CN103139860A - Uniting switch method based on speed self-adaption in isomerism cognition radio network - Google Patents

Abstract

The invention requests for protecting a uniting switch method based on speed self-adaption in an isomerism cognition radio network. The invention relates to the field of wireless communication. Due to the fact that the switch problem is produced by the movement of the position of a secondary level user, the invention provides the uniting switch method of frequency spectrum broker auxiliary selecting network and secondary level user terminal controlling switch. Attribute weight is obtained based on the layer analysis method by the frequency spectrum broker on a network side, thus the network is selected. The switch controlling method of the speed self-adaption is provided on a terminal side, thus the switch from an original network to the target network is controlled. Different switch judgment threshold methods are provided according to different switch types. According to the uniting switch method based on speed self-adaption in an isomerism cognition radio network, a suitable network can be selected according to business requirements and network conditions of the secondary level user. Switch is well controlled. Switch times and suspending probability of the secondary level user are reduced. Switch performance of the secondary level user is improved.

Description

In a kind of isomery cognitive radio networks based on the joint handoff method of speed adaptive

Technical field

The present invention relates generally to wireless communication field, especially about the frequency spectrum switching problem in the isomery cognitive radio networks.

Background technology

Along with the develop rapidly of wireless communication technology and the continuous increase of people's demand, traditional single network can not to meet the need, and the technology of a plurality of heterogeneous wireless network convergence is becoming study hotspot.Along with the fast development of cognitive radio networks research, its demand to heterogeneous wireless network is also further obvious.If secondary user's moves to another place from a place, the frequency spectrum that may carry out because of the variation of available frequency band between heterogeneous network switches.

In heterogeneous wireless network, internetwork vertical switch decision is key technology.Existing vertical switch determining method mainly contains: the first kind adopts the horizontal changing method in similar cellular communication system, the signal strength signal intensity from the heterogeneous networks access point that terminal use's measurement receives, when received signal strength during greater than certain predetermined threshold value, select the network of signal strength signal intensity maximum as the preferred network of access, but these class methods are due to the method for simply using the level switching, the decision criteria that uses is single, often can not satisfy the demand of user's various aspects of performance; Equations of The Second Kind adopts Dynamic Programming and the artificial intelligence theory accurate vertical switch decision of making comparisons, such as pattern recognition, neural net, fuzzy reasoning, and in conjunction with network characteristic and user's request, the complexity of this type of algorithm is high, and will consume the long period and learn in advance, be not suitable for the limited mobile device of computing capability and electric weight; The 3rd class is from network agent, and research is process and the property indices of switching vertically, and based on contextual information is carried out perception, analyzes the impact of switch decision with this; The 4th class adopts the cost function method, general consideration affects the combination of one of condition of vertical handoff method performance or several conditions, these conditions mainly comprise network side system's available bandwidth, connect time-delay etc., and the received signal strength of user's side, user terminal electric quantity consumption, user velocity and service fee etc., determine vertical judgement of switching and the selection of optimum network by the optimization aim function.The cost function method can realize the combination property of the vertical validity of switching and assurance network better due to the multiple attribute that has considered access network and user terminal compared with other several class methods.

In cognitive radio networks, the primary user is using its legal frequency range that has any time, and some idle frequency spectrum situations can appear, general these idle frequency spectrums can put together (being called spectrum pool), managed by a frequency spectrum broker entity, frequency spectrum broker entity also include spectrum allocation may algorithm and management policy (Xie Xianzhong. " cognitive radio power technology and application thereof ". Beijing: Electronic Industry Press 2008.4).

To a plurality of heterogeneous wireless networks based on cognitive radio, because can taking the frequency range of primary user's free time to the chance formula, secondary user's communicates, generally often select from the spectrum pool of frequency spectrum broker entity management when selecting idle frequency range, so the optional frequency range in spectrum pool is more, the communication opportunity of secondary user's is just larger.It is limited that but single network brings the communication condition of secondary user's, and the opportunity available in the mandate frequency range is less, so adopt a plurality of heterogeneous wireless networks to bring more available frequency band for secondary user's in cognitive radio networks, and higher-quality network.On the other hand, the secondary user's terminal has reconfigurability, can be when not revising hardware configuration, in the situation that radio parameter is adjusted in not interrupt communication in real time, as frequency, symbol rate, through-put power, modulation and type of coding etc.Due to such characteristic, communicate by letter also for secondary user's in heterogeneous wireless network and bring convenience; Secondary user's can need according to performance and the application of oneself, selects suitable network and frequency, then the various parameters of correct configuration.

