i,$ μ wherein _iIt is the link capacity of link i;

(8) according to above-mentioned probability respectively to shortest path and second shortest path distribution service.

2. according to the process of claim 1 wherein that described travelling carriage utilizes on-demand routing protocol or table to drive the information of the described mobile communications network of route real-time collecting.

3. according to the process of claim 1 wherein described channel quality factor Q _i(t)≤1.

4. according to the process of claim 1 wherein that the peak transfer rate under the best channel conditions condition passes through the highest order of modulation realization.

5. according to the process of claim 1 wherein described travelling carriage based on described Link State matrix, and utilize based on Bel graceful-the distance vector method of Ford shortest path first determines the shortest path of described travelling carriage to destination node.

Images