CN103338471A - Service quality index evaluating method for wireless multi-hop network based on model - Google Patents

Abstract

The invention builds a two-dimension FSMC series queue model and applies the limited retransmission ARQ (Automatic Repeat Request) error recovery technology aiming at the situations that the end-to-end service quality is evaluated precisely under the wireless multi-hop network environment, the QoS (Quality of Service) index of a link circuit is dynamically estimated, the packet loss probability of the link circuit, time delay and routing requirement of available bandwidth are comprehensively considered, so that the computation of the packet loss probability is more precise, the queue state and the service state are combined, not only can the processing time delay be computed precisely, but also the sending time delay can be evaluated precisely, therefore, more accurate QoS index of the wireless network can be obtained. Based on the model, the service quality index evaluating method applies the weighted average QoS measurement to select an optimal path in a bandwidth available link circuit; three performance indexes, namely the end-to-end bandwidth, packet loss probability and the time delay are comprehensively considered; the method can be applied to a dynamic system, can find out a route hop-by-hop as the method is benefited from the excellent distributed characteristics of the model, so that only the expected available path can be utilized in the future, the computation in certain nodes of the network is reduced, and the complexity of finding a route is reduced.

Description

Wireless multi-hop network Service Quality Metrics evaluation method based on model

Technical field

The invention belongs to wireless communication field, relate in particular to a kind of under wireless network environment the accurate scene of estimating peer-to-peer qos parameter, consider professional feature, the wireless channel of arriving, and physical layer and link layer parameter are set up the tandem queue model, and develop the routing algorithm based on this model, guarantee the end-to-end QoS demand of wireless many network multimedia business.

Background technology

Wireless multi-hop network becomes the important component part of future broadband wireless communication systems gradually, comprises WiMAX metropolitan area network, wireless Mesh netword etc.In wired-wireless communication system, the two-forty of data and the demand of QoS are increasing substantially.With respect to cable network, wireless network not only resource (comprising bandwidth and power) is deficienter, and the temporal dispersion that causes of multipath fading, Doppler frequency shift, radio transmission etc. all can cause the performance of whole system to descend.Therefore, it is significant that research has wireless network analysis and the optimization problem of QoS restriction.

In the wireless multi-hop network, the QoS of radio multimedium business restriction makes route become one and has much challenging problem.A pith of any routing algorithm is the route discovery stage, will carry out link QoS index tolerance and resource reservation here, makes and satisfies the QoS demand.Typical end-to-end QoS tolerance (as end-to-end packet loss and end-to-end time delay) depends on the dynamic change of each node place quene state in the path and the state of link, thereby depend on professional arrival feature, physical layer and link layer design, and wireless channel state.

Characteristic complicated and changeable in view of present wireless network, to estimate exactly that the link quality index is a relatively problem of difficulty, especially general link QoS index calculating method is developed, and makes it to be applicable to the network of various different low layer technology, especially a difficult point.

Wireless network analysis and optimization problem at having the QoS restriction solve by cross-level analysis and modeling, not only can study systematic function more all sidedly, and applicability are strong.Prior art is as follows:

1) be the unit interval with a time frame in the link, under piece fading channel condition, when channel status was separate between each time frame, people such as LongLe had proposed a kind of one dimension FSMC tandem queue model based on quene state.This model is considered on the arrival process, the many speed rates on the physical layer, link layer of the bernoulli traffic in batches and is recovered based on the mistake of ideally unlimited re-transmission ARQ agreement.Based on this model, can estimate every jumping average packet loss ratio and processing delay (comprising queuing delay and retransmission delay time).Thereby, this one dimension FSMC tandem queue model is applied in the QoS route discovery algorithm.

Although this method adopts the method for cross-level analysis to set up a kind of tandem queue model, can estimate average packet loss ratio and processing delay, also there are some problems, comprise two aspects.(1) this model has adopted and has recovered based on the mistake of ideally unlimited re-transmission ARQ agreement on the link layer, yet is to adopt limited re-transmission ARQ in the reality, thereby, all with actual conditions gap is being arranged aspect the end-to-end packet loss of estimating and the end-to-end time delay; (2) owing to just be to consider object with the quene state aspect modeling, ignored Link State, thereby can only estimate processing delay at aspect of performance.Yet in the reality, transmission delay (propagation delay time) also is (the ignoring propagation delay) of can not ignore.