In the method that network is selected, often adopt analytic hierarchy process (AHP) (AHP) to obtain the weights of attribute, then carry out the calculating of network result.At document Shengmei Liu, Su Pan. " An Improved TOPSIS Vertical Handoff Algorithm for Heterogeneous Wireless Networks ", IEEE International Conference on Communication Technology, in 2010:750-754, analytic hierarchy process (AHP) (AHP) is improved, switching effect has had lifting, and throughput has also improved a lot.But it combines the comentropy of attribute, does not consider the variation of the objective state of attribute, and its switching times is higher, the court verdict inaccuracy.

At the terminal switching control part, decision parameter commonly used is generally the sluggish level factor and resident chronon, can be used for reducing the ping-pong of terminal, improves the performance of switching.At document Min Liu, Zhongcheng Li. " Performance Analysis and Optimization of Handoff Algorithms in Heterogeneous Wireless Networks ", IEEE Transactions on Mobile Computing, 2008,7 (7): in 846-857 top two kinds of method for handover control combinations commonly used, propose a kind of method for handover control, reduced the outage probability of terminal when switching.But the coefficient in its controlled condition is got fixed value simply, does not change in conjunction with the variation of actual conditions, is not inconsistent and actual network environment.

Summary of the invention

The present invention is directed to the defects that in prior art, in cognitive radio networks, the vertical handoff method between heterogeneous wireless network exists, for the characteristic of cognitive radio networks, the present invention proposes a kind of joint handoff method of frequency spectrum broker entity assisted Selection and the switching of secondary user's terminal control.At network side, frequency spectrum broker entity obtains the method for attribute weights based on improved analytic hierarchy process (AHP) (AHP), carries out network and selects; In end side, the method for handover control of speed adaptive has been proposed, the switching from former network to objective network is controlled.

The technical scheme that the present invention solves the problems of the technologies described above is: propose a joint handoff method, when the position of secondary user's changes, just the information of secondary user's oneself and frequency spectrum handover request are sent to frequency spectrum broker entity, information comprises: as the received signal strength of each network, the type of service of using, user preference and speed etc.; The frequency spectrum broker receives the received signal strength that at first information number analyze each network, ignore the ineligible network of signal strength signal intensity, for the qualified network of signal strength signal intensity, utilize the network attribute value such as available bandwidth, Channel holding time, network charges of network side, the attribute weights of computing terminal, use network selecting method to calculate the relative degree of closeness of eligible network, select a best switching purpose network, and selection result is returned to secondary user's; Secondary user's is controlled at switching between former network and objective network according to self translational speed and received signal strength.

In the selection course of network side, network selecting method is specially: analyze secondary user's selection to network in moving process, use multiple attributive decision making method, need at first getattr weights for this reason.The present invention proposes to be combined with comentropy and attribute change trend by analytic hierarchy process (AHP) (AHP) and determines, by the objective weight-values that comentropy combines with attribute change trend, revises the subjective weights that obtained by analytic hierarchy process (AHP) (AHP).Like this, both overcome the shortcoming that subjective randomness is large, calculating is coarse and precision is low of analytic hierarchy process (AHP), considered again the objective difference between network, and made the network selection result more accurate.

In the switching controls of end side, the present invention proposes in conjunction with the jointly controlling of the sluggish level factor and resident chronon, and make the switching controls condition with velocity variations adaptively changing, improve the performance of controlling when switching.The present invention adopts logarithm to adjust control method or index replacement control method dynamically-adjusting parameter, makes parameter according to the variation of moving velocity of terminal and rule changes, and the contrast properties result shows that outage probability obviously reduces.

