CN105743562A - Satellite network access method based on predicted dynamic bandwidth allocation - Google Patents

Abstract

The present invention provides a satellite network access method based on predicted dynamic bandwidth allocation. A TCP protocol 3-handshake method is used to enable free access of users; a service request is sent to a satellite in a reserving manner; the satellite sets different priorities for the users according to reserved time slots; initial allocation is performed in a manner of combining need-based allocation and free allocation; after the initial allocation, a traffic prediction method is used to predict traffic of a next frame; and then, the satellite adjusts a bandwidth allocation scheme in time according to a prediction result of a last frame. A dynamic bandwidth allocation scheme is implemented, and the utilization rate of resources of a satellite system is increased.

Description

Satellite network cut-in method based on prediction Dynamic Bandwidth Allocation

Technical field

The present invention relates to a kind of satellite network cut-in method based on prediction Dynamic Bandwidth Allocation, utilize the thought of chaos volume forecasting to realize satellite Dynamic Bandwidth Allocation, improve channel utilization, belong to technical field of satellite communication.

Background technology

Satellite communication is compared to other communication mode, there is the advantages such as communication distance remote (such as utilizing fixed statellite to communicate, its communication distance is maximum can reach 18000km), area coverage is big, communication band width, transmission capacity are big, maneuverability, communication line are reliable and stable.Currently, wireless communication technology is maked rapid progress, the technology such as satellite communication, mobile communication is in the high speed development stage, low speed data that previous generation satellite network can carry and voice service can not meet the demand of this continuous expansion, and a new generation's satellite broadband communication becomes the basic application process solving this problem.

Broadband satellite network system is compared with a lot of satellite communication, the most prominent advantage is in that a satellite repeater can provide real-time access for numerous users interior on a large scale, carry out the communication of point to multi--point, how so multiple users should share limited bandwidth, reasonably utilize the communication resource, here it is the problem that wideband satellite communication must take into.Simultaneously, modern communication networks data traffic transmission speed is higher, and trafficwise also variation gradually, when limited upstream and downstream bandwidth, power and time delay, how providing better service for more users in broadband satellite network system, this is the problem that must solve.In the practical application of broadband satellite network system, when user sends data, how to send request, adopts which kind of efficient frame structure, and how satellite distributes the reasonable access that request data ensures user, be all access protocol it is also contemplated that problem.Nowadays, network communicating system popularization, it is achieved mode, also in continuous complication, is only constantly improved user's multiple access protocol and could be realized communication process efficiently fair in application.

Summary of the invention

For solving prior art Problems existing, the present invention provides a kind of satellite network cut-in method based on prediction Dynamic Bandwidth Allocation, utilizes the thought of chaos volume forecasting to realize satellite Dynamic Bandwidth Allocation, improves channel utilization.

In the present invention, user adopts the method freely accessed to access satellite, the mode adopting reservation sends service request to satellite, user is divided into different priority according to reserving time slots number by satellite, adopting distributes according to need and freely distribute the mode combined carries out original allocation, and next frame flow is predicted by the method adopting volume forecasting after original allocation, hereafter, satellite predicts the outcome according to previous frame and adjusts bandwidth allocation scheme in time.By accessing, preengage, distribute, predicting, four steps realize the multiple access protocol of satellite network to the present invention, improve the resource utilization of whole satellite system.

Satellite network cut-in method based on prediction Dynamic Bandwidth Allocation provided by the invention, comprises the following steps:

Step 1: satellite broadcasts oneself identification number ID in space, terminal by receive satellite identification number id information and ON status information return to respective satellite, ON status information represents that active user's terminal has packet to produce；

Step 2: after satellite receives the information that terminal is sent, detects identification number id information therein, if identification number ID mates with the identification number ID of oneself, then satellite distribution request time slot is to terminal, and request gap information is issued terminal；

Step 3: after terminal is successfully accessed, sends request to satellite, comprises Subscriber Queue length, i.e. terminal queuing packet count in solicited message；

Step 4: satellite arranges priority number from big to small according to terminal queuing packet count, and inserts reservation table and free allocation table；

Step 5: satellite carries out time slot and initializes distribution, i.e. after terminal accesses satellite first and sends reserve requests, satellite is allocated according to each terminal queuing packet count；

Whether remaining time slots is more than 0 particularly as follows: satellite calculates after pressing reservation table distribution to initialize distribution, if it is, according to priority distribute time slot for reservation table, for free allocation table poll distribution remaining time slots after distributing；Otherwise, being reservation table distribution time slot according to weight assignment formula, weighted formula is:

${\mathrm{allocate}}_i=\frac{data\_num}{\underset{i=1}{\overset{m}{\Σ}}req\_{\mathrm{num}}_i}req\_{\mathrm{num}}_i,$

Wherein, allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresenting terminal i reservation data timeslot number, m represents the total number of terminals currently having time slot request；

Step 6: predict the terminal timeslot number of next frame while initializing distribution, when next frame arrives satellite, the satellite adjustment time slot distribution that predicts the outcome according to previous frame, and each frame later repeats.