2) in the wireless single-hop network, consider that Adaptive Modulation and Coding (AMC) and limited re-transmission ARQ set up the two-dimentional FSMC queuing model based on quene state and service state associating, and carry out QoS performance cross-level analysis and design by model;

But the method only is applicable to the wireless single-hop network, can not be used for wireless multi-hop network.

Summary of the invention

Technical problem: the present invention is directed to the end-to-end QoS problem in the wireless multi-hop network, adopt the method for cross-level analysis, propose a kind of wireless multi-hop network Service Quality Metrics evaluation method based on model, be about to the two-dimentional FSMC tandem queue model of quene state and service state associating.Based on this model, can estimate average packet loss ratio and time delay (comprise processing delay and transmission delay,, we ignore propagation delay) here.Because this model has good distributed nature, can estimate the typical performance of every link hop-by-hop.Thereby, can apply it in the QoS routing algorithm.

Technical scheme: the inventive method is with the limited re-transmission ARQ mechanism associating on the AMC technology in the wireless network physical layer and the data link layer, and take all factors into consideration quene state and the service state at radio node place, set up a kind of improved two-dimentional FSMC tandem queue model, pass through Mathematical Modeling, obtain the steady-state distribution of state-transition matrix and the matrix of system, and then calculate the end-to-end time delay of system by this steady-state distribution, QoS indexs such as packet loss, this model can also accurately draw chain-circuit time delay (comprising processing delay and transmission delay) in accurate estimating bag-losing ratio; Based on above model, develop a kind of routing algorithm, to be taken all factors into consideration by dynamic end-to-end time delay, packet loss and link available bandwidth that model draws, but the link of selecting to have less weighted average QoS tolerance in the wide line link of multi-ribbon has reduced the complexity of seeking route effectively as optimal path.

This method comprises:

Process one: at the wireless multi-hop network of the automatic repeat requests ARQ error recovery technique that retransmits based on limited number of time of the Adaptive Modulation and Coding AMC technology with physical layer and link layer, the service state of the quene state of transmission data and system is set up a kind of based on the two-dimensional finite state Markov Chain FSMC model of state to characterizing in the associating link, channel status transition probability and the formation average arrival rate used in this model obtain the system mode transfer matrix, estimate packet loss and the time delay of link by the steady-state distribution of state-transition matrix;

Process two: the model that route discovery stage application process one is set up, estimate service quality QoS indexs such as packet loss on each node/link and time delay hop-by-hop, and the QoS index that calculates is brought in the weighted average QoS tolerance of definition, in the feasible path of bandwidth, select to have the path of minimum degree value as optimal path.

Automatic repeat requests ARQ error recovery technique asks transmit leg to retransmit the data message of makeing mistakes by the recipient, recovers wrong message, and the detailed process of estimation link packet drop rate and time delay is: the quene state U of transmission data in the associating t moment link _tService state C with system _tSet up a kind of based on state to (U _t, C _t) the two-dimensional finite state Markov Chain FSMC model that characterizes, by formula (1) computing node corresponding queues and the road combined state-transition matrix of chain:

$P_{(u,c),(v,d)}^{\left(k\right)}=P_{d|c}^{\left(k\right)}{P}^{\left(k\right)}(U_t=v|U_{t-1}=u,C_t=c)---\left(1\right)$

Wherein

Represent that a general system mode transition probability is that formation k is in state u constantly at t-1, and link k is in state c constantly at t, transfer to formation k and be in state v constantly at t, and link k is in the united state transition probability of state d constantly at t+1; State u and state v are packet number in the formation, state c and state d by link energy data packets for transmission number; Transferred to the probability of state d by state c for channel k; Calculated the steady-state distribution of corresponding united state transfer matrix by formula (2):

$\left\{\begin{array}{cc}{\mathrm{\π}}_{(U,C)}^{\left(k\right)}={\mathrm{\π}}_{(U,C)}^{\left(k\right)}{P}_{(U,C)}^{\left(k\right)}& \\ \underset{u\∈U,c\∈C}{\mathrm{\Σ}}{\mathrm{\π}}_{(u,c)}^{\left(k\right)}=1& \end{array}\right.---\left(2\right)$

Wherein

Be the associating steady-state distribution probability of formation k and link k,