Number adjusting method is adopted with velocity variations affects the Changing Pattern adjustment α of received signal strength and the relative value of β, and its citation form is logarithmic function.In the secondary user's moving process, the value of α is regulated in the variation of Negotiation speed, to reach the purpose of adjusting sluggish level method ratio, have nothing to do and the relative value of the controlled condition of resident timing method and received signal strength is big or small, do not adjust resident timing controlled factor-beta.

The index replacement method adopts the rule that affects velocity ratio with velocity variations to go to adjust the relative value of α and β, and its citation form is exponential function.In the secondary user's moving process, as v＜v _HOThe time, the slope of exponential function is larger, adjusting α that can be apparent in view and the value of β, thus can make the performance of sluggish level method and resident timing method coordinate best; Work as v〉v _HOThe time, the slope of exponential function is less, can make the value of α and β more and more approaching, and the combination of sluggish level method and resident timing method is tended to be steady.

The present invention considers that also vertical switching has asymmetry, according to upwards switching (switching from wlan network to the UMTS network) and switching (switching to wlan network from the UMTS network) downwards, adopt dissimilar switch decision thresholding, when being switching downwards, handoff procedure is easy, for selection type switches, this moment, wlan network was optional network specific digit, and secondary user's is not to switch, and its target is to improve QoS, increase decision threshold this moment, switches when performance reaches requirement again.When upwards switching, handoff procedure is complicated, and secondary user's is about to leave wlan network and covers, can produce data loss, the performance loss meeting is had a great impact, for the pressure type switches, reduce decision threshold this moment, makes its quicker switching in time, to reduce Transmission.

Technical scheme of the present invention is, in a kind of isomery cognitive radio networks based on the joint handoff method of speed adaptive, at network side, frequency spectrum broker entity carries out network according to the network attribute weights and selects based on cost type attribute and the benefit type property calculation network attribute weights of analytic hierarchy process (AHP) according to network; In end side, switch to objective network from former network according to the velocity variations self adaptation.Frequency spectrum broker entity computing network attribute weights specifically comprise: according to the accessible collection of network of isomery cognitive radio networks, and each network attribute set, set up the multiattribute matrix, the multiattribute matrix is carried out standardization, obtain the information entropy e of j attribute of network according to the standardized nature matrix _j, call formula according to the information entropy: Obtain the entropy weights of network attribute j, weight subjective according to the normalization of j attribute of network be ε j (j=1,2 ..., n) and call formula:

${\mathrm{\ω}}_{\mathrm{ij}}=\frac{{\mathrm{\ϵ}}_j\·{\mathrm{\η}}_j\·e^{b_{\mathrm{ij}}-b_{\mathrm{ij}}^n}}{\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ϵ}}_k\·{\mathrm{\η}}_k\·e^{b_{\mathrm{ik}}-b_{\mathrm{ik}}^n}},(i=\mathrm{1,2},\·,m)(j=\mathrm{1,2},\·,n)$ Obtain the actual weights of network i network attribute j, wherein,

B for a front n attribute _ijMean value, according to formula: μ _ij=ω _ijB _ijStandardized value b to attribute _ijBe weighted and obtain the weighting attribute, according to each network of weighting property calculation and ideal network and the distance of poor network, calculate the relative degree of closeness of each network and ideal network, the network of choosing relative closeness value maximum is the best switching purpose network.In end side, terminal adopts based on the joint handoff of the sluggish level factor and resident chronon and controls, determine to switch according to resident chronon and the sluggish level factor, when satisfying formula: $\mathrm{\&alpha;}\frac{D_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{h_y}+\mathrm{\&beta;}\frac{T_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{t_{\mathrm{dw}}}<-\mathrm{\&delta;},$ Terminal is sent switching controls, wherein, and T _{Φ (N-1)}(N) the expression time of staying of secondary user's in network Ф (N-1), D _{Ф (N1)}(N) signal strength signal intensity of expression network Ф (N-1) and handoff threshold is poor, t _dwRepresent resident chronon, h _yRepresent the sluggish level factor, α represents sluggish level method controlling elements, and β represents resident timing method controlling elements, and δ is decision threshold.

Handover scheme of the present invention is according to business demand and the network condition of secondary user's, select best purpose network, control better and switch, number of times and outage probability that secondary user's is switched have been reduced, improved the performance of handoffs of secondary user's, a plurality of heterogeneous wireless network convergence based on cognitive radio have been had important value.

