JP7149583B2 - Communication system, access point, communication method, and program - Google Patents

Description

The present invention relates to technology for improving communication conditions in a communication system.

In recent years, with the spread of highly functional mobile terminals such as smart phones, the demand for mobile communication traffic is rapidly increasing. As a result, in addition to the conventional expansion of the use of wireless LAN (Local Area Network), offloading to wireless LAN has progressed due to the increase in mobile data traffic due to the spread of smartphones, and unlicensed bands (2.4 GHz band, 5 GHz band) ) is seeing a surge in traffic.

In addition, technologies related to IoT (Internet Of Things) and M2M (Machine to Machine) tend to be applied to many systems. Improving efficiency is an urgent issue.

Here, since the usage of radio resources varies depending on time, place, frequency band, radio channel, etc., a situation may occur in which only some frequency bands (or radio channels) are congested.

In other words, even if there are free radio resources in the radio system as a whole, congestion may occur. An existing private wireless system (for example, IEEE802.11 wireless LAN) uses a single frequency band, or determines one band to be used in advance before performing communication. For example, IEEE802.11n is used after setting whether to use the 2.4 GHz band or the 5 GHz band. For this reason, even if there are free radio resources in the existing private radio system as a whole, congestion may occur.

Various techniques have been developed to deal with such situations. For example, in the technology disclosed in Patent Literature 1, an access point that wirelessly communicates with a plurality of stations collects information about channel usage status from each station, and selects a channel to use based on the collected information. Execute the process to change. In other words, in the technology of Patent Document 1, the access point monitors the channel usage status of each station, thereby executing optimum channel selection processing. In other words, the technique of Patent Document 1 collects information on the channel usage status of each station at the center (device (access point) that manages the whole), and the device (access point) that manages the whole centralizes the information. It optimizes the channel selection process.

U.S. Patent Application Publication No. 2002/0188723

However, in the conventional technology as described above, it is necessary to collect data on the communication status from each station in order to grasp the overall communication status of the communication system. In other words, in the conventional technology as described above, it is necessary to collect data about the communication status from each station, provide a device (central management device) for monitoring the data, and construct a centralized communication system. For this reason, it is difficult to optimize the performance of a communication system in a distributed manner with the prior art.

Therefore, in view of the above problems, it is an object of the present invention to realize a communication system capable of distributively optimizing the performance of the communication system, an access point used in the communication system, a communication method, and a program. .

In order to solve the above problems, a first invention comprises a plurality of terminals and a plurality of access points each capable of communicating with some or all of the plurality of terminals, and a plurality of frequency bands. It is a communication system that performs data communication using a plurality of channels for each.

Each access point performs a channel usage data obtaining step, an updating step, a receiving step, a fairness evaluation value obtaining step, a channel selection period adjusting step, and a channel selecting step.

The channel usage data acquiring step acquires channel usage data, which is data indicating the usage of each channel of each frequency band of the own access point, in the own access point, which is the own device.

In the updating step, based on the channel usage data, the channel probability of each frequency band of the own access point is changed so that the greater the available resources of the channel, the higher the channel probability, which is the probability that the channel will be selected. Update probabilities.

The receiving step receives data indicating usage status of each channel of each frequency band from another access point.

The fairness evaluation value obtaining step obtains a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point.

In the channel selection period adjustment step, based on the fairness evaluation value, the channel selection period, which is the period of the channel selection process that is the process of selecting a channel, and the usage status of each channel of each frequency band of the own access point are adjusted. At least one of the channel usage data acquisition cycle, which is the cycle for acquiring the channel usage data, which is the indicated data, is adjusted.

The channel selection step executes channel selection processing for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection period.

In this communication system, the concept of fairness in the use of channels for each access point is introduced, and the channel probability of each channel is set so that the throughput of each access point is increased while not impairing the fairness in the use of the channel. is adjusted (updated). As a result, it is possible to improve the throughput of each access point by adjusting the channel utilization rate of each access point in a distributed manner. , fairness in the use of each channel can also be ensured.

In other words, in this communication system, by distributing the above processing at each access point, it is possible to increase the throughput of the entire communication system while ensuring the fairness of the channel usage of each access point. can.

Therefore, in this communication system 1000, the performance of the communication system can be optimized in a distributed manner.

A second invention is the first invention, wherein the channel probability is calculated by a probability distribution according to a Gibbs distribution to which Gibbs sampling can be applied.

Accordingly, in this communication system, the channel probability can be calculated by a probability distribution following the Gibbs distribution to which Gibbs sampling can be applied.

A third invention is the first or second invention, wherein the channel utilization data is data derived from a CUR value that indicates the channel occupancy rate in a predetermined period.

As a result, in this communication system, the CUR value can be used to update the channel probability so as to improve the throughput of each access point, and the fairness evaluation value can be obtained based on the CUR value. .

により算出される。 A fourth invention is the third invention, wherein AP _b is the access point, i is the frequency band, ci is the channel included in the frequency band _{i, ζc i is} the share rate of the channel _ci , If the CUR value of channel c _i of access point AP _b is ρ(AP _b , c _i ), the fairness evaluation value is

Calculated by

Thereby, in this communication system, the fairness of the CUR value can be evaluated by the value calculated based on the Raj Jain's formula.

A fifth invention is any one of the first to fourth inventions, wherein the access point externally transmits data indicating the usage status of each channel of each frequency band in the own access point by broadcasting.

As a result, in this communication system, each access point can acquire data indicating the usage status of each channel of each frequency band in other access points through broadcast reception.

A sixth invention comprises a plurality of terminals and a plurality of access points each capable of communicating with a part or all of the plurality of terminals, and uses a plurality of channels for each of a plurality of frequency bands. An access point used in a communication system that performs data communication using a control unit, and a memory that holds at least one of commands and data in order to execute predetermined processing.

