CN107277816B - A kind of high altitude platform networking frequency spectrum distributing method - Google Patents

Abstract

This application discloses a high-altitude platform networking spectrum allocation method, including the following steps: calculating the spectrum revenue of each high-altitude platform, the factors affecting the spectrum revenue include its own spectrum resource revenue, borrowed spectrum resource revenue, borrowed spectrum resource overhead, Renting spectrum resource income; changing the amount of spectrum resources borrowed between high-altitude platforms, including each of the high-altitude platforms borrowing spectrum resources from at least one other high-altitude platform, or each of the high-altitude platforms lending spectrum resources to at least one other high-altitude platform , change the spectrum revenue of each high-altitude platform; calculate the network revenue, which is the sum of the spectrum revenues of the multiple high-altitude platforms; repeat the above steps to obtain the amount of spectrum resources borrowed between the high-altitude platforms when the network revenue is maximum. The preferred embodiment uses the Elman neural network method to change the amount of spectrum resources borrowed between high-altitude platforms. The method of the present application improves the frequency spectrum utilization efficiency of the high-altitude platform networking and improves the overall benefit.

Description

A method for allocating frequency spectrum in high-altitude platform networking

technical field

The present application relates to the communication field, and in particular to a frequency spectrum allocation method for high-altitude platform networking.

Background technique

In the space-ground integrated networking scheme, the high-altitude platform (NSP, Near Space Platform) has attracted extensive attention from the industry due to its large capacity, strong flexibility, and wide coverage, and has gradually become the center of the space-ground integrated network. Indispensable communication node. Considering that the high-altitude platform communication scene is open and the signal transmission loss attenuation is small, the resource division scheme of frequency reuse is adopted to avoid mutual interference between cells. However, with the continuous development of existing communication technologies, the proportion of broadband services such as multimedia is gradually increasing. At the same time, the location of hotspot areas and the number of users are constantly changing. It is difficult to meet the needs of broadband communication services if the traditional fixed spectrum allocation scheme is adopted.

Contents of the invention

In view of this, the present invention proposes a spectrum allocation method for high-altitude platform networking to solve the problem of low efficiency of spectrum allocation.

Embodiments of the present invention propose a high-altitude platform network spectrum allocation method, including the following steps:

Calculate the spectrum revenue of each high-altitude platform, and the influencing factors of the spectrum revenue include its own spectrum resource revenue, borrowed spectrum resource revenue, borrowed spectrum resource overhead, and leased spectrum resource revenue;

Changing the amount of spectrum resources borrowed between high-altitude platforms includes borrowing spectrum resources from each of the high-altitude platforms to at least one other high-altitude platform, or lending spectrum resources to at least one other high-altitude platform from each of the high-altitude platforms, changing the amount of spectrum resources for each high-altitude platform Spectrum benefits of the platform;

Calculate the network revenue, which is the sum of the spectrum revenue of the multiple high-altitude platforms;

Repeat the above steps to obtain the amount of spectrum resources borrowed between high-altitude platforms when the network revenue is maximum.

In one embodiment of the present invention, the spectral benefit of the kth high-altitude platform is calculated in the following manner, as

in,

Indicates the self-spectrum resource income of the kth high-altitude platform;

Indicates that the kth high-altitude platform borrows spectrum resource income;

Indicates that the kth high-altitude platform borrows spectrum resource overhead;

Indicates the income of the kth high-altitude platform from renting spectrum resources;

Among them, Q _k (i) represents the profit margin of the k-th high-altitude platform providing the i-type service, P _k (i) represents the price of the k-th high-altitude platform providing the i-type service, Indicates the transmission rate of the i-type service on k high-altitude platforms;

represents the borrowing price, Indicates the lending price, b _k (j) and b _j (k) represent the spectrum borrowed by the k-th high-altitude platform from the j-th high-altitude platform, and the spectrum leased by the k-th high-altitude platform to the j-th high-altitude platform, respectively.