Description of drawings

The handoff process figure of Fig. 1 heterogeneous network;

The access model figure of Fig. 2 heterogeneous network;

The analysis of shift figure of Fig. 3 heterogeneous wireless network;

The average cumulative switching times comparison diagram of Fig. 4 secondary user's;

The outage probability comparison diagram of the secondary user's of the different method for handover control of Fig. 5;

The average transmission data volume variation diagram of Fig. 6 secondary user's.

Embodiment

Below in conjunction with accompanying drawing and instantiation, enforcement of the present invention is described specifically.

The handoff process figure of Fig. 1 heterogeneous network.The present invention is directed to the characteristic of cognitive radio networks, proposed the joint handoff method that frequency spectrum broker assisted Selection and secondary user's terminal control switch.At network side, the frequency spectrum broker, carries out network according to the attribute weights and selects according to available bandwidth, Channel holding time, user preference, network charges computing network attribute weights based on analytic hierarchy process AHP; In end side, implement switching controls from former network to objective network according to the velocity variations self adaptation, carry out switching controls from former network to objective network.

The model of heterogeneous network of the present invention is covered the wide area network UMTS of low bandwidth and the WLAN of low covering high bandwidth and forms heterogeneous wireless network as shown in Figure 2 by the height with Typical Representative.Specific embodiments of the present invention is as follows.

When the position of secondary user's changes, secondary user's sends to frequency spectrum broker entity with the frequency spectrum handover request, and request comprises the information of the received signal strength as each network, the type of service of using, user preference and speed etc.

Frequency spectrum broker entity is analyzed the received signal strength of each network according to the secondary user's frequency spectrum handover request of receiving, ignores the ineligible network of signal strength signal intensity, the qualified network of primary election signal strength signal intensity.

The attribute weights of network side computing terminal

For the qualified network of signal strength signal intensity, the network attribute value of network side computing terminal and corresponding attribute weights.The present embodiment is selected available bandwidth, Channel holding time, and 4 network attributes of user preference and network charges are that example further illustrates:

Be provided with m qualified heterogeneous wireless network: X={x ₁, x ₂, x ₃,, x _m: m accessible collection of network }, as UMTS network, wlan network etc.If each network using n judgement attribute: S={s ₁, s ₂, s ₃,, s _n: n network attribute set }, as network network expense, user preference, available bandwidth, Channel holding time etc.Use a _ijExpression network i(i=1,2,, j(j=1 m), 2,, n) individual property value can obtain the multiattribute matrix A:

$A={\left(a_{\mathrm{ij}}\right)}_{m\&times;n}=\left[\begin{array}{cccc}{a}_{11}& a_{12}& \&CenterDot;& a_{1n}\\ a_{21}& a_{22}& \&CenterDot;& a_{2n}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ a_{m1}& a_{m2}& \&CenterDot;& a_{\mathrm{mn}}\end{array}\right]---\left(1\right)$

Network attribute is divided into cost type attribute and benefit type attribute by character, network charges, user preference etc. is classified as cost type attribute, available bandwidth, Channel holding time etc. is benefit type attribute.

According to the character of network attribute, the multiattribute matrix A is carried out standardization.Be the cost type attributes such as network charges, user preference as network attribute, the smaller the better type is adopted in standardization, according to formula:

$b_{\mathrm{ij}}=\frac{\min \left\{a_{\mathrm{ij}}\right|1\&le;i\&le;m\}}{a_{\mathrm{ij}}},(i=\mathrm{1,2},\&CenterDot;,m)$ The matrix property value is carried out standardization.

Be the benefit type attributes such as available bandwidth, Channel holding time for network attribute, the type that is the bigger the better is adopted in standardization, according to formula:

$b_{\mathrm{ij}}=\frac{a_{\mathrm{ij}}}{\max \left\{a_{\mathrm{ij}}\right|1\&le;i\&le;m\}},(i=\mathrm{1,2},\&CenterDot;,m)$ The matrix property value is carried out standardization.