Association system state transition probability for formation k and link k; U, C are respectively corresponding quene state and link service state, can calculate corresponding performance metric based on this steady-state distribution, are calculated total packet loss p of formation k by formula (3) ^(k):

p ^(k)=p ^′(k)+(1-p ^′(k))p ₀=1-(1-p ^′(k))(1-p ₀) （3）

Wherein, p ^{' (k)}Be the average overflow probability of formation k, p ₀Be the average Packet Error Ratio under the various transmission modes; When calculating by link k by formula (4), the average delay D of each packet ^(k):

$D^{\left(k\right)}=\frac{N^{\left(k\right)}}{E\left(A^{\left(k\right)}\right)(1-p^{\′\left(k\right)})}---\left(4\right)$

N wherein ^(k)Be the number of data packets sum that just is being transmitted among the average data bag number among the formation k and the link k, A ^(k)Be the arrival rate of formation k, E is for being averaging; Thereby data can be drawn by formula (5) and (6) through end-to-end total packet loss P and the overall delay D of L bar link:

$P\≈1-\underset{k=1}{\overset{L}{\mathrm{\Π}}}(1-p^{\left(k\right)})---\left(5\right)$

$D=\underset{k=1}{\overset{L}{\mathrm{\Σ}}}{D}^{\left(k\right)}---\left(6\right).$

The described route discovery stage is: at first select broadcast node by source node, form broadcasting group BG, select link among the BG and between the source node that the node of enough bandwidth is arranged then, but its link is designated as the bandwidth line link, follow the model that application process one is set up, estimate QoS indexs such as packet loss on each node/link and time delay hop-by-hop, and the QoS index that calculates is brought in the weighted average QoS tolerance of definition, the path of selecting to have the minimum degree value in the feasible path of bandwidth is as optimal path, thus the guarantee laser propagation effect.

Beneficial effect:

1) the inventive method can be estimated QoS indexs such as the average delay of Radio Link and packet loss more accurately than prior art, thereby helps to solve better wireless network performance analysis and the optimization problem with QoS restriction.

2) the inventive method can be applied in the QoS routing algorithm, helps multimedia service to select the more excellent path of QoS index.Because this model has good distributed nature, but route is found on hop-by-hop ground, has only the feasible path of expectation just can be utilized in the future.So just reduce the complexity of seeking route, reduced the calculating that some node (if introductory path is non-feasible path) is located in the network.

Description of drawings

Fig. 1 is the schematic diagram of setting up of the present invention's two dimension FSMC tandem queue model,

Fig. 2 is the routing algorithm flow chart that the present invention is based on model,

Fig. 3 one-dimensional model and two dimensional model performance comparison (packet loss),

Fig. 4 one-dimensional model and two dimensional model performance comparison (time delay),

The optimum routing performance of Fig. 5 is (packet loss) relatively,

The optimum routing performance of Fig. 6 is (time delay) relatively.

Embodiment

This method comprises:

Process one: the error recovery technique that retransmits based on limited number of time at the Adaptive Modulation and Coding AMC technology with physical layer and link layer is the wireless multi-hop network of automatic repeat requests ARQ, set up a kind of based on the two-dimensional finite state Markov Chain FSMC model of state to characterizing according to the quene state of transmission data in the link and the service state of system by supposing parameters such as professional arrival rate and queue length, channel status transition probability and the formation average arrival rate used in this model obtain the system mode transfer matrix, estimate packet loss and the time delay of link by the steady-state distribution of state-transition matrix.

Process two: the model that route discovery stage application process one is set up, estimate QoS indexs such as packet loss on each node/link and time delay hop-by-hop, and the QoS index that calculates is brought in the weighted average QoS tolerance of definition, in the feasible path of bandwidth, select to have the path of minimum degree value as optimal path.

ARQ is one of error correcting agreement of data link layer.Data message by the recipient asks transmit leg to retransmit to make mistakes recovers wrong message, be in the communication system for the treatment of one of method of mistake that channel brings, be also sometimes referred to as afterwards to error correction.The present invention adopts limited selectivity to retransmit ARQ, according to the transmission result of every frame, jumps at each, and receiving terminal all can be confirmed ACK bag or NACK unconfirmed bag to one of transmitting terminal feedback for each packet that is transmitted, and the indication message is correctly received or transmits and make mistakes.If transmitting terminal receives a NACK bag, then retransmit the bag of makeing mistakes, and set maximum retransmission.