In the further optimized embodiment of the high-altitude platform networking spectrum allocation method of the present invention, the own spectrum resources of each high-altitude platform firstly meet its own spectrum needs, then: when the own spectrum resources are left, the borrowed spectrum resources The revenue and the borrowed spectrum resource overhead take a value of 0; when the own spectrum resource is insufficient, the leased spectrum resource revenue takes a value of 0.

In the further optimized embodiment of the high-altitude platform networking spectrum allocation method of the present invention, the Elman neural network method is used to change the amount of spectrum resources borrowed between high-altitude platforms.

Specifically, the further optimized embodiment of the high-altitude platform networking resource allocation method of the present invention includes the following steps:

Predict the speed of each service of the kth high-altitude platform at time t as

The available spectrum and the borrowed spectrum at time t are expressed as B _k and ζ _k,t , the initial value of B _k is B, and the initial value of ζ _k,t is 0;

The free spectrum resource of high-altitude platform k is

η _k (j) represents the utilization rate of the frequency spectrum of platform j by platform k, that is to say, rate/spectrum utilization rate=required frequency spectrum;

judge whether

If no, take Make

Therefore there are

definition If σ _k,t >0, take

ρ _k,t = max(σ _k,t ,0);

When changing the amount of spectrum resources borrowed between high-altitude platforms, calculate the priority of k platform borrowing spectrum from j platform

and sort the priority, Represents the spectrum exceeded by the high-altitude platform k;

if then take

if then take

Recalculate the priority according to the updated value until all spectrums are allocated, and finally update _Bk according to the amount of borrowed spectrum resources between high-altitude platforms;

The leased spectrum is The borrowed spectrum is

The above-mentioned at least one technical solution adopted in the embodiment of the present application can achieve the following beneficial effects: the method of the present application improves the spectrum utilization efficiency of the high-altitude platform networking and improves the overall benefit; by flexibly controlling the bandwidth borrowed and lent by each high-altitude node, It can dynamically adapt to the situation that the proportion of broadband services such as multimedia is gradually increasing, and the location of hotspot areas and the number of users are constantly changing.

Description of drawings

The drawings described here are used to provide a further understanding of the application and constitute a part of the application. The schematic embodiments and descriptions of the application are used to explain the application and do not constitute an improper limitation to the application. In the attached picture:

Figure 1 is a schematic diagram of a high-altitude platform networking communication scenario;

Fig. 2 is the flow chart of the embodiment of spectrum allocation method of the present application;

Fig. 3 is a schematic diagram of the Elman network structure;

Fig. 4 is a schematic diagram of an embodiment of spectrum allocation using the Elman neural network method.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in conjunction with specific embodiments of the present application and corresponding drawings. Apparently, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

This patent proposes a spectrum allocation method based on the Elman neural network prediction model in the high-altitude platform networking scenario. This method firstly divides the spectrum income in the cell into four categories - "own spectrum resource income, leased spectrum resource income, leased spectrum resource overhead, leased spectrum resource income", and then by adding a central control unit (CCU, Center Control Unit), all the frequency bands are allocated in the frequency band pool, and the optimized gradient descent algorithm is used, which can not only improve the training rate of the network, but also effectively prevent the network from falling into local minimum points. The purpose of this algorithm self-learning is to use the difference between the predicted rate (actual output value) of each service in the network and the actual rate (output sample value) of each service to modify the weight and threshold, so that the square error of the network output layer and the minimum, to achieve the highest overall network revenue.

The technical solutions provided by various embodiments of the present application will be described in detail below in conjunction with the accompanying drawings.

Figure 1 is a schematic diagram of a high-altitude platform networking communication scenario. Suppose there are 7 high-altitude platforms (NSP) to form a system together, the height of each high-altitude platform is 30km, the coverage radius is 175km, and the frequency band used by each high-altitude platform is different. In this solution, by adding a central control unit (CCU), all the frequency bands are allocated in the frequency band pool. Each high-altitude platform will be allocated the bandwidth of B initially, and the total width of the frequency band pool is 7B.

The spectrum revenue of the high-altitude platform is the revenue of its own spectrum resources + the revenue of borrowed spectrum resources - the cost of borrowed spectrum resources + the revenue of leased spectrum resources.