Through after above-mentioned processing, obtain the standardized nature matrix B:

$B={\left(b_{\mathrm{ij}}\right)}_{m\&times;n}=\left[\begin{array}{cccc}{b}_{11}& b_{12}& \&CenterDot;& b_{1n}\\ b_{21}& b_{22}& \&CenterDot;& b_{2n}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ b_{m1}& b_{m2}& \&CenterDot;& b_{\mathrm{mn}}\end{array}\right]---\left(2\right)$

Call formula:

$e_j=-\frac{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\left(\frac{b_{\mathrm{ij}}}{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{b}_{\mathrm{ij}}}\right)\&CenterDot;\mathrm{In}\left(\frac{b_{\mathrm{ij}}}{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{b}_{\mathrm{ij}}}\right)}{\mathrm{Inm}},(j=\mathrm{1,2},\&CenterDot;,n)---\left(3\right)$ The information entropy e of j attribute of computing network _jIn formula (3), make xInx=0 when x=0.Obtain thus the entropy weights of j attribute:

${\mathrm{\&eta;}}_j=\frac{1-e_j}{\underset{k=1}{\overset{n}{\mathrm{\&Sigma;}}}(1-e_k)},(j=1,2,\&CenterDot;,n)---\left(4\right)$

The subjective weight of normalization of in addition, establishing j attribute of each network is ε _j(j=1,2 ..., n), can require and operator's operation strategy setting according to service quality (QoS), it has reflected the relative significance level of this attribute.

The present invention is according to the dynamic change situation of each network attribute.When certain network attribute value is increasing, just increase weights corresponding to this attribute, increase the impact of this attribute; Otherwise, when certain property value is reducing, just reduce weights corresponding to this attribute, reduce the impact of this attribute.Therefore, convolution (3) and formula (4), according to formula:

${\mathrm{\&omega;}}_{\mathrm{ij}}=\frac{{\mathrm{\&epsiv;}}_j\&CenterDot;{\mathrm{\&eta;}}_j\&CenterDot;e^{b_{\mathrm{ij}}-b_{\mathrm{ij}}^n}}{\underset{k=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&epsiv;}}_k\&CenterDot;{\mathrm{\&eta;}}_k\&CenterDot;e^{b_{\mathrm{ik}}-b_{\mathrm{ik}}^n}},(i=\mathrm{1,2},\&CenterDot;,m)(j=\mathrm{1,2},\&CenterDot;,n)---\left(5\right)$

Obtain the actual weights of network i attribute j.Wherein,

B for a front n attribute _ijMean value.

Network side is determined the best purpose network that switches

After obtaining the network attribute weights, weights actual in network and attribute use ordinal number preference method (TOPSIS) to determine the best purpose network that switches.The principle of ordinal number preference method be the gap between selected network and desirable solution minimum and with the disparity of poor solution.The Perfected process of determining is made of all possible optimum attributes value, and the poorest solution party's rule is made of the poorest all possible property value.

At first, use weights ω _ijAccording to formula: μ _ij=ω _ijB _ijTo standardized attribute b _ijBe weighted and obtain the weighting attribute.Determine respectively ideal network A according to weighted results according to following formula ⁺The poorest network A ^-:

A ⁺={(maxμ _ij|j∈J),(minμ _ij|j∈K)}(i=1,2,·,m) （6）

A ^-={(minμ _ij|j∈J),(maxμ _ij|j∈K)}(i＝1,2,·,m) （7）

Wherein, J is the set of benefit type attribute, and K is the set of cost type attribute.

Call formula (8) and (9) according to the weighting attribute and calculate each network and ideal network and the distance of poor network:

$s_i^{+}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&mu;}}_{\mathrm{ij}}-{\mathrm{\&mu;}}_j^{+})}^2}---\left(8\right)$

$s_i^{-}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&mu;}}_{\mathrm{ij}}-{\mathrm{\&mu;}}_j^{-})}^2}---\left(9\right)$

In formula,

The set of expression ideal network attribute,

Expression is the set of poor network attribute.

According to formula:

Calculate the relative degree of closeness of each network and ideal network.Choose relative degree of closeness The maximum network of value is the best purpose network that switches.I.e. choosing

The network that maximum is corresponding:

$A_{\mathrm{TOP}}^{*}=\arg \max c_i^{*},(i=\mathrm{1,2},\&CenterDot;,m)---\left(10\right)$

The adaptive switching controls of terminal velocity

Obtain best purpose network from network side after, secondary user's is considered to switch under what conditions with regard to needs.Because UMTS general overlay area is wide, WLAN often with the isolated island formal distribution in the UMTS overlay area, therefore, what need that emphasis solves is the come in and go out switching of WLAN both sides of secondary user's.