The access point performs a channel usage data acquisition step, an update step, a reception step, a fairness evaluation value acquisition step, a channel selection period adjustment step, and a channel selection step.

The channel usage data acquiring step acquires channel usage data, which is data indicating the usage of each channel of each frequency band of the own access point, in the own access point, which is the own device.

In the updating step, based on the channel usage data, the channel probability of each frequency band of the own access point is changed so that the greater the available resources of the channel, the higher the channel probability, which is the probability that the channel will be selected. Update probabilities.

The receiving step receives data indicating usage status of each channel of each frequency band from another access point.

The fairness evaluation value obtaining step obtains a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point.

In the channel selection period adjustment step, based on the fairness evaluation value, the channel selection period, which is the period of the channel selection process that is the process of selecting a channel, and the usage status of each channel of each frequency band of the own access point are adjusted. At least one of the channel usage data acquisition cycle, which is the cycle for acquiring the channel usage data, which is the indicated data, is adjusted.

The channel selection step executes channel selection processing for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection cycle.

As a result, an access point having the same effect as the first invention can be realized.

A seventh invention comprises a plurality of terminals and a plurality of access points each capable of communicating with some or all of the plurality of terminals, and uses a plurality of channels for each of a plurality of frequency bands. This is a communication method used in a communication system that performs data communication using The communication method includes a channel usage data acquisition step, an update step, a reception step, a fairness evaluation value acquisition step, a channel selection period adjustment step, and a channel selection step.

In the channel usage data acquisition step, each access point acquires channel usage data, which is data indicating the usage of each channel of each frequency band of the own access point, at the access point itself.

In the update step, each access point adjusts each frequency band of its own access point based on the channel usage data so that the channel probability, which is the probability that the channel will be selected, increases as the available resource of the channel increases. Update the channel probability for each channel in .

The receiving step receives data indicating usage status of each channel of each frequency band from another access point.

The fairness evaluation value obtaining step obtains, by each access point, a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point.

The channel selection period adjustment step is performed by each access point based on the fairness evaluation value, and the channel selection period, which is the period of the channel selection process that is the process of selecting the channel, and the frequency band of the own access point. At least one of the channel usage data acquisition cycle, which is the cycle for acquiring channel usage data, which is data indicating channel usage, is adjusted.

In the channel selection step, each access point executes channel selection processing for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection period.

Accordingly, it is possible to realize a communication method having the same effect as the first invention.

An eighth invention is a program for causing a computer to execute the communication method of the seventh invention.

Accordingly, it is possible to realize a program for causing a computer to execute a communication method having the same effect as the first invention.

ADVANTAGE OF THE INVENTION According to this invention, the communication system which can optimize the performance of a communication system in distribution, the access point used for the said communication system, the communication method, and a program are realizable.

FIG. 2 is a diagram schematically showing a space SP1 in which access points of the communication system 1000 according to the first embodiment are arranged; 1 is a schematic configuration diagram of a communication system 1000 according to a first embodiment; FIG. 1 is a schematic configuration diagram of an access point AP1 in the first embodiment; FIG. 4 is a flowchart for explaining the operation in the communication system 1000; 4 is a flowchart for explaining the operation in the communication system 1000; Figures showing channel utilization for each access point and the overall communication system (unbalanced and balanced); FIG. 4 is a diagram showing a CPU bus configuration;

[First embodiment]
A first embodiment will be described below with reference to the drawings.

<1.1: Configuration of communication system>
FIG. 1 is a diagram schematically showing a space SP1 in which access points of a communication system 1000 according to the first embodiment are arranged.

FIG. 2 is a schematic configuration diagram (one example) of the communication system 1000 according to the first embodiment.

FIG. 3 is a schematic configuration diagram of the access point AP1 according to the first embodiment.

For example, as shown in FIG. 1, the communication system 1000 includes a plurality of access points (access points AP1 to AP4 in FIG. 1) and a plurality of wireless terminals (wireless terminals ST11 to ST11 in FIG. 1) communicating with each access point. ST44). A plurality of access points (access points AP1 to AP4 in FIG. 1) and a plurality of wireless terminals (wireless terminals ST11 to ST44 in FIG. 1) communicating with each access point are arranged in the space SP1. there is

For convenience of explanation, as an example, the communication system 1000 includes four access points AP1 to AP4 and (1) four wireless terminals ST11 to ST14 communicating with the access point AP1, as shown in FIG. , (2) four wireless terminals ST21 to ST24 communicating with the access point AP2, (3) four wireless terminals ST31 to ST34 communicating with the access point AP3, and (4) communicating with the access point AP4. 4 wireless terminals ST41 to ST44 are included in the following description.

As shown in FIG. 2, the communication system 1000 includes four access points AP1 to AP4, wireless terminals ST11 to ST44 communicating with each access point, and a plurality of switch devices (switch devices SW1 to SW3 in FIG. 1). Prepare. Further, as shown in FIG. 2, the communication system 1000 includes a gateway GW1, a server Svr1 connected to the gateway GW1, and a server Svr2 connected to the gateway GW1 via the network NW1.

Note that the configuration of FIG. 2 is an example of the configuration of the communication system 1000, and the configuration of the communication system 1000 is not limited to this configuration, and may be another configuration.

The switch devices SW1 to SW3 are devices implemented by, for example, L2 switches (switching hubs) and L3 switches. Each wireless terminal connects to an external server (for example, servers Svr1 and Svr2) via each access point (access point in charge), switch devices SW1 to SW3 (switch devices on the communication path), gateway GW1, network NW1, and the like. and data communication.