In the embodiment of the high-altitude platform networking spectrum allocation method of the present invention, the own spectrum resources of each high-altitude platform firstly meet its own spectrum needs, then: when the own spectrum resources are left, the borrowed spectrum resource income, The value of the borrowed spectrum resource overhead is 0; when the own spectrum resource is insufficient, the leased spectrum resource revenue is 0.

Fig. 2 is a flow chart of an embodiment of the spectrum allocation method of the present application.

Step 1. Calculate the spectrum revenue of each high-altitude platform. The influencing factors of the spectrum revenue include the revenue of its own spectrum resources, the revenue of borrowed spectrum resources, the cost of borrowed spectrum resources, and the revenue of leased spectrum resources;

Step 2. Changing the amount of spectrum resources borrowed between high-altitude platforms, including borrowing spectrum resources from each of the high-altitude platforms to at least one other high-altitude platform, or lending spectrum resources to at least one other high-altitude platform by each of the high-altitude platforms, changing Spectrum yield for each high-altitude platform;

Step 3, calculating the network revenue, which is the sum of the spectrum revenue of the multiple high-altitude platforms;

Step 4. Repeat the above steps to obtain the amount of spectrum resources borrowed between high-altitude platforms when the network revenue is maximum.

In one embodiment of the present invention, the spectral benefit of the kth high-altitude platform is calculated in the following manner, as

in,

Indicates the self-spectrum resource income of the kth high-altitude platform;

Indicates that the kth high-altitude platform borrows spectrum resource income;

Indicates that the kth high-altitude platform borrows spectrum resource overhead;

Indicates the income of the kth high-altitude platform from renting spectrum resources;

Among them, Q _k (i) represents the profit margin of the k-th high-altitude platform providing the i-type service, P _k (i) represents the price of the k-th high-altitude platform providing the i-type service, Indicates the transmission rate of the i-type service on k high-altitude platforms;

represents the borrowing price, Indicates the lending price, b _k (j) and b _j (k) represent the spectrum borrowed by the k-th high-altitude platform from the j-th high-altitude platform, and the spectrum leased by the k-th high-altitude platform to the j-th high-altitude platform, respectively.

Under the premise that each platform meets its own spectrum needs as the first priority, it is impossible to lease and borrow spectrum at the same time, so the spectrum revenue in (1) can be divided into two cases

There is a surplus of its own spectrum, which can be leased out

There is no surplus of its own spectrum and must be borrowed

Assuming that B is the average spectrum allocated to each high-altitude platform at the beginning, and B _k is the available spectrum after k platform is actually borrowed and leased, then it can also be divided into the following two situations:

There is a surplus of its own spectrum, and when the spectrum is leased out, the available spectrum is:

There is no remaining spectrum of its own. When borrowing spectrum, the available spectrum is:

Express the utilization ratio of platform k utilizing the frequency spectrum of platform j with η _k (j), for example expressed as:

where γ _k (j) represents the matched receiver signal-to-noise ratio, Indicates the target bit error rate of the k platform.

In order to maximize the network revenue, the optimization objective is

It is necessary to determine whether each platform is borrowing spectrum or leasing spectrum, so as to calculate with the corresponding formula.

It should be noted that since the ultimate goal of each high-altitude platform is to meet the transmission rate requirements of users within the coverage of the platform, the spectrum requirements of each high-altitude platform can be equivalent to the transmission rate requirements of users within the coverage of the platform.

Figure 3 is a schematic diagram of the Elman network structure. The following proposes a method of using the Elman neural network to predict the transmission rate requirements of users within the coverage of each high-altitude platform at time t through historical information, so as to obtain the spectrum demand of each high-altitude platform at time t.

First, the transmission rate prediction needs to be performed. Provide M _k types of services on the kth high-altitude platform, assuming that the transmission rate of each type at time t-1 is If the statistics are counted from the initial moment to the time t-1, the matrix of the rate is

In this scheme, Elman neural network is used to predict the transmission rate at time t. Elman regression neuron network is generally divided into four layers: input layer, middle layer (hidden layer), receiving layer and output layer. The connection of its input layer, hidden layer and output layer is similar to the feedforward network, the unit of the input layer only plays the role of signal transmission, and the unit of the output layer plays the role of linear weighting. The transfer function of the hidden layer unit can be a linear or nonlinear function. The successor layer is also called the context layer or the state layer. It is used to remember the output value of the hidden layer unit at the previous moment, which can be considered as a one-step delay operator. .