The foundation of two dimension FSMC tandem queue model and the process of estimating corresponding QoS index are: in wireless network, consider the arrival process of the bernoulli traffic in batches, there is a time-limited formation at each node place, serves with the pattern of first in first out.According to the state of Radio Link, sending node adopts corresponding AMC technology.In data link layer, adopt based on the mistake of limited re-transmission ARQ agreement and recover.Quene state, the service state in this moment, the business in this moment that the state of each formation constantly depends on previous moment arrive situation, and queue length, the FSMC queuing model that foundation characterizes (quene state and service state) based on state, drawn the state-transition matrix of system by corresponding formation arrival rate and service speed, and then obtain the steady-state distribution of state-transition matrix, calculate parameters such as the end-to-end time delay of link and packet loss by this steady-state distribution.Owing to adopted limited re-transmission ARQ and unite and considered quene state and service state, so this model is more accurate for the estimation of packet loss and time delay (processing delay and transmission delay) end to end.

Routing algorithm specific implementation method based on model is: in the QoS route discovery algorithm, source node is broadcasted routing request packet (RRQ) to seek the path of satisfying the QoS demand that arrives destination node in network.Request bag RRQ comprises address, request bag ID number and the end-to-end QoS demand of source node and destination node.When intermediate node is received RRQ when bag, its check and source node between available bandwidth on the link, only have enough bandwidth and hold the link of new connection and could be considered later on that these links are called bandwidth feasible (BW-feasible) link.Source node is that every BW-feasible link calculates link average delay and packet loss, for the link that satisfies the QoS demand, calculate the arrival rate of forward direction next node, information such as arrival rate, link average delay and packet loss are recorded in the RRQ packet header, and the RRQ bag are forwarded to the receiving node of respective link.But the receiving node of these line links is formed a set of node, is called broadcasting group (BG).The weighted average QoS tolerance that is brought into definition for the link packet drop rate selecting an optimal path, use to estimate in the above model and time delay (as shown in the formula) in, select that littler link of corresponding value as optimal path.

$\mathrm{metric}\left(R\right)=\mathrm{\α}\frac{\mathrm{delay}\left(R\right)}{P_d}+(1-\mathrm{\α})\frac{\mathrm{loss}\left(R\right)}{P_l}$

Wherein metric (R) is defined weighted average QoS tolerance, and delay (R) and loss (R) are respectively end-to-end time delay and the end-to-end packet loss of path R, and α depends on the relative significance level of end-to-end time delay and packet loss.P _lBe the end-to-end packet loss of business demand, P _dBe the business demand end-to-end time delay.Routing algorithm of the present invention can be considered end-to-end bandwidth, packet loss and three performance index comprehensives of time delay, and can be applicable to dynamical system.

In an actual wireless network with 10 nodes, send data by source node 1 to destination node 9, use method of the present invention, concrete process of transmitting is: at first submit business information (as professional arrival rate) and end-to-end QoS demand by service route source node 1, select then and source node between have enough bandwidth next-hop node (3,5,7), but respective link is designated as bandwidth line link (L ₃, L ₅, L ₇), the model that application the present invention sets up calculates the average packet loss ratio (P of every feasible link of bandwidth ₃, P ₅, P ₇) and time delay (D ₃, D ₅, D ₇), for three links that satisfy professional end-to-end QoS demand, the packet loss of every link and time delay are brought in the weighted average QoS tolerance of the present invention's definition, draw corresponding metric size (metric (5)＜metric (7)＜metric (3)), select to have the L of less metric ₅As optimal path, repeat above-mentioned routing process at each hop node, arrive destination node 9 up to packet, and then the optimal path 1-that obtains transmitting 5-10-8-2-6-9.

This method mainly comprises two parts: the foundation of the two-dimentional FSMC tandem queue model of associating quene state and service state, use based on the routing algorithm of model.

At first, with a kind of general pattern the physical layer of wireless network is carried out modeling, make this queuing model be applicable to various physical-layer techniques.Suppose limited physics orthogonal channel, distinguish with spreading code, perhaps distinguish with frequency domain.Transmission time on every orthogonal channel is divided into the fixing one or more time slots of time span, and they are only shared by a link or many different links being closed on, calculates in the l jumping probability that can transmit i packet in the time frame.Then consider on the arrival process, the many speed rates on the physical layer, link layer of the bernoulli traffic in batches and recover based on the mistake of limited re-transmission ARQ agreement, set up a discrete time tandem queue model, based on a certain moment quene state, service state, professional arrival rate and queue length are got up quene state and service state joint to the finite state Markov Chain that characterizes with state with one.Obtain the state-transition matrix of system by the state transition probability of computing system, the steady-state distribution of this matrix exists and is unique, and steady-state distribution further calculates overall delay and total packet loss end to end thus.