The access point AP1 includes an antenna Ant1, a communication processing unit 1, a communication resource adjustment unit 2, and a data processing unit 3, as shown in FIG.

The antenna Ant1 is a wireless communication antenna for transmitting and receiving a wireless communication signal (RF signal).

The communication processing unit 1 includes an RF processing unit 11 and a baseband processing unit 12, as shown in FIG.

The RF processing unit 11 performs RF demodulation processing on the RF signal received by the antenna Ant1, acquires an RF demodulated signal, and outputs the acquired RF demodulated signal to the baseband processing unit 12 . Further, the RF processing unit 11 performs RF modulation processing on the modulated baseband signal input from the baseband processing unit 12 to obtain a wireless communication signal (RF signal). Then, the acquired RF signal is transmitted to the outside via the antenna Ant1. Also, the RF processing unit 11 acquires information on channels that can be used by the access point AP1.

The baseband processing unit 12 performs baseband demodulation processing on the RF demodulated signal output from the RF processing unit 11 (for example, when the modulation method is OFDM, guard interval (GI) removal processing, FFT conversion, demapping processing, parallel/serial conversion, etc.) to obtain a baseband demodulated signal. Further, the baseband processing unit 12 performs baseband modulation processing (for example, when the modulation method is OFDM, serial/parallel conversion, mapping processing, inverse FFT conversion, guard processing such as interval (GI) addition processing and D/A conversion), acquires a modulated baseband signal, and outputs the acquired modulated baseband signal to the RF processing unit 11 .

The communication processing unit 1 outputs data including data acquired by the RF processing unit 11 and the baseband processing unit 12 to the communication resource adjustment unit 2 as data D1. The data D1 includes (1) information such as reception sensitivity and reception power at frequencies corresponding to each channel, and (2) data received from other access points, which is the CUR value (CUR: Channel Utilization Ratio: ratio of the time the channel is used), data including information on the used channel, and the like.

The communication processing unit 1 also receives data D2 output from the communication resource adjusting unit 2, and performs baseband processing and RF processing so that a channel (communication channel) based on the data D2 is selected.

The communication resource adjustment unit 2 includes a channel probability processing unit 21, a fairness evaluation processing unit 22, and a channel selection processing unit 23, as shown in FIG.

The channel probability processing unit 21 includes an available channel detection unit 211, an average CUR value acquisition unit 212, a Gibbs distribution calculation processing unit 213, and a channel probability update processing unit 214, as shown in FIG.

Available channel detection section 211 receives data D1 output from communication processing section 1, and based on the data D1, detects channels available for access point AP1 for each frequency band. For example, the available channel detection unit 211 detects, for each frequency band, channels available for the access point AP1 from information about available channels for each frequency band detected by the RF processing unit 11 of the communication processing unit 1. To detect. Then, available channel detection section 211 outputs data including the detection result to average CUR value acquisition section 212 as data D11.

The average CUR value acquisition unit 212 receives the data D11 output from the average CUR value acquisition unit 212 . The average CUR value acquiring unit 212 acquires the CUR value for each used channel in the used frequency band (frequency band) based on the data D11. Then, the average CUR value acquisition unit 212 outputs the acquired data to the Gibbs distribution calculation processing unit 213 as data D12.

The Gibbs distribution calculation processing unit 213 receives the data D12 output from the average CUR value acquisition unit 212 . The Gibbs distribution calculation processing unit 213 calculates the Gibbs distribution based on the data D12. Then, data including the calculated Gibbs distribution data is output to channel probability update processing section 214 as data D13.

Channel probability update processing section 214 receives data D13 output from Gibbs distribution calculation processing section 213 . The channel probability update processing unit 214 executes processing for updating the channel probability of each channel based on the data D13. Then, channel probability update processing section 214 outputs data including updated channel probability Pc to channel selection processing section 23 as data D14.

The fairness evaluation processing unit 22 includes, as shown in FIG.

The other AP information acquisition unit 221 receives data D1 output from the communication processing unit 1, and obtains (1) CUR values of other access points (access points other than the access point AP1) and (2 ) to obtain information about the channel used. The other AP information acquisition unit 221 outputs data including the acquired data to the fairness evaluation value calculation unit 222 as data D15.

The fairness evaluation value calculation unit 222 receives the data D15 output from the other AP information acquisition unit 221, and acquires the fairness evaluation value ξ based on the data D15. Then, the fairness evaluation value calculation unit 222 outputs data including the acquired fairness evaluation value ξ to the detection period adjustment unit 223 as data D16.

The detection period adjustment unit 223 receives the fairness evaluation value ξ (data including the fairness evaluation value ξ) output from the fairness evaluation value calculation unit 222, and determines the detection period based on the fairness evaluation value ξ. process. Specifically, the detection period adjustment unit 223 performs (1) the number of time slots N (N: natural number) corresponding to the period for acquiring the CUR value, and (2) the channel switching process. The number of time slots Nc corresponding to the period until the channel switching process is executed is determined. Detection period adjusting section 223 outputs data including the determined numbers of time slots N and Nc to channel selection processing section 23 as data D17.

The channel selection processing unit 23 inputs the data D14 output from the channel probability processing unit 21 and the data D17 output from the fairness evaluation processing unit 22 . Then, the channel selection processing unit 23 performs channel selection processing based on data D14 (data containing the updated channel probability Pc) and data D17 (data containing the number of time slots N and Nc for determining the detection period). to run. Then, the channel selection processing unit 23 outputs data including the result of the channel selection processing to the communication processing unit 1 as data D2.

The data processing section 3 performs predetermined data processing on the data output from the communication processing section 1 . The data processing unit 3 also generates data to be transmitted to a predetermined destination and outputs the data to the communication processing unit 1 .

The access points AP2 to AP4 have the same configuration as the access point AP1.