As shown in Figure 3, the end side secondary user's is determined handoff threshold according to network received signal intensity, and when the AP of secondary user's distance W lan network i distance was di, the received signal strength of wlan network was:

RSS _i=RSS(d _i)=K ₁-K ₂lg(d _i)+Ω(d _i) （11）

In formula, K ₁Expression transmission and receiving antenna gain, K ₂The expression environment property taken advantage of characteristic, Ω (d) expression zero-mean white Gaussian noise.

If RSS ₀=K ₁-K ₂Lg (r)+Ω (r) is the handoff threshold value, and r is the handoff threshold distance.Definition D _i=RSS _i-RSS ₀, when d=r, D=0 _iMake that Ф (N) is that the network of N sampled point (constantly) is selected,

So, as not X and Y in the same time, if Ф (X) ≠ Ф (Y) (X＜Y), mean X constantly network select Ф (X) and Y constantly network select Ф (Y) difference, illustrate once switching has occured at least.Distinguishingly, Ф (X) Ф (Y)〉the level switching occured in 0 expression; Vertical switching has occured in Ф (X) Ф (Y)＜0 expression.

Distinguishingly, when secondary user's enters wlan network from the UMTS network, as long as exist the received signal strength of wlan network i to satisfy RSS _i(N)〉RSS ₀+ h _y, just switch to wlan network from the UMTS network, when secondary user's is left from wlan network i, need RSS _i(N)＜RSS ₀-h _yJust can switch.Parameter h _yRepresent the sluggish level factor.

Terminal is switched the joint handoff control that can adopt based on the sluggish level factor and resident chronon.

If T _i(N) the expression time of staying of secondary user's in wlan network i, work as T _i(N)〉0 o'clock, expression RSS _iRSS ₀Duration; Work as T _i(N)＜0 o'clock, expression RSS _i＜RSS ₀Duration.

The joint handoff control program based on the sluggish level factor and resident chronon is:

As Ф (N-1) ≠ Ф (N), according to resident chronon and the sluggish level factor, switch when satisfying formula (13) terminal.

$\mathrm{\&alpha;}\frac{D_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{h_y}+\mathrm{\&beta;}\frac{T_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{t_{\mathrm{dw}}}<-\mathrm{\&delta;}---\left(13\right)$

In formula, D _{Ф (N1)}(N) be the poor of the signal strength signal intensity of N sampled point (constantly) network Ф (N-1) and handoff threshold value, T _{Φ (N-1)}(N) be the time of staying of N sampled point (constantly) network Ф (N-1), parametric t _dwRepresent resident chronon, parameter alpha represents sluggish level method controlling elements, and parameter beta represents resident timing method controlling elements, and δ is the dynamic decision thresholding.

Be illustrated in figure 3 as heterogeneous wireless network analysis of shift figure.Normal conditions are got α=β=1, below we provide the method for dynamically determining of parameter alpha, β, the δ of speed adaptive.Can adopt number adjusting method or index replacement method are determined parameter alpha, β.Number adjusting method is determined parameter alpha, β:

Work as d=d ⁺(d ⁺＜r) time, can obtain:

RSS(d ⁺)-RSS ₀＝K ₂lg(r)-K ₂lg(d ⁺)＝h _y （14）

Represent secondary user's speed with v, make d=r-vt _dw, desirable

$\mathrm{\&alpha;}=\frac{1g\left(d^{+}\right)}{1g|r-v\&CenterDot;t_{\mathrm{dw}}|},\mathrm{\&beta;}=1---\left(15\right)$

The index replacement method is determined the method for parameter alpha, β:

Work as d=d ^-(d ^-R) time, can obtain:

RSS ₀-RSS(d ^-)=K ₂log(d ^-)-K ₂log(r)=h _y （18）

If

Represent the performance separation speed of sluggish level method and resident timing method, adopt v/v _HORatio is adjusted the ratio of two kinds of methods.Due at v＜v _HOThe time, sluggish level method is than the poor performance of resident timing method, need to push the speed to change adjusting to sluggish level method and resident timing method ratio; And at v v _HOThe time, substantially take the performance of sluggish level method as main.