Time slot distribution Adjusted Option particularly as follows:

Initialize the distribution time slot terminal more than or equal to prediction time slot for all, carry out configuration as follows:

$\left\{\begin{array}{c}{\mathrm{allocate}}_i={\mathrm{pro}}_i\\ free=\underset{i=1}{\overset{m}{\Σ}}{\mathrm{allocate}}_i-{\mathrm{pro}}_i\end{array}\right.,$

Then, total free timeslot number free being re-assigned to terminal again that initialize distribution time slot less than prediction time slot, the quantity of concrete distribution is:

${\mathrm{allocate}}_i={\mathrm{allocate}}_i+free\×\frac{{\mathrm{pro}}_i-{\mathrm{allocate}}_i}{\underset{i=1}{\overset{m}{\Σ}}{\mathrm{pro}}_i-{\mathrm{allocate}}_i},$

Wherein, allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresent terminal i reservation data timeslot number, pro_iRepresenting the timeslot number of prediction terminal i next frame, free represents total free timeslot number, and m represents the total number of terminals currently having time slot request.

The method have the advantages that the satellite network multiple access protocol of classics major downside is that the stationarity of Time Slot Occupancy, for sudden very strong service traffics, it is possible to time slot waste can be caused, reduce channel utilization.The Dynamic Bandwidth Allocation cut-in method based on prediction that the present invention proposes effectively improves this this problem, when after user terminal application to bandwidth, do not retransmit service request time slot, and utilize the bandwidth being assigned to continue to send data, satellite end adds chaos time sequence method for predicting, after initial allocation, next frame flow is predicted, satellite adjusts bandwidth allocation scheme in time according to predicting the outcome, this addresses the problem Time Slot Occupancy stationarity problem, improve the resource utilization of satellite network.Utilize mixing and demand assigned mode when original allocation, and add prioritization, preferential to high priority terminal distribution time slot, it is ensured that the reasonability of resource distribution and fairness.

Accompanying drawing explanation

Fig. 1 is low-track satellite network system assumption diagram；

Fig. 2 is the uplink downlink frame structure of cut-in method provided by the invention；

Fig. 3 is flow chart of the present invention.

Detailed description of the invention

The research of the multiple access protocol of broadband low-track satellite network is concentrated mainly on the Access Algorithm of satellite and user side, provide its study model, as shown in Figure 1, it is made up of satellite (ST) and user terminal (UT), satellite (ST) primary responsibility processes the business task of ICBM SHF satellite terminal, and each user terminal is connected on satellite network by the ICBM SHF satellite terminal specified.

Chaos flow traffic forecast model:

Network data flow has obvious self-similarity nature, and self similarity is also a key property of chaos, can be good at the practical business flow in prediction network hence with Chaotic time series forecasting model.The whole forecast model that the present invention adopts builds and is divided into three phases:

First stage is phase space reconfiguration, adopts the method that coordinate postpones, by One-dimension Time Series b_NThe different phase spaces postponing to construct real system, it will be assumed that m is Embedded dimensions, and τ is time delay, and current time series is, this sequence can be expressed as b₁,b₂…b_N, this sequence can represent that m ties up N-(m-1) τ phase point in phase space, then i-th phase point can be obtained by formula (1).

X_i={ b_i,b_i+τ…b_i+(m-1)τ, i=1,2 ... N-(m-1) τ (1)

It is considered herein that τ and m connects each other, and provide embedding time window τ_wThe concept of=(m-1) τ, the concept that the present invention is constant according to embedding time window, adopt and be uniformly embedded into mode, and choose m=2, regulate the τ parameter as phase space reconfiguration, when this method can simplify phase space reconfiguration, parameter chooses process.

Second stage is the calculating of maximum Lyapunov exponent, initial value is had extremely sensitive perceiving property by chaos system, even if therefore two initial value differences are only small, but the track that As time goes on the two initial value produces also there will be the state exponentially separated, and Lyapunov index is through quantitative mode to describe this phenomenon, the average index rate that Lyapunov index representative system is dispersed between adjacent orbit in phase space or restrained.And in the index of all Lyapunov, as long as there being the value more than zero to occur, then whole system will be all chaos.As long as therefore calculating the value of maximum Lyapunov exponent in actual applications, and namely can determine whether system whether chaos by observing the value that calculates.