The characteristic of the Elman regression neuron network is that the output of the hidden layer is self-connected to the input of the hidden layer through the delay and storage of the undertaking layer. This self-connection method makes it sensitive to the data of the historical state, and the internal feedback network Joining increases the ability of the network itself to process dynamic information, thus achieving the purpose of dynamic modeling. In addition, the Elman regression neural network can approach any nonlinear mapping with arbitrary precision, without considering the specific form of the impact of external noise on the system. If the input and output data pairs of the system are given, the system can be modeled.

As shown in Figure 3, the nonlinear state space expression of the Elman neural network is:

y(k)=g(w ³ x(k)+b ₂ ) (9)

x(k)=f(w ¹ x _c (k)+w ² (u(k-1))+b ₁ ) (10)

_xc (k)=x(k-1) (11)

Among them, k represents the moment, y, x, u, and x _c represent the 1-dimensional output node vector, the m-dimensional hidden layer node unit vector, the n-dimensional input vector and the m-dimensional feedback state vector. w ³ , w ² , and w ¹ respectively represent the connection weight matrix from the hidden layer to the input layer, from the input layer to the hidden layer, and from the receiving layer to the hidden layer. f(.) is the transfer function of hidden layer neurons, and g(.) is the transfer function of output layer. b ₁ , b ₂ are the thresholds of the input layer and the hidden layer respectively. The Elman neural network learning algorithm uses an optimized gradient descent algorithm, that is, the adaptive learning rate momentum gradient descent backpropagation algorithm, which can not only increase the training rate of the network, but also effectively inhibit the network from falling into local minimum points. The purpose of learning is to use the difference between the actual output value of the network and the output sample value to modify the weight and threshold, so that the sum of squared errors of the network output layer is minimized. Let the actual output vector of the k-th step system be y _d (k), and in the time period (0, T), define the error function as:

Taking w ³ and w ² as an example, calculate the partial derivatives of E with respect to w ³ and w ² respectively, and the weight correction formula can be obtained as follows:

Among them, φ is the learning rate, mc is the momentum factor, and the default value is 0.9. In this way, not only the current gradient direction, but also the gradient direction at the previous moment are considered when updating, which reduces the sensitivity of network performance to parameter adjustment. The local minima are effectively suppressed.

The transmission rate prediction process is

Step A, train the network, assuming that the length of the training sequence is α, this length indicates how many lines in (8) are used for training, the length of α will determine the number of training times and training accuracy of the network, and also determine how much historical information will be used to make predictions. For example, if α is 3 and t is 10, then it will be trained 6 times with {1,2,3}, {2,3,4}, {3,4,5}..., {6,7,8}, and then Estimate the rate of 10 as {7,8,9}.

Step B, the input vector will be 3×M _k dimensional, then according to experience, the number of hidden layer units is set to 2M _k -1, a relatively good prediction result will be obtained, and the output vector will be 1×M _k dimensional of.

Step C. Make predictions

Through the above method, the speed of each service of high-altitude platform k is predicted at time t as

From this, the predicted transmission rate of users within the coverage of each high-altitude platform at time t can be obtained, which is used to calculate the spectrum demand of each high-altitude platform at time t, making it possible to rent or borrow spectrum between platforms, and maximize Network earnings. A method of allocating spectrum according to the transmission rate will be described below.

Fig. 4 is a schematic diagram of an embodiment of spectrum allocation using the Elman neural network method.