<1.2: Operation of communication system>
The operation of the communication system 1000 configured as above will be described below.

4 and 5 are flowcharts for explaining the operation in communication system 1000. FIG. Specifically, it is a flow chart of processing executed by the access point AP1. The operation of the communication system 1000 will be described with reference to FIGS. 4-5.

(Step S1):
In step S1, the access point AP1 executes initialization processing. Specifically, the access point AP1 (1) sets the number of time slots corresponding to the period for acquiring the CUR value to an initial value, (2) executes the channel switching process, and then executes the channel switching process. The number of time slots Nc corresponding to the period until execution is set to an initial value, and (3) time t (discrete time step t, t: integer) is set to an initial value (t=0).

(Step S2):
In step S2, the access point AP1 performs a process of detecting available channels in frequency band i. Specifically, the available channel detection unit 211 of the access point AP1 detects the , find available channels at the access point AP1. The available channel detection unit 211 acquires data indicating available channels in the frequency band _i as ci. Note that the frequency band i is, for example, the ISM band (920 MHz band, 2.4 GHz band, 5 GHz band, 60 GHz band),
(1) 920 MHz band: i=1
(2) 2.4 GHz band: i=2
(3) 5 GHz band: i=3
(4) 60 GHz band: i=4
, and each frequency band and the value of i.

For convenience of explanation, a case where i is one (a case where one frequency band is used in the communication system 1000) will be explained below.

(Step S3):
In step S3, the average CUR value acquisition unit 212 of the access point AP1 executes a CUR value calculation process. Assuming that the CUR value of channel c of frequency band i of access point APb is ρ(AP _b , c _i ), the average CUR value acquisition unit 212 of access point AP1 obtains the CUR value ρ(AP _b , c _i ) of access point AP1. ) as ρ(AP _b , c _i )=N _busy /N
(b=1)
N: the number of time slots corresponding to the period during which the CUR value is acquired N _busy : In the period of the number of time slots N, the processing corresponding to the number of time slots determined to be “busy” (not “idle”) is performed. Get it by running

(Step S4):
In step S4, the average CUR value acquisition unit 212 of the access point AP1 performs average CUR value calculation processing. The average CUR value acquisition unit 212 of the access point AP1 obtains the average CUR value ave_ρ(AP _b , c _i ) (b=1) of the access point AP1 as
ave_ρ(AP _b ,c _i )=(1−ν)×ρ(AP _b ,c _i )+ν×ave_ρ(AP _b ,c _i )
(b=1)
ν: real number, 0≦ν≦1
Acquired by executing the process equivalent to .

に相当する処理により、取得する。 (Step S5):
In step S5, the Gibbs distribution calculation processing unit 213 of the access point AP1 calculates the Gibbs distribution Gibbs_dist(ave_ρ(AP _b , c _i )) (b=1) as

Acquired by a process equivalent to

に相当する処理を実行し、取得する。なお、ηは、忘却率であり、更新度合いを決める係数である。 (Step S6):
In step S6, the channel probability update processing unit 214 of the access point AP1 executes channel probability update processing. Specifically, channel probability update processing section 214 calculates probability Pc _i (probability of appearance of channel c _i in frequency band i) as

Acquire by executing the process equivalent to . Note that η is a forgetting rate, which is a coefficient that determines the degree of update.

ｖｅｃ＿ｃ：チャネル・アロケーション・ベクトル（各チャネルの割り当て状況を示すデータを要素とするベクトル）
Ｉ：周波数バンドの集合（通信システム１０００において使用する周波数バンドの集合）
Ｃ_ｉ：周波数バンドｉに含まれるチャネルの集合
Ｂ_ｃ：チャネルｃを使用するアクセスポイントの集合
Ｚ（Ｔ）：正規化のための定数
α：ＣＵＲの目標値
Ｒ_ｂ，ｃ：アクセスポイントＡＰ_ｂのチャネルｃについてのスループット
ξ：ＣＵＲ公平性評価値（ＣＵＲ Ｆａｉｒｎｅｓｓ ｉｎｄｅｘ）
とする。 Note that the state transition probability π(vec_c) for the channel in the communication system 1000 is

vec_c: channel allocation vector (a vector whose elements are data indicating the allocation status of each channel)
I: set of frequency bands (set of frequency bands used in communication system 1000)
C _i : set of channels included in frequency band i B _c : set of access points using channel c Z(T): constant for normalization α: target value of CUR R _b,c : access point AP _b Throughput for channel c ξ: CUR fairness index
and

Note that the channel allocation vector vec_c can be regarded as the state of a Markov chain. That is, the future (time t+1) behavior of the channel allocation vector vec_c is determined by its current value (time t) and is independent of its past behavior, thus ensuring Markovian property, and as a result, the channel allocation • The vector vec_c can be viewed as the state of a Markov chain.

Here, (1) throughput R _b,c for channel c of access point AP _b and (2) CUR fairness evaluation value ξ included in the above formula will be explained.

<<Throughput (R _b,c )>>
The throughput R _b,c for channel c of access point AP _b is defined, for example, as follows.