A kind of wheeled mobile robots algorithm can be adopted to calculate maximum Lyapunov exponent, and this algorithm amount of calculation and memory space are less, are beneficial to real-time estimate.Meet constraints (m-1) τ < | j-i | < α (m-1) τ, under the premise of 1 < α < 10, try to achieve X according to formula (2)_iNearest neighbor point X_{i_near}.Wherein j < N-(m-1) τ, | | X_j-X_i| | represent X_jAnd X_iBetween Euclidean distance.< | j-i | ensure that X to constraints (m-1) τ_iWith its consecutive points X_{i_near}On different tracks, | j-i | < (m-1) τ then ensure that the number of times of calculating Lyapunov index is within rational scope, reduces the load of system.

$\left|\right|X_{i\_near}-X_i\left|\right|=\underset{j}{min}\left|\right|X_j-X_i\left|\right|---\left(2\right)$

Thus, Xi Lyapunov index on the maximum direction of extension can be obtained when setting constraints 1 < h < N-(m-1) τ-i by formula (3).

$L\left(i\right)=\max \left\{\frac{1}{h}\ln \frac{\left|\right|X_{i\_\mathrm{near}+h}-X_{i+h}\left|\right|}{\left|\right|\underset{h}{X_{i\_\mathrm{near}}}-X_i\left|\right|}\right\}---\left(3\right)$

The average maximum Lyapunov exponent value of phase point is obtained again through formula (4).

$L=\frac{1}{N-(m-1)\mathrm{\&tau;}}\underset{i=1}{\overset{N-(m-1)\mathrm{\&tau;}}{\&Sigma;}}L\left(i\right)---\left(4\right)$

Three phases is information source prediction, the deduction in the first two stage can obtain, and in current phase space, last phase point is obtained by formula (5).

X_N-(m-1)τ={ b_N-(m-1)τ,b_N-(m-2)τ...b_N}(5)

Derive can obtain through a step

X_N+1-(m-1)τ={ b_N+1-(m-1)τ,b_N+1-(m-2)τ...b_N+1}(6)

By X_N-(m-1)τNearest neighbor point be X_near, it can be seen that X_N+1-(m-1)τNearest neighbor point be X_near+1, again because Lyapunov index is the exponential growth rate of a step conversion in space, so obtaining formula (7).

$e^{\mathrm{\&lambda;}}=\frac{\left|\right|X_{N+1-(m+1)\mathrm{\&tau;}}-X_{N-(m+1)\mathrm{\&tau;}}\left|\right|}{\left|\right|X_{near+1}-X_{near}\left|\right|}---\left(7\right)$

Thus can pass through top n time series b₁,b₂…b_NTry to achieve the N+1 seasonal effect in time series value.

The present invention proposes the satellite network access protocol based on prediction Dynamic Bandwidth Allocation, as it is shown on figure 3, mainly include user's access, user's reserve requests, satellite distribution time slot and satellite predictions adjustment time slot to distribute this Four processes, concrete execution step is as follows:

First step 1, satellite broadcast the identification number ID of oneself in space, before user needs to access respective satellite transmission data, can first receive the information of these broadcast, then the id information received and ON status information being loaded and confirm to return to respective satellite in character ACK bag, ON status information represents that active user's terminal has packet to produce.

Step 2, satellite detect id information therein after receiving the ACK bag that user sends, if ID mates with the ID of oneself, then satellite distribution request time slot is to user, and request gap information is contained in ACK bag and mails to user.

After step 3, user are successfully accessed, sending request to satellite, comprise Subscriber Queue length, i.e. terminal queuing packet count in solicited message, solicited message frame structure is as shown in Figure 2.

Step 4, satellite arrange priority number from big to small according to terminal queuing packet count, and insert reservation table and free allocation table.Reservation table is the table being stored in satellite scheduler, saves the ID preengage and reserving time slots number；Free allocation table saves the ID of all any active ues in satellite network.

Step 5, satellite carry out time slot and initialize distribution: when user accesses satellite first and sends reserve requests, satellite is allocated according to the terminal queuing packet count of each user.Particularly as follows: satellite calculates after pressing reservation table distribution, whether remaining time slots is more than 0, if it is, according to priority distribute time slot for reservation table, for free allocation table poll distribution remaining time slots after distributing；Otherwise, according to weight assignment formula be reservation table distribution time slot.