For high-altitude platform k, the available spectrum and borrowed spectrum at time t are denoted as B _k and ζ _k,t , the initial value of B _k is B, and the initial value of ζ _k,t is 0. Define the free spectrum resource of high-altitude platform k as ρ _k,t , then it can be defined as

Step 10, the prediction stage, calculating the rate of each service according to the available frequency spectrum;

If the following formula is satisfied, then the spectrum of the high-altitude platform k itself meets the demand, and there may be spectrum for sale

If it cannot be satisfied, then set the rate that can be satisfied as but

Then the excess is defined as

Step 20, updating and clearing phase, calculate the borrowed spectrum and idle spectrum resources of each high-altitude platform;

For high-altitude platform k, define σ _k,t

If σ _k,t >0, it means that the spectrum resources are sufficient, and update ζ _k,t as follows

ρ _k,t = max(σ _k,t ,0) (21)

Step 30, in the dispatching stage, calculate the borrowed spectrum and leased spectrum of the high-altitude platform, and update the available spectrum value.

Step 301. When the scheduler receives a spectrum borrowing request from a high-altitude platform, he will check whether other platforms have available spectrum and calculate the priority

(Proceeds from borrowing less rent)

Step 302, sorting the priorities, Represents the spectrum exceeded by high-altitude platform k, λ _k (i,j) represents the priority of platform k to borrow spectrum from platform j

There are two cases:

(1) The borrowable spectrum resource of platform j is greater than the demand of platform k, that is to say Then the k platform will not borrow from other platforms, and the update is as follows

(2) is larger than ρ _j,t (the borrowable spectrum resource of platform j is less than the demand of platform k), that is to say Then update as follows,

Step 303, recalculate the priority according to the updated value until all spectrums have been allocated

Finally, according to the rental and borrowing conditions of the high-altitude platform, B _k is updated by the following formula

Leased Spectrum:

Borrow spectrum:

Step 304, perform an allocation operation.

It should also be noted that the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes Other elements not expressly listed, or elements inherent in the process, method, commodity, or apparatus are also included. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

The above descriptions are only examples of the present application, and are not intended to limit the present application. For those skilled in the art, various modifications and changes may occur in this application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims (4)

1. A high-altitude platform network spectrum allocation method is characterized in that, comprising the following steps: Calculate the spectrum revenue of each high-altitude platform, and the influencing factors of the spectrum revenue include its own spectrum resource revenue, borrowed spectrum resource revenue, borrowed spectrum resource overhead, and leased spectrum resource revenue; Changing the amount of spectrum resources borrowed between high-altitude platforms includes borrowing spectrum resources from each of the high-altitude platforms to at least one other high-altitude platform, or lending spectrum resources to at least one other high-altitude platform from each of the high-altitude platforms, changing the amount of spectrum resources for each high-altitude platform Spectrum benefits of the platform; Calculate the network revenue, which is the sum of the spectrum revenue of the multiple high-altitude platforms; Repeat the above steps to obtain the amount of spectrum resources borrowed between high-altitude platforms when the network revenue is maximum. 2. the high-altitude platform network spectrum allocation method as claimed in claim 1, is characterized in that The spectrum gain of the kth high-altitude platform is in, Indicates the self-spectrum resource income of the kth high-altitude platform; Indicates that the kth high-altitude platform borrows spectrum resource income; Indicates that the kth high-altitude platform borrows spectrum resource overhead; Indicates the income of the kth high-altitude platform from renting spectrum resources; Among them, Q _k (i) represents the profit margin of the k-th high-altitude platform providing the i-type service, P _k (i) represents the price of the k-th high-altitude platform providing the i-type service, Indicates the transmission rate of the i-type service on k high-altitude platforms; represents the borrowing price, Indicates the lending price, b _k (j) and b _j (k) represent the spectrum borrowed by the k-th high-altitude platform from the j-th high-altitude platform, and the spectrum leased by the k-th high-altitude platform to the j-th high-altitude platform, respectively. 3. high-altitude platform network spectrum allocation method as claimed in claim 2, is characterized in that, Each high-altitude platform's own spectrum resources give priority to meeting its own spectrum needs; When the self-spectrum resource has surplus, the value of the borrowed spectrum resource income and the borrowed spectrum resource overhead is 0; When the self-spectrum resource is insufficient, the value of the leased spectrum resource income is 0. 4. The high-altitude platform network spectrum allocation method according to any one of claims 1 to 3, characterized in that, Using the Elman neural network method to change the amount of spectrum resources borrowed between high-altitude platforms.