ζ_ｂ，ｃ∈（０，１］
Ｓ_ｂ：アクセスポイントＡＰ_ｂと通信する無線端末ｓの集合
Ｓ_ｂ，ｃ：アクセスポイントの集合Ｂに含まれるアクセスポイントＡＰ_ｂでチャネルｃを使用する無線端末ｓの集合
ζ_ｂ，ｃ（Ｂ_ｃ）：チャネルｃの使用における、アクセスポイントＡＰ_ｂによる使用の割合（通信システム１０００において、チャネルｃを使用しているアクセスポイントがアクセスポイントＡＰ_ｂだけである場合、この値は「１」となる。）
≪ＣＵＲ公平性評価値（ξ）≫
ＣＵＲ公平性評価値（ＣＵＲ Ｆａｉｒｎｅｓｓ ｉｎｄｅｘ）ξは、例えば、以下のように定義される。

ρ_ｂ，ｃ：アクセスポイントＡＰ_ｂのチャネルｃにおけるＣＵＲ値
なお、ＣＵＲ公平性評価値ξ（Ｒａｊ Ｊａｉｎ’ｓ ｅｑｕａｔｉｏｎに基づく値、０≦ξ≦１）は、全てのアクセスポイントがチャネルを均等に（公平に）使用している状態に近づけば近づく程、大きな値となり、全てのアクセスポイントがチャネルを均等に（公平に）使用している状態で、「１」となる。 Assuming that the set of wireless terminals included in the communication system 1000 is S and the element (corresponding to each wireless terminal) is s, the transmission time D(s) of the wireless terminal s is
D(s)=1/Φ(SINR(s))
Φ(): SINR of wireless terminal s (s ) (Signal-to-interference-plus-noise ratio). In the short term, the SINR of each wireless terminal affects the transmission time, but in the long term, each terminal is expected to achieve similar throughput, so throughput R _b for channel c of access point AP _{b , c} can be defined by the following equations.

ζ _b,c ∈(0,1)
S _b : set of wireless terminals s communicating with access point AP _b S _b,c : set of wireless terminals s using channel c at access point AP _b included in access point set B ζ _b,c (B _c ): percentage of use by access point AP _b in use of channel c (If access point AP _b is the only access point using channel c in communication system 1000, this value is '1'. )
<<CUR fairness evaluation value (ξ)>>
A CUR fairness index ξ is defined as follows, for example.

ρ _b,c : CUR value in channel c of access point AP _b Note that the CUR fairness evaluation value ξ (value based on Raj Jain's equation, 0≦ξ≦1) The closer the channel is used (fairly), the larger the value becomes.

上式の第１項は、通信システム１０００の全体のスループットを示しており、第２項は、公平性が失われたときのペナルティ値を示している。λ（λ＞０）は、ペナルティ値のための乗算係数である。 ≪Optimization Function≫
An optimization function for optimizing (maximizing) the overall performance of communication system 1000 will now be described. This optimization function is defined, for example, as follows.

The first term in the above equation indicates the overall throughput of communication system 1000, and the second term indicates the penalty value when fairness is lost. λ (λ>0) is the multiplication factor for the penalty value.

vec_c is a channel allocation vector, the j-th element of which indicates the status of the channel allocated to access point AP _j (j-th access point).

In order to optimize (maximize) the overall performance of the communication system 1000, the overall throughput of the communication system 1000 should be maximized while ensuring a balance of the CUR values of the available channels in the communication system 1000. Just do it. Let _{vec_copt} be the channel allocation vector that satisfies this, that is, maximizes the optimization function in the above equation (vec_copt _εC ^B ).

In order to solve the above optimization problem, a model using a Markov chain may be set in which the channel allocation vector is the state and the state transition probability is defined by the above (Equation 4). In order to solve an optimization problem using a Markov chain model, its probability distribution (target distribution) must converge to a steady state (equilibrium state) (it must be a stationary distribution). Gibbs sampling is effective in solving optimization problems using Markov chain models.

ｖｅｃ＿ｃ_－ｂ：アクセスポイントＡＰ_ｂ以外の全アクセスポイントのチャネル・アロケーション・ベクトル
上記数式により、状態が更新されるマルコフ連鎖によるモデルの確率分布は、定常状態（平衡状態）に収束する（ステーショナリ分布（stationary distribution）である）。 For example, if the current state vec_c is vec_c=vec_c(t), the access point AP _b updates its channel to c _b =c _b (t+1) according to the Markov chain update probability shown in the following formula.

vec_c _{− b} : Channel allocation vector of all access points other than access point AP _b According to the above formula, the probability distribution of the Markov chain model whose state is updated converges to a steady state (equilibrium state) (stationary distribution ( stationary distribution).

The reason is as follows.

When the access point AP _b updates its utilization resource based on the observation result of c _b (t), a state transition from c _b (t)=m to c _b (t+1)=n occurs, so this state transition depends only on the current channel allocation vector.

In addition, since the space CB that defines the channel allocation vector is finite at each time (discrete time) t, the model based on the Markov chain whose state is updated by the above formula is modeled as a time ^- inhomogeneous Markov chain. become. Since the state transition probability from an arbitrary state to another state has a positive value, it is possible to transition from an arbitrary state m to an arbitrary state n with a positive state transition probability. Therefore, the Markov chain whose state is updated by the above formula is irreducible and recurrent.

Furthermore, the Markov chain whose states are updated by the above equation is aperiodic due to self-loops with positive probabilities.

As described above, the Markov chain whose state is updated by the above formula has the above three properties, that is, (1) irreducible, and (2) recursive and (3) aperiodic. Therefore, in the Markov chain whose state is updated by the above formula, when T approaches 0, the state transition probability is guaranteed to converge to a static distribution (stationary distribution) (steady state (equilibrium state) ).

Gibbs sampling is useful for sampling in models based on Markov chains. In Gibbs sampling, according to the conditional probability of state transition (for example, the conditional probability of state transition in (Equation 8)), only one dimension of the state vector is changed to converge to a static distribution (stationary distribution). (so that the energy function is minimized), it is particularly effective when the conditional probability of state transition (for example, the conditional probability of state transition in (Formula 8)) is not complicated.