For velocity variations is become steadily to the adjusting of sluggish level method and resident timing method, to sum up analyze, desirable

$\mathrm{\&alpha;}=1+e^{-\frac{2v\&CenterDot;t_{\mathrm{dw}}}{d^{-}-d^{+}}},$ $\mathrm{\&beta;}=1-e^{-\frac{2v\&CenterDot;t_{\mathrm{dw}}}{d^{-}-d^{+}}}---\left(19\right)$

At last, the present invention adopts dynamic decision thresholding δ to adjust controlled condition:

$\mathrm{\&delta;}=\left\{\begin{array}{cc}1+\mathrm{\&theta;}& \mathrm{UMTS}\&RightArrow;\mathrm{WLAN}\\ 1-\mathrm{\&theta;}& \mathrm{WLAN}\&RightArrow;\mathrm{UMTS}\\ 1& \mathrm{ELSE}\end{array}\right.---\left(20\right)$

According to quality of service requirement, consider the requirement of transmitted data amount and outage probability in the present invention, generally get 0＜θ＜0.5.

The present invention uses MATLAB software to carry out emulation experiment.UMTS network and wlan network arrange respectively according to 3GPP and IEEE802.11 standard, and other simulation parameter is as follows: R=150m, d ⁺=120m, d ^-=140m, r=129.6m, sample frequency interval T=0.05s, resident chronon t _dw=5s, perception user's initial velocity v=30m/s.The total bandwidth of setting the UMTS network is 2Mbps, and the total bandwidth of wlan network is 11Mbps.Simulation result such as accompanying drawing Fig. 4, Fig. 5 and Fig. 6, wherein: AHP+TOPSIS is algorithm in document [3], SAE+TOPSIS is algorithm in document [2], MODM be document (Shi Wenxiao etc. based on the heterogeneous wireless network of multiobjective decision-making access selection algorithm) in algorithm, and the inventive method is improved one's methods with AHP+TOPSIS and is represented.Fig. 4, Fig. 5 and Fig. 6 show that the present invention reduces switching times and the outage probability of secondary user's effectively, has improved transmitted data amount.

Claims (9)

1. In an isomery cognitive radio networks based on the joint handoff method of speed adaptive, it is characterized in that, at network side, frequency spectrum broker entity carries out network according to the network attribute weights and selects based on cost type attribute and the benefit type property calculation network attribute weights of analytic hierarchy process (AHP) according to network; In end side, change according to resident chronon and the sluggish level factor based on terminal velocity and determine to switch, complete from former network and switch to objective network.
2. 2. joint handoff method according to claim 1, it is characterized in that, frequency spectrum broker entity computing network attribute weights specifically comprise: according to the accessible collection of network of isomery cognitive radio networks, and each network attribute set, set up the multiattribute matrix, the multiattribute matrix is carried out standardization, obtain the information entropy e of network attribute according to the standardized nature matrix _j, call formula according to the information entropy: ${\mathrm{\&eta;}}_i=\frac{1-e_j}{\underset{k=1}{\overset{n}{\mathrm{\&Sigma;}}}(1-e_k)},(j=\mathrm{1,2},\&CenterDot;,n)$ Obtain the entropy weights of j network attribute; Be ε according to the normalized weight of network _j(j=1,2 ..., n), call formula:

   ${\mathrm{\&omega;}}_{\mathrm{ij}}=\frac{{\mathrm{\&epsiv;}}_j\&CenterDot;{\mathrm{\&eta;}}_j\&CenterDot;e^{b_{\mathrm{ij}}-b_{\mathrm{ij}}^n}}{\underset{k=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&epsiv;}}_k\&CenterDot;{\mathrm{\&eta;}}_k\&CenterDot;e^{b_{\mathrm{ik}}-b_{\mathrm{ik}}^n}},(i=\mathrm{1,2},\&CenterDot;,m)(j=\mathrm{1,2},\&CenterDot;,n)$ Obtain the actual weights of j attribute of network i; According to formula: μ _ij=ω _ijB _ijStandardized value b to attribute _ijBe weighted and obtain the weighting attribute, according to each network of weighting property calculation and ideal network and the distance of poor network, calculate the relative degree of closeness of each network and ideal network, the network of choosing relative closeness value maximum is the best purpose network that switches, wherein