Weighted formula is:

${\mathrm{allocate}}_i=\frac{data\_num}{\underset{i=1}{\overset{m}{\&Sigma;}}req\_{\mathrm{num}}_i}req\_{\mathrm{num}}_i,$

Wherein, allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresenting terminal i reservation data timeslot number, m represents the total number of terminals currently having time slot request.

The terminal timeslot number of prediction next frame while step 6, initialization distribution, when next frame arrives satellite, satellite distributes according to the adjustment time slot that predicts the outcome of previous frame, and repeats in each frame later, and Adjusted Option is as follows:

Initialize the distribution time slot terminal more than or equal to prediction time slot for all, carry out configuration as follows:

$\left\{\begin{array}{c}{\mathrm{allocate}}_i={\mathrm{pro}}_i\\ free=\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{allocate}}_i-{\mathrm{pro}}_i\end{array}\right.$

Then, total free timeslot number free being re-assigned to again and initializes the distribution time slot terminal less than prediction time slot, the quantity of concrete distribution is:

${\mathrm{allocate}}_i={\mathrm{allocate}}_i+free\&times;\frac{{\mathrm{pro}}_i-{\mathrm{allocate}}_i}{\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{pro}}_i-{\mathrm{allocate}}_i}$

Wherein allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresent terminal i reservation data timeslot number, pro_iRepresenting the timeslot number of prediction terminal i, free represents total free timeslot number.

Claims (3)

1. the satellite network cut-in method based on prediction Dynamic Bandwidth Allocation, it is characterised in that comprise the following steps:

Step 1: satellite broadcasts oneself identification number ID in space, terminal by receive satellite identification number id information and ON status information be sent to respective satellite, ON status information represents that present terminal has packet to produce；

Step 2: after satellite receives the information that terminal is sent, detects identification number id information therein, if identification number ID mates with the identification number ID of oneself, then satellite distribution request time slot is to terminal, and request gap information is issued terminal；

Step 3: after terminal is successfully accessed, sends request to satellite, comprises Subscriber Queue length, i.e. terminal queuing packet count in solicited message；

Step 4: satellite arranges priority number from big to small according to terminal queuing packet count, and inserts reservation table and free allocation table；

Step 5: satellite carries out time slot and initializes distribution, i.e. after terminal accesses satellite first and sends reserve requests, satellite is allocated according to each terminal queuing packet count；

Step 6: predict the terminal timeslot number of next frame while initializing distribution, when next frame arrives satellite, the satellite adjustment time slot distribution that predicts the outcome according to previous frame, and each frame later repeats.

2. the satellite network cut-in method based on prediction Dynamic Bandwidth Allocation as claimed in claim 1, it is characterised in that initialize distribution particularly as follows:

Satellite calculates remaining time slots after pressing reservation table distribution and, whether more than 0, if it is, according to priority distribute time slot for reservation table, distributes remaining time slots for free allocation table poll after distributing；Otherwise, being reservation table distribution time slot according to weight assignment formula, weighted formula is:

${\mathrm{allocate}}_i=\frac{data\_num}{\underset{i=1}{\overset{m}{\&Sigma;}}req\_{\mathrm{num}}_i}req\_{\mathrm{num}}_i,$

Wherein, allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresenting terminal i reservation data timeslot number, m represents the total number of terminals currently having time slot request.

3. the satellite network cut-in method based on prediction Dynamic Bandwidth Allocation as claimed in claim 1, it is characterised in that time slot distribution Adjusted Option in described step 6 particularly as follows:

Initialize the distribution time slot terminal more than or equal to prediction time slot for all, carry out configuration as follows:

$\left\{\begin{array}{c}{\mathrm{allocate}}_i={\mathrm{pro}}_i\\ free=\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{allocate}}_i-{\mathrm{pro}}_i\end{array}\right.,$

Then, total free timeslot number free being re-assigned to terminal again that initialize distribution time slot less than prediction time slot, the quantity of concrete distribution is:

${\mathrm{allocate}}_i={\mathrm{allocate}}_i+free\&times;\frac{{\mathrm{pro}}_i-{\mathrm{allocate}}_i}{\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{pro}}_i-{\mathrm{allocate}}_i},$

Wherein, allocate_iRepresenting the timeslot number of terminal i distribution, data_num represents total data slot number, req_num_iRepresent terminal i reservation data timeslot number, pro_iRepresenting the timeslot number of prediction terminal i next frame, free represents total free timeslot number, and m represents the total number of terminals currently having time slot request.