ｃ∈Ｃ
Ｔ：温度パラメータ（Ｔ＞０）
上式の確率分布は、マルコフ連鎖の確率密度関数の静的な分布（ステーショナリ分布）を有するという性質を有している。このことは、（Ａ）通信システム１０００におけるチャネルについての状態遷移確率π（ｖｅｃ＿ｃ）が（数式４）により規定されること、および、（Ｂ）（数式８）により状態が更新されるマルコフ連鎖が、（１）既約（可約でない：ｉｒｒｅｄｕｃｉｂｌｅ）であり、（２）再帰的（ｒｅｃｕｒｒｅｎｔ）であり、かつ、（３）非周期的（ａｐｅｒｉｏｄｉｃ）であることから証明される。 In the communication system 1000, the state transition probability to the target state c of the channel is calculated by sampling random numbers according to the probability distribution according to the following formula.

c∈C
T: temperature parameter (T>0)
The probability distribution of the above formula has the property of having a static distribution (stationary distribution) of the probability density function of the Markov chain. This means that (A) the state transition probability π(vec_c) for a channel in communication system 1000 is defined by (Equation 4), and (B) the Markov chain whose state is updated by (Equation 8) is , is (1) irreducible, (2) recurrent, and (3) aperiodic.

As described above, in step S6, the channel probability update processing unit 214 of the access point AP1 performs processing corresponding to (Equation 3) to update the channel probability. In the communication system 1000, optimization processing is performed by the Markov chain model modeled as described above. State )) is guaranteed to be updated.

(Step S7):
In step S7, the channel selection processing unit 23 calculates the remainder mod(Nc, t) of the number of time slots Nc for the discrete time t, and if mod(Nc, t)=0, the process proceeds to step S8. , mod (Nc, t)=0, the process proceeds to step S9.

(Step S8):
In step S8, the channel selection processor 23 performs channel selection based on the channel probability Pc. That is, the channel selection processing unit 23 selects the channel with the smallest channel probability Pc (the channel with the smallest CUR value and the largest available resource), and sends data D2 including the selection result to the communication processing unit 1. Output. The communication processing unit 1 performs baseband processing and RF processing so as to perform data communication using the channel selected by the channel selection processing unit 23 based on the data D2. As a result, the access point AP1 performs data communication using the selected channel.

(Step S9):
In step S9, the fairness evaluation processing unit 22 of the access point AP1 performs fairness evaluation processing (processing of steps S91 to S95).

(Step S91):
In step S91, the other AP information acquiring unit 221 acquires (1) the CUR value of the other access point (access point other than the access point AP1) and (2) the used channel from the data D1 output from the communication processing unit 1. Get information and Then, the other AP information acquisition unit 221 outputs data including the acquired data to the fairness evaluation value calculation unit 222 as data D15.

Note that the (1) CUR value and (2) information on the used channel of another access point (an access point other than the access point AP1) are data containing the information when the other access point accesses the access point AP1. Acquired by transmitting (for example, transmitting by broadcasting) and receiving the data at the access point AP1.

ζ_ｃｉ：チャネルｃ_ｉのシェア率
なお、上記数式による公平性評価値ξ（０≦ξ≦１）は、通信システム１０００に含まれるアクセスポイントにおいて、全てのチャネルが公平に使用されている場合、「１」に近づく。 (Step S92):
In step S<b>92 , the fairness evaluation value calculation unit 222 acquires the fairness evaluation value ξ based on the data D<b>15 output from the other AP information acquisition unit 221 . Specifically, the fairness evaluation value calculation unit 222 acquires the fairness evaluation value ξ by executing a process corresponding to the following formula.

ζ _ci : Share ratio of channel _ci Note that the fairness evaluation value ξ (0≦ξ≦1) obtained by the above formula is given by Close to "1".

(Steps S93-S95):
In step S93, the detection period adjustment unit 223 compares the fairness evaluation value ξ acquired by the fairness evaluation value calculation unit 222 with the threshold TH1, and if the fairness evaluation value ξ is equal to or greater than the threshold TH1, the process is started. Proceeding to step S94, the access point AP1 is made to wait for, for example, _Nn time slots, and does not perform the fairness evaluation process for a while. Then, after the period for _Nn time slots has passed, the process returns to step S91.

On the other hand, if the fairness evaluation value ξ is smaller than the threshold TH1, the process proceeds to step S95.

In step S95, the detection period adjustment unit 223 performs processing for determining the detection period. Specifically, the detection period adjustment unit 223 performs (1) the number of time slots N (N: natural number) corresponding to the period for acquiring the CUR value, and (2) the channel switching process. The number of time slots Nc corresponding to the period until the channel switching process is executed is determined.

This detection period determination process is performed by, for example, a fuzzy inference system in which the input is the fairness evaluation value ξ and the output is the number of time slots N and Nc, a system using a PNN (Probabilistic Neural Network), and other neural networks. It may be executed by the system or the like used.

For example, when the fairness evaluation value ξ is low, it can be seen that there is an imbalance in the usage of available channels in the communication system 1000. In this case, the detection period adjustment unit 223 reduces the value Nc to a small value. Adjust the value Nc so that Then, the channel selection processing unit 23 executes channel switching processing for each period corresponding to the value Nc adjusted by the detection period adjustment unit 223 . As a result, in the access point AP1, the number of channel selection processes executed in a certain period of time increases. As a result, in the entire system of the communication system 1000, there is a high possibility that the imbalance in the use of available channels will be eliminated.

After executing the above fairness evaluation processing, the fairness evaluation processing unit 22 advances the processing to step S10.

(Steps S10, S11):
In step S10, the access point AP1 advances the time (time step) by one (increments the discrete time step by one).

In step S11, the access point AP1 determines whether or not to end the channel probability update process, fairness evaluation process, channel selection process, etc. (processes in steps S2 to S10). If not, the process returns to step S2.

Other access points included in the communication system 1000 also perform the same processing as the above-described processing performed by the access point AP1.