In addition, in the routing algorithm based on model, at first in network, broadcast routing request packet to seek the path of satisfying the QoS demand that arrives destination node by source node.When intermediate node is received routing request packet, available bandwidth between its inspection and the source node on the link, only having enough bandwidth holds the link of new connection and could continue be considered, for the feasible link of these bandwidth, the weighted average QoS that uses the present invention's definition measures to compare, wherein use the end-to-end overall delay and the total packet loss that go out according to model assessment, when the metric of a link during less than another link, less that link of selectance value guarantees that with this source node is to the optimal transmission quality of terminal node as optimal path.Routing algorithm of the present invention can be considered end-to-end bandwidth, packet loss and three performance index comprehensives of time delay, and can be applicable to dynamical system.

Simulation result of the present invention

Emulation experiment is finished at the Linux+NS-2.29 platform.Suppose the professional Bernoulli Jacob's arrival process in batches of obeying, the probability that arrives 0,1,2,3,4 packet of source node formation in each time frame is respectively 1-25A/48, A/4, A/8, A/12, A/16, and A is average arrival rate (establishing A=1.8); Queue length is made as identical, is expressed as QL; On every link in the series system, for each time frame distributes a time slot, channel status is separate between the adjacent time frame, average Signal to Interference plus Noise Ratio

Be 15, adopt the Nakagami channel, and Nakagami parameter m=1; For service speed, suppose under the transmission mode n that every frame can transmit n packet; Maximum retransmission is 4 times.In the soundness verification part of model, we choose 6 series connection nodes (i.e. 6 formations), and Packet Error Ratio is made as 0.2.

At the experimental section of QoS route, we set up a random topology with 10 nodes, and area size is 1000m * 1000m, if the distance between two nodes less than 300m, they can intercom (namely having a link) mutually so.Here, we ignore the mobility of node.Each node uses fixing transmitted power, and link (i, the average Signal to Interference plus Noise Ratio of receiving terminal j) is defined as

Wherein, P is that an index that comprises factors such as transmitted power and antenna gain (is established P=9 * 10 ⁷); σ (i, j) represent link (i, the influence of average environment noise/interference j), σ (2,10)=σ (10,2)=0.9 wherein, σ (2,9)=σ (9,2)=0.5, other all give tacit consent to σ=1 except specifying; (i, the j) distance between representation node i and the node j suppose that the path fading index is 3 to d.Here we adopt single Business Stream, suppose that Packet Error Ratio is 0.1, and maximum retransmission is 4 times.

The two-dimentional queuing model that the present invention proposes is compared with the one-dimensional model of existing method, and advantage applies is aspect two.On the one hand, adopt limited number of time re-transmission ARQ mechanism in the reality at link layer; On the other hand, can analyze the behaviour of systems better based on the two-dimentional queuing model of quene state and service state associating, not only can estimate processing delay, and can estimate transmission delay.Thereby, in the estimation of packet loss and end-to-end time delay, all than one-dimensional model more accurately (here with simulation value as actual value), be embodied in respectively among Fig. 3 and Fig. 4.

Based on the route discovery algorithm of queuing model, when selecting optimum route, path metric is about the QoS performance.Here, and we and minimum hop count route (MHR, MinimalHopRouting) algorithm compares, and source node and destination node are respectively node 1 and node 9 in the topological diagram.Use when carrying out routing based on the route discovery algorithm of queuing model of the present invention, get α=0 or α=0.5 or α=1 respectively, the optimal path that obtains is 1-〉5-〉10-〉8-〉2-〉6-〉9; When carrying out routing with the minimum hop count routing algorithm, the transmission path that obtains is 1-〉5-〉10-〉2-〉9(MHR path 1) or 1-5-8-2-9(MHR path 2).Wherein, in the MHR path 2 because link 5-8 poor quality (namely

Be lower than certain minimum threshold), actual can not communicating of most of the time in the emulation.The simulation result in MHR path 1 is seen Fig. 5 and Fig. 6.Obviously, at the end-to-end QoS aspect of performance, algorithm of the present invention better (packet loss and end-to-end time delay are all lower), thus can satisfy professional QoS demand better.