   

   Be a front n b _ijMean value, m is the network number, n is the attribute number of each network using.
3. 3. joint handoff method according to claim 1, is characterized in that, in end side, terminal adopts based on the joint handoff of the sluggish level factor and resident chronon and controls, determine to switch according to resident chronon and the sluggish level factor, when satisfying formula: $\mathrm{\&alpha;}\frac{D_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{h_y}+\mathrm{\&beta;}\frac{T_{\mathrm{\&Phi;}(N-1)}\left(N\right)}{t_{\mathrm{dw}}}<-\mathrm{\&delta;},$ Terminal is sent switching controls, wherein, and D _{Ф (N-1)}(N) be N the constantly signal strength signal intensity of network i and handoff threshold poor, T _{Φ (N-1)}(N) N the time of staying of secondary user's in network i in the moment of expression, t _dwRepresent resident chronon, α represents the sluggish level control ratio factor, and β represents resident timing controlled scale factor, and δ is decision threshold.
4. 4. joint handoff method according to claim 2, is characterized in that, described ideal network and the poorest network are specially: call formula:

   A ⁺={(maxμ _ij|j∈J)，(minμ _ij|j∈K)} (i＝1,2,·,m)

   A ^-={(minμ _ij|j∈J),(maxμ _ij|j∈K)} (i＝1,2,·,m)

   Determine respectively ideal network A ⁺The poorest network A ^-, wherein, J is the set of benefit type attribute, K is the set of cost type attribute.
5. 5. joint handoff method according to claim 2, is characterized in that, calls formula:

   $s_i^{+}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&mu;}}_{\mathrm{ij}}-{\mathrm{\&mu;}}_j^{+})}^2},$ $S_I^{-}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&mu;}}_{\mathrm{ij}}-{\mathrm{\&mu;}}_j^{-})}^2}$ Computing network i and ideal network and the distance of poor network

   

   With

   

   In formula, The set of expression ideal network attribute,

   

   Expression is the set of poor network attribute, according to formula: The relative degree of closeness of computing network and ideal network.
6. 6. joint handoff method according to claim 2, is characterized in that, described standardization is specially: when network attribute is cost type attribute, according to formula: $b_{\mathrm{ij}}=\frac{\min \left\{a_{\mathrm{ij}}\right|1\&le;i\&le;m\}}{a_{\mathrm{ij}}},(i=\mathrm{1,2},\&CenterDot;,m)$ To element a in the multiattribute matrix _ijCarry out standardization; Be benefit type attribute for network attribute, according to formula:

   $b_{\mathrm{ij}}=\frac{a_{\mathrm{ij}}}{\max \left\{a_{\mathrm{ij}}\right|1\&le;i\&le;m\}},(i=\mathrm{1,2},\&CenterDot;,m)$ To element a in the multiattribute matrix _ijCarry out standardization, obtain the standardized nature matrix.
7. 7. joint handoff method according to claim 3, is characterized in that, adopts number adjusting method according to formula:

   

   The sluggish level control ratio factor and resident timing controlled scale factor are determined in β=1.
8. 8. joint handoff method according to claim 3, is characterized in that, adopts the index replacement method according to formula:

   

   

   Determine the sluggish level control ratio factor and resident timing controlled scale factor.
9. 9. joint handoff method according to claim 3, is characterized in that, calls formula according to service quality: $\mathrm{\&delta;}=\left\{\begin{array}{cc}1+\mathrm{\&theta;}& \mathrm{UMTS}\&RightArrow;\mathrm{WLAN}\\ 1-\mathrm{\&theta;}& \mathrm{WLAN}\&RightArrow;\mathrm{UMTS}\\ 1& \mathrm{ELSE}\end{array}\right.$ Determine the handoff threshold δ between heterogeneous networks, wherein, 0＜θ＜0.5.

Images