As described above, in the communication system 1000, the concept of fairness of channel usage for each access point is introduced, and while the fairness of channel usage is not impaired, the throughput is calculated based on the sum of the throughput of each access point. The allocation of communication resources is determined so that the overall throughput of the communication system 1000 to be processed is increased. In other words, the communication system 1000 achieves optimization of communication resource distribution by obtaining a channel allocation vector that maximizes the evaluation function shown in (Formula 7). The first term of (Equation 7) is the overall throughput of communication system 1000 calculated based on the sum of the throughputs of each access point, and the second term of (Equation 7) is the channel usage at each access point. is the penalty value when the fairness of Therefore, by distributing the communication resources of the communication system 1000 based on the channel allocation vector that maximizes the evaluation function shown in (Formula 7), the overall throughput of the communication system 1000 is high, and each access point It is possible to realize a state in which the fairness of channel usage is ensured.

Then, in the communication system 1000, a model using a Markov chain is set in which the channel allocation vector is the state and the state transition probability is defined by (Equation 4), and the probability Pc of the channel utilization rate of each access point is set. is updated, it is guaranteed to converge to the above-described optimized state (the state in which the overall throughput of the communication system 1000 is high and the fairness of channel usage of each access point is ensured).

In the communication system 1000, each access point collects (1) the CUR value and (2) information on the used channel from other access points, and uses the collected data to determine the channel defined by the above model. By executing the process of updating the utilization rate probability Pc, the above-mentioned optimized state (the state in which the overall throughput of the communication system 1000 is high and the fairness of the channel utilization of each access point is ensured ), it is possible to converge to the state (channel utilization state) of the own device (own access point) according to ).

That is, in the communication system 1000, by processing as described above, distributed processing at each access point realizes local optimization (a state in which fairness in throughput and channel usage is ensured at each access point). In addition, optimization of the entire system (a state in which the throughput of the entire system and the fairness of channel usage of each access point are ensured) can be realized.

Accordingly, the communication system 1000 allows distributed optimization of the performance of the communication system 1000 .

A processing result in communication system 1000 will be described using FIG. For convenience of explanation, it is assumed that each device of the communication system 1000 is in the following state. In the communication system 1000, each access point (access points AP1 to AP4) has the same communication performance, each access point has the same amount of data processing per predetermined time, and each wireless terminal (wireless terminals ST11 to ST44) have the same communication performance, each wireless terminal has the same amount of data processing per predetermined time, and the distance from each wireless terminal to the access point in charge is the same. The available channels in communication system 1000 are channels Ch_1 and Ch_2, each having the same bandwidth.

It is also assumed that the channel usage status (channel occupancy) of each access point (access points AP1 to AP4) at time t=t0 is in the state (unbalanced state) shown in the left diagram of FIG. In the left diagram of FIG. 6, each channel utilization rate of the entire communication system 1000 is shown. As can be seen from the left diagram of FIG. 6, the channel utilization rate of each access point varies.

At this time, by executing the above processing at the access points AP1 to AP4 of the communication system 1000, the channel probability Pc is updated so that the utilization rate of each channel at each access point becomes equal, and sufficient time has passed. Then, it converges to the state (equilibrium state) shown in the right diagram of FIG. In the right diagram of FIG. 6, each channel utilization rate (balanced state) of the entire communication system 1000 is shown. As can be seen from the right diagram of FIG. 6, the channel utilization rate of each access point is uniform.

In this way, in the communication system 1000, it is guaranteed that the overall system throughput is increased and that the channel utilization rate converges to a fair state. Achieving optimization of the entire system (a state in which the throughput of the entire system and fairness in the use of channels for each access point is ensured) while realizing fairness in the use of channels at access points). can do.

Accordingly, the communication system 1000 allows distributed optimization of the performance of the communication system 1000 .

[Other embodiments]
Although the network configuration of the communication system 1000 shown in FIG. 2 has been described in the above embodiment, the network configuration of the communication system 1000 is not limited to that shown in FIG. Also, the number and arrangement of access points and the number and arrangement of wireless terminals are not limited to those shown above, and communication system 1000 may include other numbers of access points and wireless terminals. may Also, the arrangement of access points and the arrangement of wireless terminals may be arbitrary.

Also, the network topology of the communication system 1000 is not limited to that shown in the above embodiment, and may be another topology.

In the above embodiment, the case of optimizing the performance of the communication system 1000 based on the channel utilization rate has been described, but the present invention is not limited to this, and other physical quantities ( For example, the above technique may be applied to optimize the performance of the communication system 1000 based on the radio field strength, etc.).

Further, in the above embodiment, the case of optimizing the performance of the communication system 1000 for one frequency band has been described, but it is not limited to this, and in the communication system 1000, using a plurality of frequency bands, A plurality of channels may be set for each frequency band.

Further, in the communication system and access point described in the above embodiments, each block may be individually integrated into one chip by a semiconductor device such as LSI, or may be integrated into one chip so as to include part or all of the blocks. good.

Although LSI is used here, it may also be called IC, system LSI, super LSI, or ultra LSI depending on the degree of integration.

Further, the method of circuit integration is not limited to LSI, and may be implemented by a dedicated circuit or a general-purpose processor. An FPGA (Field Programmable Gate Array) that can be programmed after the LSI is manufactured, or a reconfigurable processor that can reconfigure the connections and settings of the circuit cells inside the LSI may be used.

Also, part or all of the processing of each functional block in each of the above embodiments may be implemented by a program. Part or all of the processing of each functional block in each of the above embodiments is performed by a central processing unit (CPU) in a computer. A program for performing each process is stored in a storage device such as a hard disk or ROM, and is read from the ROM or RAM and executed.