Claims (3)

1. wireless multi-hop network Service Quality Metrics evaluation method based on model is characterized in that this method comprises:

Process one: at the wireless multi-hop network of the automatic repeat requests ARQ error recovery technique that retransmits based on limited number of time of the Adaptive Modulation and Coding AMC technology with physical layer and link layer, the service state of the quene state of transmission data and system is set up a kind of based on the two-dimensional finite state Markov Chain FSMC model of state to characterizing in the associating link, channel status transition probability and the formation average arrival rate used in this model obtain the system mode transfer matrix, estimate packet loss and the time delay of link by the steady-state distribution of state-transition matrix;

Process two: the model that route discovery stage application process one is set up, estimate service quality QoS indexs such as packet loss on each node/link and time delay hop-by-hop, and the QoS index that calculates is brought in the weighted average QoS tolerance of definition, in the feasible path of bandwidth, select to have the path of minimum degree value as optimal path.

2. the wireless multi-hop network Service Quality Metrics evaluation method based on model as claimed in claim 1, it is characterized in that: repeat requests ARQ error recovery technique asks transmit leg to retransmit the data message of makeing mistakes by the recipient automatically, recover wrong message, the detailed process of estimation link packet drop rate and time delay is: the quene state U of transmission data in the associating t moment link _tService state C with system _tSet up a kind of based on state to (U _t, C _t) the two-dimensional finite state Markov Chain FSMC model that characterizes, by formula (1) computing node corresponding queues and the road combined state-transition matrix of chain:

$P_{(u,c),(v,d)}^{\left(k\right)}=P_{d|c}^{\left(k\right)}{P}^{\left(k\right)}(U_t=v|U_{t-1}=u,C_t=c)---\left(1\right)$

Wherein Represent that a general system mode transition probability is that formation k is in state u constantly at t-1, and link k is in state c constantly at t, transfer to formation k and be in state v constantly at t, and link k is in the united state transition probability of state d constantly at t+1; State u and state v are packet number in the formation, state c and state d by link energy data packets for transmission number;

Transferred to the probability of state d by state c for channel k; Calculated the steady-state distribution of corresponding united state transfer matrix by formula (2):

$\left\{\begin{array}{cc}{\mathrm{\π}}_{(U,C)}^{\left(k\right)}={\mathrm{\π}}_{(U,C)}^{\left(k\right)}{P}_{(U,C)}^{\left(k\right)}& \\ \underset{u\∈U,c\∈C}{\mathrm{\Σ}}{\mathrm{\π}}_{(u,c)}^{\left(k\right)}=1& \end{array}\right.---\left(2\right)$

Wherein Be the associating steady-state distribution probability of formation k and link k, Association system state transition probability for formation k and link k; U, C are respectively corresponding quene state and link service state, can calculate corresponding performance metric based on this steady-state distribution, are calculated total packet loss p of formation k by formula (3) ^(k):

p ^(k)=p ^′(k)+(1-p ^′(k))p ₀=1-(1-p ^′(k))(1-p ₀) （3）

Wherein, p ^{' (k)}Be the average overflow probability of formation k, p ₀Be the average Packet Error Ratio under the various transmission modes; When calculating by link k by formula (4), the average delay D of each packet ^(k):

$D^{\left(k\right)}=\frac{N^{\left(k\right)}}{E\left(A^{\left(k\right)}\right)(1-p^{\′\left(k\right)})}---\left(4\right)$

N wherein ^(k)Be the number of data packets sum that just is being transmitted among the average data bag number among the formation k and the link k, A ^(k)Be the arrival rate of formation k, E is for being averaging; Thereby data can be drawn by formula (5) and (6) through end-to-end total packet loss P and the overall delay D of L bar link:

$P\≈1-\underset{k=1}{\overset{L}{\mathrm{\Π}}}(1-p^{\left(k\right)})---\left(5\right)$

$D=\underset{k=1}{\overset{L}{\mathrm{\Σ}}}{D}^{\left(k\right)}---\left(6\right).$

3. the wireless multi-hop network Service Quality Metrics evaluation method based on model as claimed in claim 1, it is characterized in that: the described route discovery stage is: at first select broadcast node by source node, form broadcasting group BG, select link among the BG and between the source node that the node of enough bandwidth is arranged then, but its link is designated as the bandwidth line link, follow the model that application process one is set up, estimate QoS indexs such as packet loss on each node/link and time delay hop-by-hop, and the QoS index that calculates is brought in the weighted average QoS tolerance of definition, the path of selecting to have the minimum degree value in the feasible path of bandwidth is as optimal path, thus the guarantee laser propagation effect.

Images