Further, each process of the above embodiments may be realized by hardware, or may be realized by software (including cases where it is realized together with an OS (operating system), middleware, or a predetermined library). Furthermore, it may be realized by mixed processing of software and hardware.

For example, when each functional unit of the above embodiment is realized by software, the hardware configuration shown in FIG. may be used to realize each functional unit by software processing.

Further, when each functional unit of the above embodiment is realized by software, the software may be realized using a single computer having the hardware configuration shown in FIG. may be realized by distributed processing using .

Also, the execution order of the processing methods in the above embodiments is not necessarily limited to the description of the above embodiments, and the execution order can be changed without departing from the gist of the invention.

A computer program that causes a computer to execute the method described above and a computer-readable recording medium that records the program are included in the scope of the present invention. Examples of computer-readable recording media include flexible disks, hard disks, CD-ROMs, MOs, DVDs, DVD-ROMs, DVD-RAMs, large-capacity DVDs, next-generation DVDs, and semiconductor memories. .

The computer program is not limited to being recorded on the recording medium, and may be transmitted via an electric communication line, a wireless or wired communication line, a network represented by the Internet, or the like.

The specific configuration of the present invention is not limited to the above-described embodiment, and various changes and modifications are possible without departing from the gist of the invention.

1000 communication system AP1 to AP4 access points ST11 to ST44 wireless terminal 1 communication processing unit 2 communication resource adjustment unit 21 channel probability processing unit 22 fairness evaluation processing unit 23 channel selection processing unit

Claims (8)

Communication comprising a plurality of terminals and a plurality of access points each capable of communicating with some or all of the plurality of terminals, and performing data communication using a plurality of channels for each of a plurality of frequency bands a system,
Each access point
a channel usage data acquiring step of acquiring channel usage data, which is data indicating the usage status of each channel of each frequency band of the own access point, in the own access point that is the own device;
Based on the channel usage data, the channel probability of each channel of each frequency band of the own access point is set so that the greater the available resource of the channel, the higher the channel probability, which is the probability that the channel will be selected. an update step to update;
a receiving step of receiving data indicating usage status of each channel of each frequency band from other access points;
a fairness evaluation value obtaining step of obtaining a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point;
Based on the fairness evaluation value, a channel selection period that is the period of channel selection processing that is processing for selecting a channel, and a channel that is data indicating the usage status of each channel of each frequency band of the access point itself. a channel selection period adjustment step of adjusting at least one of the channel usage data acquisition cycle, which is a cycle for acquiring usage data;
a channel selection step of executing a channel selection process for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection period;
run the
Communications system. The channel probability is calculated by a probability distribution according to a Gibbs distribution to which Gibbs sampling can be applied.
A communication system according to claim 1. The channel usage data is data derived from a CUR value that indicates the occupancy rate of a channel in a predetermined period.
A communication system according to claim 1 or 2. Let AP _b be the access point, i be the frequency band, c _i be the channel included in frequency band i, ζ c _i be the share rate of channel c _{i, and ρ be the CUR value of channel c i of} access point AP _b . (AP _b , c _i ), then
The fairness evaluation value is

calculated by
A communication system according to claim 3. The access point
Sending data indicating the usage status of each channel of each frequency band at the own access point to the outside by broadcasting,
A communication system according to any one of claims 1 to 4. Communication comprising a plurality of terminals and a plurality of access points each capable of communicating with some or all of the plurality of terminals, and performing data communication using a plurality of channels for each of a plurality of frequency bands An access point used in the system,
a control unit;
a memory holding at least one of instructions and data for performing predetermined processing;
with
a channel usage data acquiring step of acquiring channel usage data, which is data indicating the usage status of each channel of each frequency band of the own access point, in the own access point that is the own device;
Based on the channel usage data, the channel probability of each channel of each frequency band of the own access point is set so that the greater the available resource of the channel, the higher the channel probability, which is the probability that the channel will be selected. an update step to update;
a receiving step of receiving data indicating usage status of each channel of each frequency band from other access points;
a fairness evaluation value obtaining step of obtaining a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point;
Based on the fairness evaluation value, a channel selection period that is the period of channel selection processing that is processing for selecting a channel, and a channel that is data indicating the usage status of each channel of each frequency band of the access point itself. a channel selection period adjustment step of adjusting at least one of the channel usage data acquisition cycle, which is a cycle for acquiring usage data;
a channel selection step of executing a channel selection process for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection period;
run the
access point. Communication comprising a plurality of terminals and a plurality of access points each capable of communicating with some or all of the plurality of terminals, and performing data communication using a plurality of channels for each of a plurality of frequency bands A communication method for use in a system, comprising:
a channel usage data acquisition step of acquiring, by each access point, channel usage data, which is data indicating the usage status of each channel of each frequency band of the own access point, in the own access point, which is the own device;
Each channel in each frequency band of each access point is set so that the channel probability, which is the probability that the channel is selected, increases as the available resources of the channel increase based on the channel usage data. an updating step of updating the channel probability of
a receiving step of receiving data indicating usage status of each channel of each frequency band from other access points;
a fairness evaluation value obtaining step of obtaining, by each access point, a fairness evaluation value indicating the fairness of the value indicating the usage status of each channel of each frequency band for each access point;
Each access point, based on the fairness evaluation value, determines the channel selection period, which is the period of channel selection processing, and the usage status of each channel of each frequency band of the own access point. a channel selection period adjustment step of adjusting at least one of a channel usage data acquisition cycle that is a cycle for acquiring channel usage data that is the data indicated;
a channel selection step in which each access point performs channel selection processing for selecting a channel for data communication based on the channel probability at a timing determined by the channel selection period;
A communication method comprising: A program for causing a computer to execute the communication method according to claim 7.

Images